[{"data":1,"prerenderedAt":29202},["ShallowReactive",2],{"lesson:\u002Fdeep-learning\u002Farchitectures\u002Fattention-and-transformers":3,"course-wordcounts":23424,"ref-card-index":24336,"nav:deep-learning":29017,"tikz:275eb1406d1142114f692c18df30c935f3f30b3502e653aee85f0b98064c0a44":29193,"tikz:98babe1adf74ccc313bc8d01040b34a1d80d5de4bd89ed7550429f54c7bdd188":29194,"tikz:e1ce3af2d701a9d05bee403986619251fea3b30dced65a3e76c4ee9255eb5c28":29195,"tikz:f1304c22c8b585b2a338c2145110fcb5c41434dba20c62581f1d8aa3db154a18":29196,"tikz:084444461fe4f278f3000a37434b8480999198ffaf9bd2347d72be5e67ede4b1":29197,"tikz:685cf0eaa8fde0a43cac0b1af564addde3061d23fa6822815d8442148858c098":29198,"tikz:1081d93c9d7cfa7915db69065a0b79c97f6afe1c41346cf0827b9cef51f21481":29199,"tikz:f28890382bf7cbd4ae1d5275638982b367015f083a1d98df7fc7281fbe5f1b0e":29200,"tikz:ccb2861b1d8e505198e6b51f70fe7b1c94937c2596a9ad750b31991176c26408":29201},{"id":4,"title":5,"blurb":6,"body":7,"brief":23396,"category":23397,"description":23398,"draft":23399,"extension":23400,"meta":23401,"module":23403,"navigation":23404,"path":23405,"practice":23406,"rawbody":23407,"readingTime":23408,"seo":23413,"sources":23414,"status":23419,"stem":23420,"summary":23421,"topics":23422,"__hash__":23423},"course\u002F06.deep-learning\u002F05.architectures\u002F05.attention-and-transformers.md","Attention & Transformers","",{"type":8,"value":9,"toc":23375},"minimark",[10,176,615,620,635,1505,1579,2127,2131,2135,2502,3676,4367,5116,5229,5331,5451,6367,7814,8350,8353,8417,8421,8550,9174,9645,10051,10236,10642,10942,11392,11556,11942,12260,12267,12271,12352,12596,12602,12980,13144,13147,13151,13174,13500,14246,14256,14260,14289,15005,15008,15672,16436,16641,17211,17347,17350,17354,17431,17958,18040,18043,18046,18504,18508,18534,18858,18946,18949,18953,18960,19009,19257,19704,19708,20120,21199,21246,21249,21253,21267,21702,21881,21884,21888,21891,22453,22460,22464,23093,23371],[11,12,13,14,19,20,99,100,117,118,135,136,151,152,156,157,161,162,166,167],"p",{},"A ",[15,16,18],"a",{"href":17},"\u002Fdeep-learning\u002Farchitectures\u002Frecurrent-networks","recurrent network"," reads a\nsequence one step at a time, squeezing the entire past into a single hidden state\n",[21,22,25],"span",{"className":23},[24],"katex",[21,26,30],{"className":27,"ariaHidden":29},[28],"katex-html","true",[21,31,34,39],{"className":32},[33],"base",[21,35],{"className":36,"style":38},[37],"strut","height:0.8444em;vertical-align:-0.15em;",[21,40,43,48],{"className":41},[42],"mord",[21,44,47],{"className":45},[42,46],"mathnormal","h",[21,49,52],{"className":50},[51],"msupsub",[21,53,57,90],{"className":54},[55,56],"vlist-t","vlist-t2",[21,58,61,85],{"className":59},[60],"vlist-r",[21,62,66],{"className":63,"style":65},[64],"vlist","height:0.2806em;",[21,67,69,74],{"style":68},"top:-2.55em;margin-left:0em;margin-right:0.05em;",[21,70],{"className":71,"style":73},[72],"pstrut","height:2.7em;",[21,75,81],{"className":76},[77,78,79,80],"sizing","reset-size6","size3","mtight",[21,82,84],{"className":83},[42,46,80],"t",[21,86,89],{"className":87},[88],"vlist-s","​",[21,91,93],{"className":92},[60],[21,94,97],{"className":95,"style":96},[64],"height:0.15em;",[21,98],{},". That bottleneck is the defect: information from position ",[21,101,103],{"className":102},[24],[21,104,106],{"className":105,"ariaHidden":29},[28],[21,107,109,113],{"className":108},[33],[21,110],{"className":111,"style":112},[37],"height:0.6444em;",[21,114,116],{"className":115},[42],"1"," reaches\nposition ",[21,119,121],{"className":120},[24],[21,122,124],{"className":123,"ariaHidden":29},[28],[21,125,127,131],{"className":126},[33],[21,128],{"className":129,"style":130},[37],"height:0.4306em;",[21,132,134],{"className":133},[42,46],"n"," only after ",[21,137,139],{"className":138},[24],[21,140,142],{"className":141,"ariaHidden":29},[28],[21,143,145,148],{"className":144},[33],[21,146],{"className":147,"style":130},[37],[21,149,134],{"className":150},[42,46]," sequential hops, and the gradient connecting them\ndecays along the way: the very problem ",[15,153,155],{"href":154},"\u002Fdeep-learning\u002Farchitectures\u002Flstm-and-gru","LSTMs and\nGRUs"," only partially patch.\n",[158,159,160],"strong",{},"Attention"," removes the hops entirely. Instead of routing information through a\nchain, every position reads directly from every other position in ",[163,164,165],"em",{},"one"," step,\nweighting what it reads by learned, content-dependent relevance.",[168,169,170],"sup",{},[15,171,116],{"href":172,"ariaDescribedBy":173,"dataFootnoteRef":6,"id":175},"#user-content-fn-gf-attention",[174],"footnote-label","user-content-fnref-gf-attention",[177,178,180],"callout",{"type":179},"definition",[11,181,182,185,186,204,205,275,276,336,337,486,487,540,541,556,557,610,611,614],{},[158,183,184],{},"Definition (Attention)."," A differentiable lookup. Given a query vector ",[21,187,189],{"className":188},[24],[21,190,192],{"className":191,"ariaHidden":29},[28],[21,193,195,199],{"className":194},[33],[21,196],{"className":197,"style":198},[37],"height:0.625em;vertical-align:-0.1944em;",[21,200,203],{"className":201,"style":202},[42,46],"margin-right:0.0359em;","q",", a\nset of keys ",[21,206,208],{"className":207},[24],[21,209,211],{"className":210,"ariaHidden":29},[28],[21,212,214,218,223,270],{"className":213},[33],[21,215],{"className":216,"style":217},[37],"height:1.0361em;vertical-align:-0.2861em;",[21,219,222],{"className":220},[221],"mopen","{",[21,224,226,231],{"className":225},[42],[21,227,230],{"className":228,"style":229},[42,46],"margin-right:0.0315em;","k",[21,232,234],{"className":233},[51],[21,235,237,261],{"className":236},[55,56],[21,238,240,258],{"className":239},[60],[21,241,244],{"className":242,"style":243},[64],"height:0.3117em;",[21,245,247,250],{"style":246},"top:-2.55em;margin-left:-0.0315em;margin-right:0.05em;",[21,248],{"className":249,"style":73},[72],[21,251,253],{"className":252},[77,78,79,80],[21,254,257],{"className":255,"style":256},[42,46,80],"margin-right:0.0572em;","j",[21,259,89],{"className":260},[88],[21,262,264],{"className":263},[60],[21,265,268],{"className":266,"style":267},[64],"height:0.2861em;",[21,269],{},[21,271,274],{"className":272},[273],"mclose","}",", and an associated set of values ",[21,277,279],{"className":278},[24],[21,280,282],{"className":281,"ariaHidden":29},[28],[21,283,285,288,291,333],{"className":284},[33],[21,286],{"className":287,"style":217},[37],[21,289,222],{"className":290},[221],[21,292,294,298],{"className":293},[42],[21,295,297],{"className":296,"style":202},[42,46],"v",[21,299,301],{"className":300},[51],[21,302,304,325],{"className":303},[55,56],[21,305,307,322],{"className":306},[60],[21,308,310],{"className":309,"style":243},[64],[21,311,313,316],{"style":312},"top:-2.55em;margin-left:-0.0359em;margin-right:0.05em;",[21,314],{"className":315,"style":73},[72],[21,317,319],{"className":318},[77,78,79,80],[21,320,257],{"className":321,"style":256},[42,46,80],[21,323,89],{"className":324},[88],[21,326,328],{"className":327},[60],[21,329,331],{"className":330,"style":267},[64],[21,332],{},[21,334,274],{"className":335},[273],", attention\nreturns a convex combination ",[21,338,340],{"className":339},[24],[21,341,343],{"className":342,"ariaHidden":29},[28],[21,344,346,350,398,403,446],{"className":345},[33],[21,347],{"className":348,"style":349},[37],"height:1.1858em;vertical-align:-0.4358em;",[21,351,354,361],{"className":352},[353],"mop",[21,355,360],{"className":356,"style":359},[353,357,358],"op-symbol","small-op","position:relative;top:0em;","∑",[21,362,364],{"className":363},[51],[21,365,367,389],{"className":366},[55,56],[21,368,370,386],{"className":369},[60],[21,371,374],{"className":372,"style":373},[64],"height:0.162em;",[21,375,377,380],{"style":376},"top:-2.4003em;margin-left:0em;margin-right:0.05em;",[21,378],{"className":379,"style":73},[72],[21,381,383],{"className":382},[77,78,79,80],[21,384,257],{"className":385,"style":256},[42,46,80],[21,387,89],{"className":388},[88],[21,390,392],{"className":391},[60],[21,393,396],{"className":394,"style":395},[64],"height:0.4358em;",[21,397],{},[21,399],{"className":400,"style":402},[401],"mspace","margin-right:0.1667em;",[21,404,406,411],{"className":405},[42],[21,407,410],{"className":408,"style":409},[42,46],"margin-right:0.0037em;","α",[21,412,414],{"className":413},[51],[21,415,417,438],{"className":416},[55,56],[21,418,420,435],{"className":419},[60],[21,421,423],{"className":422,"style":243},[64],[21,424,426,429],{"style":425},"top:-2.55em;margin-left:-0.0037em;margin-right:0.05em;",[21,427],{"className":428,"style":73},[72],[21,430,432],{"className":431},[77,78,79,80],[21,433,257],{"className":434,"style":256},[42,46,80],[21,436,89],{"className":437},[88],[21,439,441],{"className":440},[60],[21,442,444],{"className":443,"style":267},[64],[21,445],{},[21,447,449,452],{"className":448},[42],[21,450,297],{"className":451,"style":202},[42,46],[21,453,455],{"className":454},[51],[21,456,458,478],{"className":457},[55,56],[21,459,461,475],{"className":460},[60],[21,462,464],{"className":463,"style":243},[64],[21,465,466,469],{"style":312},[21,467],{"className":468,"style":73},[72],[21,470,472],{"className":471},[77,78,79,80],[21,473,257],{"className":474,"style":256},[42,46,80],[21,476,89],{"className":477},[88],[21,479,481],{"className":480},[60],[21,482,484],{"className":483,"style":267},[64],[21,485],{}," of the values, where the\nweights ",[21,488,490],{"className":489},[24],[21,491,493],{"className":492,"ariaHidden":29},[28],[21,494,496,500],{"className":495},[33],[21,497],{"className":498,"style":499},[37],"height:0.7167em;vertical-align:-0.2861em;",[21,501,503,506],{"className":502},[42],[21,504,410],{"className":505,"style":409},[42,46],[21,507,509],{"className":508},[51],[21,510,512,532],{"className":511},[55,56],[21,513,515,529],{"className":514},[60],[21,516,518],{"className":517,"style":243},[64],[21,519,520,523],{"style":425},[21,521],{"className":522,"style":73},[72],[21,524,526],{"className":525},[77,78,79,80],[21,527,257],{"className":528,"style":256},[42,46,80],[21,530,89],{"className":531},[88],[21,533,535],{"className":534},[60],[21,536,538],{"className":537,"style":267},[64],[21,539],{}," measure the compatibility between ",[21,542,544],{"className":543},[24],[21,545,547],{"className":546,"ariaHidden":29},[28],[21,548,550,553],{"className":549},[33],[21,551],{"className":552,"style":198},[37],[21,554,203],{"className":555,"style":202},[42,46]," and each key ",[21,558,560],{"className":559},[24],[21,561,563],{"className":562,"ariaHidden":29},[28],[21,564,566,570],{"className":565},[33],[21,567],{"className":568,"style":569},[37],"height:0.9805em;vertical-align:-0.2861em;",[21,571,573,576],{"className":572},[42],[21,574,230],{"className":575,"style":229},[42,46],[21,577,579],{"className":578},[51],[21,580,582,602],{"className":581},[55,56],[21,583,585,599],{"className":584},[60],[21,586,588],{"className":587,"style":243},[64],[21,589,590,593],{"style":246},[21,591],{"className":592,"style":73},[72],[21,594,596],{"className":595},[77,78,79,80],[21,597,257],{"className":598,"style":256},[42,46,80],[21,600,89],{"className":601},[88],[21,603,605],{"className":604},[60],[21,606,608],{"className":607,"style":267},[64],[21,609],{},". It\nis a ",[163,612,613],{},"soft"," dictionary: a hard lookup returns the one value whose key matches;\nattention returns a weighted blend of all of them.",[616,617,619],"h2",{"id":618},"queries-keys-and-values","Queries, keys, and values",[11,621,622,623,626,627,630,631,634],{},"The vocabulary is borrowed from retrieval. A ",[158,624,625],{},"query"," is what a position is\nlooking for; a ",[158,628,629],{},"key"," advertises what each position offers; a ",[158,632,633],{},"value"," is the\ncontent actually returned. All three are linear projections of the inputs, so the\nmechanism is fully learned: three weight matrices determine what each token\nsearches for, what it exposes for matching, and what content it returns.",[636,637,638,657],"table",{},[639,640,641],"thead",{},[642,643,644,648,651,654],"tr",{},[645,646,647],"th",{},"Object",[645,649,650],{},"Symbol",[645,652,653],{},"Shape",[645,655,656],{},"Role",[658,659,660,848,1018,1189,1291],"tbody",{},[642,661,662,666,751,845],{},[663,664,665],"td",{},"Query",[663,667,668],{},[21,669,671],{"className":670},[24],[21,672,674,695],{"className":673,"ariaHidden":29},[28],[21,675,677,680,683,687,692],{"className":676},[33],[21,678],{"className":679,"style":198},[37],[21,681,203],{"className":682,"style":202},[42,46],[21,684],{"className":685,"style":686},[401],"margin-right:0.2778em;",[21,688,691],{"className":689},[690],"mrel","=",[21,693],{"className":694,"style":686},[401],[21,696,698,702,706],{"className":697},[33],[21,699],{"className":700,"style":701},[37],"height:0.9694em;vertical-align:-0.2861em;",[21,703,705],{"className":704},[42,46],"x",[21,707,709,714],{"className":708},[42],[21,710,713],{"className":711,"style":712},[42,46],"margin-right:0.1389em;","W",[21,715,717],{"className":716},[51],[21,718,720,743],{"className":719},[55,56],[21,721,723,740],{"className":722},[60],[21,724,727],{"className":725,"style":726},[64],"height:0.3283em;",[21,728,730,733],{"style":729},"top:-2.55em;margin-left:-0.1389em;margin-right:0.05em;",[21,731],{"className":732,"style":73},[72],[21,734,736],{"className":735},[77,78,79,80],[21,737,739],{"className":738},[42,46,80],"Q",[21,741,89],{"className":742},[88],[21,744,746],{"className":745},[60],[21,747,749],{"className":748,"style":267},[64],[21,750],{},[663,752,753],{},[21,754,756],{"className":755},[24],[21,757,759],{"className":758,"ariaHidden":29},[28],[21,760,762,766],{"className":761},[33],[21,763],{"className":764,"style":765},[37],"height:0.8491em;",[21,767,769,774],{"className":768},[42],[21,770,773],{"className":771},[42,772],"mathbb","R",[21,775,777],{"className":776},[51],[21,778,780],{"className":779},[55],[21,781,783],{"className":782},[60],[21,784,786],{"className":785,"style":765},[64],[21,787,789,792],{"style":788},"top:-3.063em;margin-right:0.05em;",[21,790],{"className":791,"style":73},[72],[21,793,795],{"className":794},[77,78,79,80],[21,796,798],{"className":797},[42,80],[21,799,801,805],{"className":800},[42,80],[21,802,804],{"className":803},[42,46,80],"d",[21,806,808],{"className":807},[51],[21,809,811,836],{"className":810},[55,56],[21,812,814,833],{"className":813},[60],[21,815,818],{"className":816,"style":817},[64],"height:0.3448em;",[21,819,821,825],{"style":820},"top:-2.3488em;margin-left:0em;margin-right:0.0714em;",[21,822],{"className":823,"style":824},[72],"height:2.5em;",[21,826,830],{"className":827},[77,828,829,80],"reset-size3","size1",[21,831,230],{"className":832,"style":229},[42,46,80],[21,834,89],{"className":835},[88],[21,837,839],{"className":838},[60],[21,840,843],{"className":841,"style":842},[64],"height:0.1512em;",[21,844],{},[663,846,847],{},"what this position is searching for",[642,849,850,853,932,1015],{},[663,851,852],{},"Key",[663,854,855],{},[21,856,858],{"className":857},[24],[21,859,861,880],{"className":860,"ariaHidden":29},[28],[21,862,864,868,871,874,877],{"className":863},[33],[21,865],{"className":866,"style":867},[37],"height:0.6944em;",[21,869,230],{"className":870,"style":229},[42,46],[21,872],{"className":873,"style":686},[401],[21,875,691],{"className":876},[690],[21,878],{"className":879,"style":686},[401],[21,881,883,887,890],{"className":882},[33],[21,884],{"className":885,"style":886},[37],"height:0.8333em;vertical-align:-0.15em;",[21,888,705],{"className":889},[42,46],[21,891,893,896],{"className":892},[42],[21,894,713],{"className":895,"style":712},[42,46],[21,897,899],{"className":898},[51],[21,900,902,924],{"className":901},[55,56],[21,903,905,921],{"className":904},[60],[21,906,908],{"className":907,"style":726},[64],[21,909,910,913],{"style":729},[21,911],{"className":912,"style":73},[72],[21,914,916],{"className":915},[77,78,79,80],[21,917,920],{"className":918,"style":919},[42,46,80],"margin-right:0.0715em;","K",[21,922,89],{"className":923},[88],[21,925,927],{"className":926},[60],[21,928,930],{"className":929,"style":96},[64],[21,931],{},[663,933,934],{},[21,935,937],{"className":936},[24],[21,938,940],{"className":939,"ariaHidden":29},[28],[21,941,943,946],{"className":942},[33],[21,944],{"className":945,"style":765},[37],[21,947,949,952],{"className":948},[42],[21,950,773],{"className":951},[42,772],[21,953,955],{"className":954},[51],[21,956,958],{"className":957},[55],[21,959,961],{"className":960},[60],[21,962,964],{"className":963,"style":765},[64],[21,965,966,969],{"style":788},[21,967],{"className":968,"style":73},[72],[21,970,972],{"className":971},[77,78,79,80],[21,973,975],{"className":974},[42,80],[21,976,978,981],{"className":977},[42,80],[21,979,804],{"className":980},[42,46,80],[21,982,984],{"className":983},[51],[21,985,987,1007],{"className":986},[55,56],[21,988,990,1004],{"className":989},[60],[21,991,993],{"className":992,"style":817},[64],[21,994,995,998],{"style":820},[21,996],{"className":997,"style":824},[72],[21,999,1001],{"className":1000},[77,828,829,80],[21,1002,230],{"className":1003,"style":229},[42,46,80],[21,1005,89],{"className":1006},[88],[21,1008,1010],{"className":1009},[60],[21,1011,1013],{"className":1012,"style":842},[64],[21,1014],{},[663,1016,1017],{},"what each position offers for matching",[642,1019,1020,1023,1100,1186],{},[663,1021,1022],{},"Value",[663,1024,1025],{},[21,1026,1028],{"className":1027},[24],[21,1029,1031,1049],{"className":1030,"ariaHidden":29},[28],[21,1032,1034,1037,1040,1043,1046],{"className":1033},[33],[21,1035],{"className":1036,"style":130},[37],[21,1038,297],{"className":1039,"style":202},[42,46],[21,1041],{"className":1042,"style":686},[401],[21,1044,691],{"className":1045},[690],[21,1047],{"className":1048,"style":686},[401],[21,1050,1052,1055,1058],{"className":1051},[33],[21,1053],{"className":1054,"style":886},[37],[21,1056,705],{"className":1057},[42,46],[21,1059,1061,1064],{"className":1060},[42],[21,1062,713],{"className":1063,"style":712},[42,46],[21,1065,1067],{"className":1066},[51],[21,1068,1070,1092],{"className":1069},[55,56],[21,1071,1073,1089],{"className":1072},[60],[21,1074,1076],{"className":1075,"style":726},[64],[21,1077,1078,1081],{"style":729},[21,1079],{"className":1080,"style":73},[72],[21,1082,1084],{"className":1083},[77,78,79,80],[21,1085,1088],{"className":1086,"style":1087},[42,46,80],"margin-right:0.2222em;","V",[21,1090,89],{"className":1091},[88],[21,1093,1095],{"className":1094},[60],[21,1096,1098],{"className":1097,"style":96},[64],[21,1099],{},[663,1101,1102],{},[21,1103,1105],{"className":1104},[24],[21,1106,1108],{"className":1107,"ariaHidden":29},[28],[21,1109,1111,1114],{"className":1110},[33],[21,1112],{"className":1113,"style":765},[37],[21,1115,1117,1120],{"className":1116},[42],[21,1118,773],{"className":1119},[42,772],[21,1121,1123],{"className":1122},[51],[21,1124,1126],{"className":1125},[55],[21,1127,1129],{"className":1128},[60],[21,1130,1132],{"className":1131,"style":765},[64],[21,1133,1134,1137],{"style":788},[21,1135],{"className":1136,"style":73},[72],[21,1138,1140],{"className":1139},[77,78,79,80],[21,1141,1143],{"className":1142},[42,80],[21,1144,1146,1149],{"className":1145},[42,80],[21,1147,804],{"className":1148},[42,46,80],[21,1150,1152],{"className":1151},[51],[21,1153,1155,1177],{"className":1154},[55,56],[21,1156,1158,1174],{"className":1157},[60],[21,1159,1162],{"className":1160,"style":1161},[64],"height:0.1645em;",[21,1163,1165,1168],{"style":1164},"top:-2.357em;margin-left:0em;margin-right:0.0714em;",[21,1166],{"className":1167,"style":824},[72],[21,1169,1171],{"className":1170},[77,828,829,80],[21,1172,297],{"className":1173,"style":202},[42,46,80],[21,1175,89],{"className":1176},[88],[21,1178,1180],{"className":1179},[60],[21,1181,1184],{"className":1182,"style":1183},[64],"height:0.143em;",[21,1185],{},[663,1187,1188],{},"the content returned when a key matches",[642,1190,1191,1194,1269,1272],{},[663,1192,1193],{},"Score",[663,1195,1196],{},[21,1197,1199],{"className":1198},[24],[21,1200,1202,1223],{"className":1201,"ariaHidden":29},[28],[21,1203,1205,1209,1212,1215,1220],{"className":1204},[33],[21,1206],{"className":1207,"style":1208},[37],"height:0.6389em;vertical-align:-0.1944em;",[21,1210,203],{"className":1211,"style":202},[42,46],[21,1213],{"className":1214,"style":1087},[401],[21,1216,1219],{"className":1217},[1218],"mbin","⋅",[21,1221],{"className":1222,"style":1087},[401],[21,1224,1226,1229],{"className":1225},[33],[21,1227],{"className":1228,"style":569},[37],[21,1230,1232,1235],{"className":1231},[42],[21,1233,230],{"className":1234,"style":229},[42,46],[21,1236,1238],{"className":1237},[51],[21,1239,1241,1261],{"className":1240},[55,56],[21,1242,1244,1258],{"className":1243},[60],[21,1245,1247],{"className":1246,"style":243},[64],[21,1248,1249,1252],{"style":246},[21,1250],{"className":1251,"style":73},[72],[21,1253,1255],{"className":1254},[77,78,79,80],[21,1256,257],{"className":1257,"style":256},[42,46,80],[21,1259,89],{"className":1260},[88],[21,1262,1264],{"className":1263},[60],[21,1265,1267],{"className":1266,"style":267},[64],[21,1268],{},[663,1270,1271],{},"scalar",[663,1273,1274,1275],{},"raw compatibility of the query with key ",[21,1276,1278],{"className":1277},[24],[21,1279,1281],{"className":1280,"ariaHidden":29},[28],[21,1282,1284,1288],{"className":1283},[33],[21,1285],{"className":1286,"style":1287},[37],"height:0.854em;vertical-align:-0.1944em;",[21,1289,257],{"className":1290,"style":256},[42,46],[642,1292,1293,1296,1350,1389],{},[663,1294,1295],{},"Weight",[663,1297,1298],{},[21,1299,1301],{"className":1300},[24],[21,1302,1304],{"className":1303,"ariaHidden":29},[28],[21,1305,1307,1310],{"className":1306},[33],[21,1308],{"className":1309,"style":499},[37],[21,1311,1313,1316],{"className":1312},[42],[21,1314,410],{"className":1315,"style":409},[42,46],[21,1317,1319],{"className":1318},[51],[21,1320,1322,1342],{"className":1321},[55,56],[21,1323,1325,1339],{"className":1324},[60],[21,1326,1328],{"className":1327,"style":243},[64],[21,1329,1330,1333],{"style":425},[21,1331],{"className":1332,"style":73},[72],[21,1334,1336],{"className":1335},[77,78,79,80],[21,1337,257],{"className":1338,"style":256},[42,46,80],[21,1340,89],{"className":1341},[88],[21,1343,1345],{"className":1344},[60],[21,1346,1348],{"className":1347,"style":267},[64],[21,1349],{},[663,1351,1352,1353],{},"scalar in ",[21,1354,1356],{"className":1355},[24],[21,1357,1359],{"className":1358,"ariaHidden":29},[28],[21,1360,1362,1366,1370,1374,1379,1382,1385],{"className":1361},[33],[21,1363],{"className":1364,"style":1365},[37],"height:1em;vertical-align:-0.25em;",[21,1367,1369],{"className":1368},[221],"(",[21,1371,1373],{"className":1372},[42],"0",[21,1375,1378],{"className":1376},[1377],"mpunct",",",[21,1380],{"className":1381,"style":402},[401],[21,1383,116],{"className":1384},[42],[21,1386,1388],{"className":1387},[273],")",[663,1390,1391,1392],{},"softmax-normalized score, ",[21,1393,1395],{"className":1394},[24],[21,1396,1398,1496],{"className":1397,"ariaHidden":29},[28],[21,1399,1401,1404,1444,1447,1487,1490,1493],{"className":1400},[33],[21,1402],{"className":1403,"style":349},[37],[21,1405,1407,1410],{"className":1406},[353],[21,1408,360],{"className":1409,"style":359},[353,357,358],[21,1411,1413],{"className":1412},[51],[21,1414,1416,1436],{"className":1415},[55,56],[21,1417,1419,1433],{"className":1418},[60],[21,1420,1422],{"className":1421,"style":373},[64],[21,1423,1424,1427],{"style":376},[21,1425],{"className":1426,"style":73},[72],[21,1428,1430],{"className":1429},[77,78,79,80],[21,1431,257],{"className":1432,"style":256},[42,46,80],[21,1434,89],{"className":1435},[88],[21,1437,1439],{"className":1438},[60],[21,1440,1442],{"className":1441,"style":395},[64],[21,1443],{},[21,1445],{"className":1446,"style":402},[401],[21,1448,1450,1453],{"className":1449},[42],[21,1451,410],{"className":1452,"style":409},[42,46],[21,1454,1456],{"className":1455},[51],[21,1457,1459,1479],{"className":1458},[55,56],[21,1460,1462,1476],{"className":1461},[60],[21,1463,1465],{"className":1464,"style":243},[64],[21,1466,1467,1470],{"style":425},[21,1468],{"className":1469,"style":73},[72],[21,1471,1473],{"className":1472},[77,78,79,80],[21,1474,257],{"className":1475,"style":256},[42,46,80],[21,1477,89],{"className":1478},[88],[21,1480,1482],{"className":1481},[60],[21,1483,1485],{"className":1484,"style":267},[64],[21,1486],{},[21,1488],{"className":1489,"style":686},[401],[21,1491,691],{"className":1492},[690],[21,1494],{"className":1495,"style":686},[401],[21,1497,1499,1502],{"className":1498},[33],[21,1500],{"className":1501,"style":112},[37],[21,1503,116],{"className":1504},[42],[11,1506,1507,1508,1578],{},"Compatibility is measured by a dot product: ",[21,1509,1511],{"className":1510},[24],[21,1512,1514,1532],{"className":1513,"ariaHidden":29},[28],[21,1515,1517,1520,1523,1526,1529],{"className":1516},[33],[21,1518],{"className":1519,"style":1208},[37],[21,1521,203],{"className":1522,"style":202},[42,46],[21,1524],{"className":1525,"style":1087},[401],[21,1527,1219],{"className":1528},[1218],[21,1530],{"className":1531,"style":1087},[401],[21,1533,1535,1538],{"className":1534},[33],[21,1536],{"className":1537,"style":569},[37],[21,1539,1541,1544],{"className":1540},[42],[21,1542,230],{"className":1543,"style":229},[42,46],[21,1545,1547],{"className":1546},[51],[21,1548,1550,1570],{"className":1549},[55,56],[21,1551,1553,1567],{"className":1552},[60],[21,1554,1556],{"className":1555,"style":243},[64],[21,1557,1558,1561],{"style":246},[21,1559],{"className":1560,"style":73},[72],[21,1562,1564],{"className":1563},[77,78,79,80],[21,1565,257],{"className":1566,"style":256},[42,46,80],[21,1568,89],{"className":1569},[88],[21,1571,1573],{"className":1572},[60],[21,1574,1576],{"className":1575,"style":267},[64],[21,1577],{}," is large when the query\nand key point the same way. Normalizing the scores with a softmax turns them into\na probability distribution over positions, and the output is the expectation of\nthe values under that distribution,",[21,1580,1583],{"className":1581},[1582],"katex-display",[21,1584,1586],{"className":1585},[24],[21,1587,1589,1644,1979],{"className":1588,"ariaHidden":29},[28],[21,1590,1592,1595,1635,1638,1641],{"className":1591},[33],[21,1593],{"className":1594,"style":499},[37],[21,1596,1598,1601],{"className":1597},[42],[21,1599,410],{"className":1600,"style":409},[42,46],[21,1602,1604],{"className":1603},[51],[21,1605,1607,1627],{"className":1606},[55,56],[21,1608,1610,1624],{"className":1609},[60],[21,1611,1613],{"className":1612,"style":243},[64],[21,1614,1615,1618],{"style":425},[21,1616],{"className":1617,"style":73},[72],[21,1619,1621],{"className":1620},[77,78,79,80],[21,1622,257],{"className":1623,"style":256},[42,46,80],[21,1625,89],{"className":1626},[88],[21,1628,1630],{"className":1629},[60],[21,1631,1633],{"className":1632,"style":267},[64],[21,1634],{},[21,1636],{"className":1637,"style":686},[401],[21,1639,691],{"className":1640},[690],[21,1642],{"className":1643,"style":686},[401],[21,1645,1647,1651,1952,1955,1959,1962,1970,1973,1976],{"className":1646},[33],[21,1648],{"className":1649,"style":1650},[37],"height:2.5488em;vertical-align:-1.1218em;",[21,1652,1654,1658,1949],{"className":1653},[42],[21,1655],{"className":1656},[221,1657],"nulldelimiter",[21,1659,1662],{"className":1660},[1661],"mfrac",[21,1663,1665,1940],{"className":1664},[55,56],[21,1666,1668,1937],{"className":1667},[60],[21,1669,1672,1856,1867],{"className":1670,"style":1671},[64],"height:1.427em;",[21,1673,1675,1679],{"style":1674},"top:-2.314em;",[21,1676],{"className":1677,"style":1678},[72],"height:3em;",[21,1680,1682,1758,1761,1765,1768,1771,1774,1777,1780,1853],{"className":1681},[42],[21,1683,1685,1688],{"className":1684},[353],[21,1686,360],{"className":1687,"style":359},[353,357,358],[21,1689,1691],{"className":1690},[51],[21,1692,1694,1750],{"className":1693},[55,56],[21,1695,1697,1747],{"className":1696},[60],[21,1698,1701],{"className":1699,"style":1700},[64],"height:0.1783em;",[21,1702,1703,1706],{"style":376},[21,1704],{"className":1705,"style":73},[72],[21,1707,1709],{"className":1708},[77,78,79,80],[21,1710,1712],{"className":1711},[42,80],[21,1713,1715,1718],{"className":1714},[42,80],[21,1716,257],{"className":1717,"style":256},[42,46,80],[21,1719,1721],{"className":1720},[51],[21,1722,1724],{"className":1723},[55],[21,1725,1727],{"className":1726},[60],[21,1728,1731],{"className":1729,"style":1730},[64],"height:0.6828em;",[21,1732,1734,1737],{"style":1733},"top:-2.786em;margin-right:0.0714em;",[21,1735],{"className":1736,"style":824},[72],[21,1738,1740],{"className":1739},[77,828,829,80],[21,1741,1743],{"className":1742},[42,80],[21,1744,1746],{"className":1745},[42,80],"′",[21,1748,89],{"className":1749},[88],[21,1751,1753],{"className":1752},[60],[21,1754,1756],{"className":1755,"style":395},[64],[21,1757],{},[21,1759],{"className":1760,"style":402},[401],[21,1762,1764],{"className":1763},[353],"exp",[21,1766,1369],{"className":1767},[221],[21,1769,203],{"className":1770,"style":202},[42,46],[21,1772],{"className":1773,"style":1087},[401],[21,1775,1219],{"className":1776},[1218],[21,1778],{"className":1779,"style":1087},[401],[21,1781,1783,1786],{"className":1782},[42],[21,1784,230],{"className":1785,"style":229},[42,46],[21,1787,1789],{"className":1788},[51],[21,1790,1792,1845],{"className":1791},[55,56],[21,1793,1795,1842],{"className":1794},[60],[21,1796,1799],{"className":1797,"style":1798},[64],"height:0.328em;",[21,1800,1801,1804],{"style":246},[21,1802],{"className":1803,"style":73},[72],[21,1805,1807],{"className":1806},[77,78,79,80],[21,1808,1810],{"className":1809},[42,80],[21,1811,1813,1816],{"className":1812},[42,80],[21,1814,257],{"className":1815,"style":256},[42,46,80],[21,1817,1819],{"className":1818},[51],[21,1820,1822],{"className":1821},[55],[21,1823,1825],{"className":1824},[60],[21,1826,1828],{"className":1827,"style":1730},[64],[21,1829,1830,1833],{"style":1733},[21,1831],{"className":1832,"style":824},[72],[21,1834,1836],{"className":1835},[77,828,829,80],[21,1837,1839],{"className":1838},[42,80],[21,1840,1746],{"className":1841},[42,80],[21,1843,89],{"className":1844},[88],[21,1846,1848],{"className":1847},[60],[21,1849,1851],{"className":1850,"style":267},[64],[21,1852],{},[21,1854,1388],{"className":1855},[273],[21,1857,1859,1862],{"style":1858},"top:-3.23em;",[21,1860],{"className":1861,"style":1678},[72],[21,1863],{"className":1864,"style":1866},[1865],"frac-line","border-bottom-width:0.04em;",[21,1868,1870,1873],{"style":1869},"top:-3.677em;",[21,1871],{"className":1872,"style":1678},[72],[21,1874,1876,1879,1882,1885,1888,1891,1894,1934],{"className":1875},[42],[21,1877,1764],{"className":1878},[353],[21,1880,1369],{"className":1881},[221],[21,1883,203],{"className":1884,"style":202},[42,46],[21,1886],{"className":1887,"style":1087},[401],[21,1889,1219],{"className":1890},[1218],[21,1892],{"className":1893,"style":1087},[401],[21,1895,1897,1900],{"className":1896},[42],[21,1898,230],{"className":1899,"style":229},[42,46],[21,1901,1903],{"className":1902},[51],[21,1904,1906,1926],{"className":1905},[55,56],[21,1907,1909,1923],{"className":1908},[60],[21,1910,1912],{"className":1911,"style":243},[64],[21,1913,1914,1917],{"style":246},[21,1915],{"className":1916,"style":73},[72],[21,1918,1920],{"className":1919},[77,78,79,80],[21,1921,257],{"className":1922,"style":256},[42,46,80],[21,1924,89],{"className":1925},[88],[21,1927,1929],{"className":1928},[60],[21,1930,1932],{"className":1931,"style":267},[64],[21,1933],{},[21,1935,1388],{"className":1936},[273],[21,1938,89],{"className":1939},[88],[21,1941,1943],{"className":1942},[60],[21,1944,1947],{"className":1945,"style":1946},[64],"height:1.1218em;",[21,1948],{},[21,1950],{"className":1951},[273,1657],[21,1953,1378],{"className":1954},[1377],[21,1956],{"className":1957,"style":1958},[401],"margin-right:2em;",[21,1960],{"className":1961,"style":402},[401],[21,1963,1966],{"className":1964},[42,1965],"text",[21,1967,1969],{"className":1968},[42],"out",[21,1971],{"className":1972,"style":686},[401],[21,1974,691],{"className":1975},[690],[21,1977],{"className":1978,"style":686},[401],[21,1980,1982,1986,2037,2040,2080,2083,2123],{"className":1981},[33],[21,1983],{"className":1984,"style":1985},[37],"height:2.4638em;vertical-align:-1.4138em;",[21,1987,1990],{"className":1988},[353,1989],"op-limits",[21,1991,1993,2028],{"className":1992},[55,56],[21,1994,1996,2025],{"className":1995},[60],[21,1997,2000,2013],{"className":1998,"style":1999},[64],"height:1.05em;",[21,2001,2003,2007],{"style":2002},"top:-1.8723em;margin-left:0em;",[21,2004],{"className":2005,"style":2006},[72],"height:3.05em;",[21,2008,2010],{"className":2009},[77,78,79,80],[21,2011,257],{"className":2012,"style":256},[42,46,80],[21,2014,2016,2019],{"style":2015},"top:-3.05em;",[21,2017],{"className":2018,"style":2006},[72],[21,2020,2021],{},[21,2022,360],{"className":2023},[353,357,2024],"large-op",[21,2026,89],{"className":2027},[88],[21,2029,2031],{"className":2030},[60],[21,2032,2035],{"className":2033,"style":2034},[64],"height:1.4138em;",[21,2036],{},[21,2038],{"className":2039,"style":402},[401],[21,2041,2043,2046],{"className":2042},[42],[21,2044,410],{"className":2045,"style":409},[42,46],[21,2047,2049],{"className":2048},[51],[21,2050,2052,2072],{"className":2051},[55,56],[21,2053,2055,2069],{"className":2054},[60],[21,2056,2058],{"className":2057,"style":243},[64],[21,2059,2060,2063],{"style":425},[21,2061],{"className":2062,"style":73},[72],[21,2064,2066],{"className":2065},[77,78,79,80],[21,2067,257],{"className":2068,"style":256},[42,46,80],[21,2070,89],{"className":2071},[88],[21,2073,2075],{"className":2074},[60],[21,2076,2078],{"className":2077,"style":267},[64],[21,2079],{},[21,2081],{"className":2082,"style":402},[401],[21,2084,2086,2089],{"className":2085},[42],[21,2087,297],{"className":2088,"style":202},[42,46],[21,2090,2092],{"className":2091},[51],[21,2093,2095,2115],{"className":2094},[55,56],[21,2096,2098,2112],{"className":2097},[60],[21,2099,2101],{"className":2100,"style":243},[64],[21,2102,2103,2106],{"style":312},[21,2104],{"className":2105,"style":73},[72],[21,2107,2109],{"className":2108},[77,78,79,80],[21,2110,257],{"className":2111,"style":256},[42,46,80],[21,2113,89],{"className":2114},[88],[21,2116,2118],{"className":2117},[60],[21,2119,2121],{"className":2120,"style":267},[64],[21,2122],{},[21,2124,2126],{"className":2125},[42],".",[2128,2129],"tikz-figure",{"hash":2130},"275eb1406d1142114f692c18df30c935f3f30b3502e653aee85f0b98064c0a44",[616,2132,2134],{"id":2133},"scaled-dot-product-attention","Scaled dot-product attention",[11,2136,2137,2138,2153,2154,2170,2171,2187,2188,2204,2205,2220,2221,2236,2237,2252,2253,2268,2269,2302,2303,2318,2319,2393,2394,2447,2448,2501],{},"Stack the ",[21,2139,2141],{"className":2140},[24],[21,2142,2144],{"className":2143,"ariaHidden":29},[28],[21,2145,2147,2150],{"className":2146},[33],[21,2148],{"className":2149,"style":130},[37],[21,2151,134],{"className":2152},[42,46]," queries into the rows of ",[21,2155,2157],{"className":2156},[24],[21,2158,2160],{"className":2159,"ariaHidden":29},[28],[21,2161,2163,2167],{"className":2162},[33],[21,2164],{"className":2165,"style":2166},[37],"height:0.8778em;vertical-align:-0.1944em;",[21,2168,739],{"className":2169},[42,46],", the ",[21,2172,2174],{"className":2173},[24],[21,2175,2177],{"className":2176,"ariaHidden":29},[28],[21,2178,2180,2183],{"className":2179},[33],[21,2181],{"className":2182,"style":130},[37],[21,2184,2186],{"className":2185},[42,46],"m"," keys into the rows of ",[21,2189,2191],{"className":2190},[24],[21,2192,2194],{"className":2193,"ariaHidden":29},[28],[21,2195,2197,2201],{"className":2196},[33],[21,2198],{"className":2199,"style":2200},[37],"height:0.6833em;",[21,2202,920],{"className":2203,"style":919},[42,46],",\nand the ",[21,2206,2208],{"className":2207},[24],[21,2209,2211],{"className":2210,"ariaHidden":29},[28],[21,2212,2214,2217],{"className":2213},[33],[21,2215],{"className":2216,"style":130},[37],[21,2218,2186],{"className":2219},[42,46]," values into the rows of ",[21,2222,2224],{"className":2223},[24],[21,2225,2227],{"className":2226,"ariaHidden":29},[28],[21,2228,2230,2233],{"className":2229},[33],[21,2231],{"className":2232,"style":2200},[37],[21,2234,1088],{"className":2235,"style":1087},[42,46],". Keeping the query count ",[21,2238,2240],{"className":2239},[24],[21,2241,2243],{"className":2242,"ariaHidden":29},[28],[21,2244,2246,2249],{"className":2245},[33],[21,2247],{"className":2248,"style":130},[37],[21,2250,134],{"className":2251},[42,46]," and the\nkey\u002Fvalue count ",[21,2254,2256],{"className":2255},[24],[21,2257,2259],{"className":2258,"ariaHidden":29},[28],[21,2260,2262,2265],{"className":2261},[33],[21,2263],{"className":2264,"style":130},[37],[21,2266,2186],{"className":2267},[42,46]," separate matters: in self-attention ",[21,2270,2272],{"className":2271},[24],[21,2273,2275,2293],{"className":2274,"ariaHidden":29},[28],[21,2276,2278,2281,2284,2287,2290],{"className":2277},[33],[21,2279],{"className":2280,"style":130},[37],[21,2282,134],{"className":2283},[42,46],[21,2285],{"className":2286,"style":686},[401],[21,2288,691],{"className":2289},[690],[21,2291],{"className":2292,"style":686},[401],[21,2294,2296,2299],{"className":2295},[33],[21,2297],{"className":2298,"style":130},[37],[21,2300,2186],{"className":2301},[42,46],", but in\ncross-attention a decoder of length ",[21,2304,2306],{"className":2305},[24],[21,2307,2309],{"className":2308,"ariaHidden":29},[28],[21,2310,2312,2315],{"className":2311},[33],[21,2313],{"className":2314,"style":130},[37],[21,2316,134],{"className":2317},[42,46]," reads an encoder of length ",[21,2320,2322],{"className":2321},[24],[21,2323,2325,2384],{"className":2324,"ariaHidden":29},[28],[21,2326,2328,2332,2335,2338,2381],{"className":2327},[33],[21,2329],{"className":2330,"style":2331},[37],"height:0.8889em;vertical-align:-0.1944em;",[21,2333,2186],{"className":2334},[42,46],[21,2336],{"className":2337,"style":686},[401],[21,2339,2341,2374,2378],{"className":2340},[690],[21,2342,2344],{"className":2343},[690],[21,2345,2348],{"className":2346},[42,2347],"vbox",[21,2349,2352],{"className":2350},[2351],"thinbox",[21,2353,2356,2359,2370],{"className":2354},[2355],"rlap",[21,2357],{"className":2358,"style":2331},[37],[21,2360,2363],{"className":2361},[2362],"inner",[21,2364,2366],{"className":2365},[42],[21,2367,2369],{"className":2368},[690],"",[21,2371],{"className":2372},[2373],"fix",[21,2375],{"className":2376},[401,2377],"nobreak",[21,2379,691],{"className":2380},[690],[21,2382],{"className":2383,"style":686},[401],[21,2385,2387,2390],{"className":2386},[33],[21,2388],{"className":2389,"style":130},[37],[21,2391,134],{"className":2392},[42,46],".\nQueries and keys must share a width ",[21,2395,2397],{"className":2396},[24],[21,2398,2400],{"className":2399,"ariaHidden":29},[28],[21,2401,2403,2406],{"className":2402},[33],[21,2404],{"className":2405,"style":38},[37],[21,2407,2409,2412],{"className":2408},[42],[21,2410,804],{"className":2411},[42,46],[21,2413,2415],{"className":2414},[51],[21,2416,2418,2439],{"className":2417},[55,56],[21,2419,2421,2436],{"className":2420},[60],[21,2422,2425],{"className":2423,"style":2424},[64],"height:0.3361em;",[21,2426,2427,2430],{"style":68},[21,2428],{"className":2429,"style":73},[72],[21,2431,2433],{"className":2432},[77,78,79,80],[21,2434,230],{"className":2435,"style":229},[42,46,80],[21,2437,89],{"className":2438},[88],[21,2440,2442],{"className":2441},[60],[21,2443,2445],{"className":2444,"style":96},[64],[21,2446],{}," so their dot product is defined; values\ncarry an independent width ",[21,2449,2451],{"className":2450},[24],[21,2452,2454],{"className":2453,"ariaHidden":29},[28],[21,2455,2457,2460],{"className":2456},[33],[21,2458],{"className":2459,"style":38},[37],[21,2461,2463,2466],{"className":2462},[42],[21,2464,804],{"className":2465},[42,46],[21,2467,2469],{"className":2468},[51],[21,2470,2472,2493],{"className":2471},[55,56],[21,2473,2475,2490],{"className":2474},[60],[21,2476,2479],{"className":2477,"style":2478},[64],"height:0.1514em;",[21,2480,2481,2484],{"style":68},[21,2482],{"className":2483,"style":73},[72],[21,2485,2487],{"className":2486},[77,78,79,80],[21,2488,297],{"className":2489,"style":202},[42,46,80],[21,2491,89],{"className":2492},[88],[21,2494,2496],{"className":2495},[60],[21,2497,2499],{"className":2498,"style":96},[64],[21,2500],{},". The shapes chain as follows.",[21,2503,2505],{"className":2504},[1582],[21,2506,2508],{"className":2507},[24],[21,2509,2511],{"className":2510,"ariaHidden":29},[28],[21,2512,2514,2518,2694,2697,2700,2853,2856,2859,3009,3012,3015,3161,3164,3167,3511,3514,3517,3670,3673],{"className":2513},[33],[21,2515],{"className":2516,"style":2517},[37],"height:3.1147em;vertical-align:-1.9647em;",[21,2519,2523],{"className":2520},[2521,2522],"minner","munder",[21,2524,2526,2685],{"className":2525},[55,56],[21,2527,2529,2682],{"className":2528},[60],[21,2530,2532,2591],{"className":2531,"style":2200},[64],[21,2533,2535,2538],{"style":2534},"top:-1.4715em;",[21,2536],{"className":2537,"style":1678},[72],[21,2539,2541],{"className":2540},[77,78,79,80],[21,2542,2544,2547,2551],{"className":2543},[42,80],[21,2545,134],{"className":2546},[42,46,80],[21,2548,2550],{"className":2549},[1218,80],"×",[21,2552,2554,2557],{"className":2553},[42,80],[21,2555,804],{"className":2556},[42,46,80],[21,2558,2560],{"className":2559},[51],[21,2561,2563,2583],{"className":2562},[55,56],[21,2564,2566,2580],{"className":2565},[60],[21,2567,2569],{"className":2568,"style":817},[64],[21,2570,2571,2574],{"style":820},[21,2572],{"className":2573,"style":824},[72],[21,2575,2577],{"className":2576},[77,828,829,80],[21,2578,230],{"className":2579,"style":229},[42,46,80],[21,2581,89],{"className":2582},[88],[21,2584,2586],{"className":2585},[60],[21,2587,2589],{"className":2588,"style":842},[64],[21,2590],{},[21,2592,2594,2597],{"style":2593},"top:-3em;",[21,2595],{"className":2596,"style":1678},[72],[21,2598,2600],{"className":2599},[2521,2522],[21,2601,2603,2673],{"className":2602},[55,56],[21,2604,2606,2670],{"className":2605},[60],[21,2607,2609,2659],{"className":2608,"style":2200},[64],[21,2610,2614,2617],{"className":2611,"style":2613},[2612],"svg-align","top:-2.1576em;",[21,2615],{"className":2616,"style":1678},[72],[21,2618,2622,2639,2649],{"className":2619,"style":2621},[2620],"stretchy","height:0.548em;min-width:1.6em;",[21,2623,2627],{"className":2624,"style":2626},[2625],"brace-left","height:0.548em;",[2628,2629,2635],"svg",{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},"http:\u002F\u002Fwww.w3.org\u002F2000\u002Fsvg","400em","0.548em","0 0 400000 548","xMinYMin slice",[2636,2637],"path",{"d":2638},"M0 6l6-6h17c12.688 0 19.313.3 20 1 4 4 7.313 8.3 10 13\n 35.313 51.3 80.813 93.8 136.5 127.5 55.688 33.7 117.188 55.8 184.5 66.5.688\n 0 2 .3 4 1 18.688 2.7 76 4.3 172 5h399450v120H429l-6-1c-124.688-8-235-61.7\n-331-161C60.687 138.7 32.312 99.3 7 54L0 41V6z",[21,2640,2643],{"className":2641,"style":2626},[2642],"brace-center",[2628,2644,2646],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},"xMidYMin slice",[2636,2647],{"d":2648},"M199572 214\nc100.7 8.3 195.3 44 280 108 55.3 42 101.7 93 139 153l9 14c2.7-4 5.7-8.7 9-14\n 53.3-86.7 123.7-153 211-199 66.7-36 137.3-56.3 212-62h199568v120H200432c-178.3\n 11.7-311.7 78.3-403 201-6 8-9.7 12-11 12-.7.7-6.7 1-18 1s-17.3-.3-18-1c-1.3 0\n-5-4-11-12-44.7-59.3-101.3-106.3-170-141s-145.3-54.3-229-60H0V214z",[21,2650,2653],{"className":2651,"style":2626},[2652],"brace-right",[2628,2654,2656],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},"xMaxYMin slice",[2636,2657],{"d":2658},"M399994 0l6 6v35l-6 11c-56 104-135.3 181.3-238 232-57.3\n 28.7-117 45-179 50H-300V214h399897c43.3-7 81-15 113-26 100.7-33 179.7-91 237\n-174 2.7-5 6-9 10-13 .7-1 7.3-1 20-1h17z",[21,2660,2661,2664],{"style":2593},[21,2662],{"className":2663,"style":1678},[72],[21,2665,2667],{"className":2666},[42],[21,2668,739],{"className":2669},[42,46],[21,2671,89],{"className":2672},[88],[21,2674,2676],{"className":2675},[60],[21,2677,2680],{"className":2678,"style":2679},[64],"height:0.8424em;",[21,2681],{},[21,2683,89],{"className":2684},[88],[21,2686,2688],{"className":2687},[60],[21,2689,2692],{"className":2690,"style":2691},[64],"height:1.6344em;",[21,2693],{},[21,2695],{"className":2696,"style":1958},[401],[21,2698],{"className":2699,"style":402},[401],[21,2701,2703],{"className":2702},[2521,2522],[21,2704,2706,2844],{"className":2705},[55,56],[21,2707,2709,2841],{"className":2708},[60],[21,2710,2712,2770],{"className":2711,"style":2200},[64],[21,2713,2715,2718],{"style":2714},"top:-1.6659em;",[21,2716],{"className":2717,"style":1678},[72],[21,2719,2721],{"className":2720},[77,78,79,80],[21,2722,2724,2727,2730],{"className":2723},[42,80],[21,2725,2186],{"className":2726},[42,46,80],[21,2728,2550],{"className":2729},[1218,80],[21,2731,2733,2736],{"className":2732},[42,80],[21,2734,804],{"className":2735},[42,46,80],[21,2737,2739],{"className":2738},[51],[21,2740,2742,2762],{"className":2741},[55,56],[21,2743,2745,2759],{"className":2744},[60],[21,2746,2748],{"className":2747,"style":817},[64],[21,2749,2750,2753],{"style":820},[21,2751],{"className":2752,"style":824},[72],[21,2754,2756],{"className":2755},[77,828,829,80],[21,2757,230],{"className":2758,"style":229},[42,46,80],[21,2760,89],{"className":2761},[88],[21,2763,2765],{"className":2764},[60],[21,2766,2768],{"className":2767,"style":842},[64],[21,2769],{},[21,2771,2772,2775],{"style":2593},[21,2773],{"className":2774,"style":1678},[72],[21,2776,2778],{"className":2777},[2521,2522],[21,2779,2781,2832],{"className":2780},[55,56],[21,2782,2784,2829],{"className":2783},[60],[21,2785,2787,2818],{"className":2786,"style":2200},[64],[21,2788,2791,2794],{"className":2789,"style":2790},[2612],"top:-2.352em;",[21,2792],{"className":2793,"style":1678},[72],[21,2795,2797,2804,2811],{"className":2796,"style":2621},[2620],[21,2798,2800],{"className":2799,"style":2626},[2625],[2628,2801,2802],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},[2636,2803],{"d":2638},[21,2805,2807],{"className":2806,"style":2626},[2642],[2628,2808,2809],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},[2636,2810],{"d":2648},[21,2812,2814],{"className":2813,"style":2626},[2652],[2628,2815,2816],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},[2636,2817],{"d":2658},[21,2819,2820,2823],{"style":2593},[21,2821],{"className":2822,"style":1678},[72],[21,2824,2826],{"className":2825},[42],[21,2827,920],{"className":2828,"style":919},[42,46],[21,2830,89],{"className":2831},[88],[21,2833,2835],{"className":2834},[60],[21,2836,2839],{"className":2837,"style":2838},[64],"height:0.648em;",[21,2840],{},[21,2842,89],{"className":2843},[88],[21,2845,2847],{"className":2846},[60],[21,2848,2851],{"className":2849,"style":2850},[64],"height:1.44em;",[21,2852],{},[21,2854],{"className":2855,"style":1958},[401],[21,2857],{"className":2858,"style":402},[401],[21,2860,2862],{"className":2861},[2521,2522],[21,2863,2865,3000],{"className":2864},[55,56],[21,2866,2868,2997],{"className":2867},[60],[21,2869,2871,2928],{"className":2870,"style":2200},[64],[21,2872,2873,2876],{"style":2714},[21,2874],{"className":2875,"style":1678},[72],[21,2877,2879],{"className":2878},[77,78,79,80],[21,2880,2882,2885,2888],{"className":2881},[42,80],[21,2883,2186],{"className":2884},[42,46,80],[21,2886,2550],{"className":2887},[1218,80],[21,2889,2891,2894],{"className":2890},[42,80],[21,2892,804],{"className":2893},[42,46,80],[21,2895,2897],{"className":2896},[51],[21,2898,2900,2920],{"className":2899},[55,56],[21,2901,2903,2917],{"className":2902},[60],[21,2904,2906],{"className":2905,"style":1161},[64],[21,2907,2908,2911],{"style":1164},[21,2909],{"className":2910,"style":824},[72],[21,2912,2914],{"className":2913},[77,828,829,80],[21,2915,297],{"className":2916,"style":202},[42,46,80],[21,2918,89],{"className":2919},[88],[21,2921,2923],{"className":2922},[60],[21,2924,2926],{"className":2925,"style":1183},[64],[21,2927],{},[21,2929,2930,2933],{"style":2593},[21,2931],{"className":2932,"style":1678},[72],[21,2934,2936],{"className":2935},[2521,2522],[21,2937,2939,2989],{"className":2938},[55,56],[21,2940,2942,2986],{"className":2941},[60],[21,2943,2945,2975],{"className":2944,"style":2200},[64],[21,2946,2948,2951],{"className":2947,"style":2790},[2612],[21,2949],{"className":2950,"style":1678},[72],[21,2952,2954,2961,2968],{"className":2953,"style":2621},[2620],[21,2955,2957],{"className":2956,"style":2626},[2625],[2628,2958,2959],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},[2636,2960],{"d":2638},[21,2962,2964],{"className":2963,"style":2626},[2642],[2628,2965,2966],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},[2636,2967],{"d":2648},[21,2969,2971],{"className":2970,"style":2626},[2652],[2628,2972,2973],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},[2636,2974],{"d":2658},[21,2976,2977,2980],{"style":2593},[21,2978],{"className":2979,"style":1678},[72],[21,2981,2983],{"className":2982},[42],[21,2984,1088],{"className":2985,"style":1087},[42,46],[21,2987,89],{"className":2988},[88],[21,2990,2992],{"className":2991},[60],[21,2993,2995],{"className":2994,"style":2838},[64],[21,2996],{},[21,2998,89],{"className":2999},[88],[21,3001,3003],{"className":3002},[60],[21,3004,3007],{"className":3005,"style":3006},[64],"height:1.4342em;",[21,3008],{},[21,3010],{"className":3011,"style":1958},[401],[21,3013],{"className":3014,"style":402},[401],[21,3016,3018],{"className":3017},[2521,2522],[21,3019,3021,3152],{"className":3020},[55,56],[21,3022,3024,3149],{"className":3023},[60],[21,3025,3028,3049],{"className":3026,"style":3027},[64],"height:0.8991em;",[21,3029,3031,3034],{"style":3030},"top:-1.5492em;",[21,3032],{"className":3033,"style":1678},[72],[21,3035,3037],{"className":3036},[77,78,79,80],[21,3038,3040,3043,3046],{"className":3039},[42,80],[21,3041,134],{"className":3042},[42,46,80],[21,3044,2550],{"className":3045},[1218,80],[21,3047,2186],{"className":3048},[42,46,80],[21,3050,3051,3054],{"style":2593},[21,3052],{"className":3053,"style":1678},[72],[21,3055,3057],{"className":3056},[2521,2522],[21,3058,3060,3141],{"className":3059},[55,56],[21,3061,3063,3138],{"className":3062},[60],[21,3064,3066,3096],{"className":3065,"style":3027},[64],[21,3067,3069,3072],{"className":3068,"style":2613},[2612],[21,3070],{"className":3071,"style":1678},[72],[21,3073,3075,3082,3089],{"className":3074,"style":2621},[2620],[21,3076,3078],{"className":3077,"style":2626},[2625],[2628,3079,3080],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},[2636,3081],{"d":2638},[21,3083,3085],{"className":3084,"style":2626},[2642],[2628,3086,3087],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},[2636,3088],{"d":2648},[21,3090,3092],{"className":3091,"style":2626},[2652],[2628,3093,3094],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},[2636,3095],{"d":2658},[21,3097,3098,3101],{"style":2593},[21,3099],{"className":3100,"style":1678},[72],[21,3102,3104,3107],{"className":3103},[42],[21,3105,739],{"className":3106},[42,46],[21,3108,3110,3113],{"className":3109},[42],[21,3111,920],{"className":3112,"style":919},[42,46],[21,3114,3116],{"className":3115},[51],[21,3117,3119],{"className":3118},[55],[21,3120,3122],{"className":3121},[60],[21,3123,3125],{"className":3124,"style":3027},[64],[21,3126,3128,3131],{"style":3127},"top:-3.113em;margin-right:0.05em;",[21,3129],{"className":3130,"style":73},[72],[21,3132,3134],{"className":3133},[77,78,79,80],[21,3135,3137],{"className":3136},[42,80],"⊤",[21,3139,89],{"className":3140},[88],[21,3142,3144],{"className":3143},[60],[21,3145,3147],{"className":3146,"style":2679},[64],[21,3148],{},[21,3150,89],{"className":3151},[88],[21,3153,3155],{"className":3154},[60],[21,3156,3159],{"className":3157,"style":3158},[64],"height:1.5091em;",[21,3160],{},[21,3162],{"className":3163,"style":1958},[401],[21,3165],{"className":3166,"style":402},[401],[21,3168,3170],{"className":3169},[2521,2522],[21,3171,3173,3502],{"className":3172},[55,56],[21,3174,3176,3499],{"className":3175},[60],[21,3177,3180,3202],{"className":3178,"style":3179},[64],"height:1.15em;",[21,3181,3183,3187],{"style":3182},"top:-1.2436em;",[21,3184],{"className":3185,"style":3186},[72],"height:3.15em;",[21,3188,3190],{"className":3189},[77,78,79,80],[21,3191,3193,3196,3199],{"className":3192},[42,80],[21,3194,134],{"className":3195},[42,46,80],[21,3197,2550],{"className":3198},[1218,80],[21,3200,2186],{"className":3201},[42,46,80],[21,3203,3205,3208],{"style":3204},"top:-3.15em;",[21,3206],{"className":3207,"style":3186},[72],[21,3209,3211],{"className":3210},[2521,2522],[21,3212,3214,3490],{"className":3213},[55,56],[21,3215,3217,3487],{"className":3216},[60],[21,3218,3220,3251],{"className":3219,"style":3179},[64],[21,3221,3224,3227],{"className":3222,"style":3223},[2612],"top:-1.852em;",[21,3225],{"className":3226,"style":3186},[72],[21,3228,3230,3237,3244],{"className":3229,"style":2621},[2620],[21,3231,3233],{"className":3232,"style":2626},[2625],[2628,3234,3235],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},[2636,3236],{"d":2638},[21,3238,3240],{"className":3239,"style":2626},[2642],[2628,3241,3242],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},[2636,3243],{"d":2648},[21,3245,3247],{"className":3246,"style":2626},[2652],[2628,3248,3249],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},[2636,3250],{"d":2658},[21,3252,3253,3256],{"style":3204},[21,3254],{"className":3255,"style":3186},[72],[21,3257,3259,3267,3271,3274],{"className":3258},[42],[21,3260,3262],{"className":3261},[353],[21,3263,3266],{"className":3264},[42,3265],"mathrm","softmax",[21,3268],{"className":3269,"style":3270},[401],"margin-right:-0.1667em;",[21,3272],{"className":3273,"style":402},[401],[21,3275,3277,3287,3481],{"className":3276},[2521],[21,3278,3282],{"className":3279,"style":3281},[221,3280],"delimcenter","top:0em;",[21,3283,1369],{"className":3284},[3285,3286],"delimsizing","size2",[21,3288,3290,3293,3478],{"className":3289},[42],[21,3291],{"className":3292},[221,1657],[21,3294,3296],{"className":3295},[1661],[21,3297,3299,3469],{"className":3298},[55,56],[21,3300,3302,3466],{"className":3301},[60],[21,3303,3306,3412,3420],{"className":3304,"style":3305},[64],"height:1.095em;",[21,3307,3309,3312],{"style":3308},"top:-2.5864em;",[21,3310],{"className":3311,"style":1678},[72],[21,3313,3315],{"className":3314},[77,78,79,80],[21,3316,3318],{"className":3317},[42,80],[21,3319,3322],{"className":3320},[42,3321,80],"sqrt",[21,3323,3325,3403],{"className":3324},[55,56],[21,3326,3328,3400],{"className":3327},[60],[21,3329,3332,3382],{"className":3330,"style":3331},[64],"height:0.8622em;",[21,3333,3335,3338],{"className":3334,"style":2593},[2612],[21,3336],{"className":3337,"style":1678},[72],[21,3339,3342],{"className":3340,"style":3341},[42,80],"padding-left:0.833em;",[21,3343,3345,3348],{"className":3344},[42,80],[21,3346,804],{"className":3347},[42,46,80],[21,3349,3351],{"className":3350},[51],[21,3352,3354,3374],{"className":3353},[55,56],[21,3355,3357,3371],{"className":3356},[60],[21,3358,3360],{"className":3359,"style":817},[64],[21,3361,3362,3365],{"style":820},[21,3363],{"className":3364,"style":824},[72],[21,3366,3368],{"className":3367},[77,828,829,80],[21,3369,230],{"className":3370,"style":229},[42,46,80],[21,3372,89],{"className":3373},[88],[21,3375,3377],{"className":3376},[60],[21,3378,3380],{"className":3379,"style":842},[64],[21,3381],{},[21,3383,3385,3388],{"style":3384},"top:-2.8222em;",[21,3386],{"className":3387,"style":1678},[72],[21,3389,3393],{"className":3390,"style":3392},[3391,80],"hide-tail","min-width:0.853em;height:1.08em;",[2628,3394,3397],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},"1.08em","0 0 400000 1080",[2636,3398],{"d":3399},"M95,702\nc-2.7,0,-7.17,-2.7,-13.5,-8c-5.8,-5.3,-9.5,-10,-9.5,-14\nc0,-2,0.3,-3.3,1,-4c1.3,-2.7,23.83,-20.7,67.5,-54\nc44.2,-33.3,65.8,-50.3,66.5,-51c1.3,-1.3,3,-2,5,-2c4.7,0,8.7,3.3,12,10\ns173,378,173,378c0.7,0,35.3,-71,104,-213c68.7,-142,137.5,-285,206.5,-429\nc69,-144,104.5,-217.7,106.5,-221\nl0 -0\nc5.3,-9.3,12,-14,20,-14\nH400000v40H845.2724\ns-225.272,467,-225.272,467s-235,486,-235,486c-2.7,4.7,-9,7,-19,7\nc-6,0,-10,-1,-12,-3s-194,-422,-194,-422s-65,47,-65,47z\nM834 80h400000v40h-400000z",[21,3401,89],{"className":3402},[88],[21,3404,3406],{"className":3405},[60],[21,3407,3410],{"className":3408,"style":3409},[64],"height:0.1778em;",[21,3411],{},[21,3413,3414,3417],{"style":1858},[21,3415],{"className":3416,"style":1678},[72],[21,3418],{"className":3419,"style":1866},[1865],[21,3421,3423,3426],{"style":3422},"top:-3.4461em;",[21,3424],{"className":3425,"style":1678},[72],[21,3427,3429],{"className":3428},[77,78,79,80],[21,3430,3432,3435],{"className":3431},[42,80],[21,3433,739],{"className":3434},[42,46,80],[21,3436,3438,3441],{"className":3437},[42,80],[21,3439,920],{"className":3440,"style":919},[42,46,80],[21,3442,3444],{"className":3443},[51],[21,3445,3447],{"className":3446},[55],[21,3448,3450],{"className":3449},[60],[21,3451,3454],{"className":3452,"style":3453},[64],"height:0.927em;",[21,3455,3457,3460],{"style":3456},"top:-2.931em;margin-right:0.0714em;",[21,3458],{"className":3459,"style":824},[72],[21,3461,3463],{"className":3462},[77,828,829,80],[21,3464,3137],{"className":3465},[42,80],[21,3467,89],{"className":3468},[88],[21,3470,3472],{"className":3471},[60],[21,3473,3476],{"className":3474,"style":3475},[64],"height:0.538em;",[21,3477],{},[21,3479],{"className":3480},[273,1657],[21,3482,3484],{"className":3483,"style":3281},[273,3280],[21,3485,1388],{"className":3486},[3285,3286],[21,3488,89],{"className":3489},[88],[21,3491,3493],{"className":3492},[60],[21,3494,3497],{"className":3495,"style":3496},[64],"height:1.298em;",[21,3498],{},[21,3500,89],{"className":3501},[88],[21,3503,3505],{"className":3504},[60],[21,3506,3509],{"className":3507,"style":3508},[64],"height:1.9647em;",[21,3510],{},[21,3512],{"className":3513,"style":1958},[401],[21,3515],{"className":3516,"style":402},[401],[21,3518,3520],{"className":3519},[2521,2522],[21,3521,3523,3662],{"className":3522},[55,56],[21,3524,3526,3659],{"className":3525},[60],[21,3527,3529,3586],{"className":3528,"style":2200},[64],[21,3530,3531,3534],{"style":2714},[21,3532],{"className":3533,"style":1678},[72],[21,3535,3537],{"className":3536},[77,78,79,80],[21,3538,3540,3543,3546],{"className":3539},[42,80],[21,3541,134],{"className":3542},[42,46,80],[21,3544,2550],{"className":3545},[1218,80],[21,3547,3549,3552],{"className":3548},[42,80],[21,3550,804],{"className":3551},[42,46,80],[21,3553,3555],{"className":3554},[51],[21,3556,3558,3578],{"className":3557},[55,56],[21,3559,3561,3575],{"className":3560},[60],[21,3562,3564],{"className":3563,"style":1161},[64],[21,3565,3566,3569],{"style":1164},[21,3567],{"className":3568,"style":824},[72],[21,3570,3572],{"className":3571},[77,828,829,80],[21,3573,297],{"className":3574,"style":202},[42,46,80],[21,3576,89],{"className":3577},[88],[21,3579,3581],{"className":3580},[60],[21,3582,3584],{"className":3583,"style":1183},[64],[21,3585],{},[21,3587,3588,3591],{"style":2593},[21,3589],{"className":3590,"style":1678},[72],[21,3592,3594],{"className":3593},[2521,2522],[21,3595,3597,3651],{"className":3596},[55,56],[21,3598,3600,3648],{"className":3599},[60],[21,3601,3603,3633],{"className":3602,"style":2200},[64],[21,3604,3606,3609],{"className":3605,"style":2790},[2612],[21,3607],{"className":3608,"style":1678},[72],[21,3610,3612,3619,3626],{"className":3611,"style":2621},[2620],[21,3613,3615],{"className":3614,"style":2626},[2625],[2628,3616,3617],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},[2636,3618],{"d":2638},[21,3620,3622],{"className":3621,"style":2626},[2642],[2628,3623,3624],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},[2636,3625],{"d":2648},[21,3627,3629],{"className":3628,"style":2626},[2652],[2628,3630,3631],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},[2636,3632],{"d":2658},[21,3634,3635,3638],{"style":2593},[21,3636],{"className":3637,"style":1678},[72],[21,3639,3641,3645],{"className":3640},[42],[21,3642,3644],{"className":3643},[42,46],"A",[21,3646,1088],{"className":3647,"style":1087},[42,46],[21,3649,89],{"className":3650},[88],[21,3652,3654],{"className":3653},[60],[21,3655,3657],{"className":3656,"style":2838},[64],[21,3658],{},[21,3660,89],{"className":3661},[88],[21,3663,3665],{"className":3664},[60],[21,3666,3668],{"className":3667,"style":3006},[64],[21,3669],{},[21,3671],{"className":3672,"style":402},[401],[21,3674,2126],{"className":3675},[42],[11,3677,3678,3679,3779,3780,3811,3812,3919,3920,3936,3937,3952,3953,4053,4054,4069,4070,4085,4086,4101,4102,4117,4118,4136,4137,4171,4172,4242,4243,4313,4314,4366],{},"Every query-key score is a single matrix product ",[21,3680,3682],{"className":3681},[24],[21,3683,3685,3734],{"className":3684,"ariaHidden":29},[28],[21,3686,3688,3692,3695,3724,3727,3731],{"className":3687},[33],[21,3689],{"className":3690,"style":3691},[37],"height:1.0435em;vertical-align:-0.1944em;",[21,3693,739],{"className":3694},[42,46],[21,3696,3698,3701],{"className":3697},[42],[21,3699,920],{"className":3700,"style":919},[42,46],[21,3702,3704],{"className":3703},[51],[21,3705,3707],{"className":3706},[55],[21,3708,3710],{"className":3709},[60],[21,3711,3713],{"className":3712,"style":765},[64],[21,3714,3715,3718],{"style":788},[21,3716],{"className":3717,"style":73},[72],[21,3719,3721],{"className":3720},[77,78,79,80],[21,3722,3137],{"className":3723},[42,80],[21,3725],{"className":3726,"style":686},[401],[21,3728,3730],{"className":3729},[690],"∈",[21,3732],{"className":3733,"style":686},[401],[21,3735,3737,3741],{"className":3736},[33],[21,3738],{"className":3739,"style":3740},[37],"height:0.7713em;",[21,3742,3744,3747],{"className":3743},[42],[21,3745,773],{"className":3746},[42,772],[21,3748,3750],{"className":3749},[51],[21,3751,3753],{"className":3752},[55],[21,3754,3756],{"className":3755},[60],[21,3757,3759],{"className":3758,"style":3740},[64],[21,3760,3761,3764],{"style":788},[21,3762],{"className":3763,"style":73},[72],[21,3765,3767],{"className":3766},[77,78,79,80],[21,3768,3770,3773,3776],{"className":3769},[42,80],[21,3771,134],{"className":3772},[42,46,80],[21,3774,2550],{"className":3775},[1218,80],[21,3777,2186],{"className":3778},[42,46,80],": entry ",[21,3781,3783],{"className":3782},[24],[21,3784,3786],{"className":3785,"ariaHidden":29},[28],[21,3787,3789,3792,3795,3799,3802,3805,3808],{"className":3788},[33],[21,3790],{"className":3791,"style":1365},[37],[21,3793,1369],{"className":3794},[221],[21,3796,3798],{"className":3797},[42,46],"i",[21,3800,1378],{"className":3801},[1377],[21,3803],{"className":3804,"style":402},[401],[21,3806,257],{"className":3807,"style":256},[42,46],[21,3809,1388],{"className":3810},[273]," is the dot product ",[21,3813,3815],{"className":3814},[24],[21,3816,3818,3873],{"className":3817,"ariaHidden":29},[28],[21,3819,3821,3824,3864,3867,3870],{"className":3820},[33],[21,3822],{"className":3823,"style":1208},[37],[21,3825,3827,3830],{"className":3826},[42],[21,3828,203],{"className":3829,"style":202},[42,46],[21,3831,3833],{"className":3832},[51],[21,3834,3836,3856],{"className":3835},[55,56],[21,3837,3839,3853],{"className":3838},[60],[21,3840,3842],{"className":3841,"style":243},[64],[21,3843,3844,3847],{"style":312},[21,3845],{"className":3846,"style":73},[72],[21,3848,3850],{"className":3849},[77,78,79,80],[21,3851,3798],{"className":3852},[42,46,80],[21,3854,89],{"className":3855},[88],[21,3857,3859],{"className":3858},[60],[21,3860,3862],{"className":3861,"style":96},[64],[21,3863],{},[21,3865],{"className":3866,"style":1087},[401],[21,3868,1219],{"className":3869},[1218],[21,3871],{"className":3872,"style":1087},[401],[21,3874,3876,3879],{"className":3875},[33],[21,3877],{"className":3878,"style":569},[37],[21,3880,3882,3885],{"className":3881},[42],[21,3883,230],{"className":3884,"style":229},[42,46],[21,3886,3888],{"className":3887},[51],[21,3889,3891,3911],{"className":3890},[55,56],[21,3892,3894,3908],{"className":3893},[60],[21,3895,3897],{"className":3896,"style":243},[64],[21,3898,3899,3902],{"style":246},[21,3900],{"className":3901,"style":73},[72],[21,3903,3905],{"className":3904},[77,78,79,80],[21,3906,257],{"className":3907,"style":256},[42,46,80],[21,3909,89],{"className":3910},[88],[21,3912,3914],{"className":3913},[60],[21,3915,3917],{"className":3916,"style":267},[64],[21,3918],{}," of query ",[21,3921,3923],{"className":3922},[24],[21,3924,3926],{"className":3925,"ariaHidden":29},[28],[21,3927,3929,3933],{"className":3928},[33],[21,3930],{"className":3931,"style":3932},[37],"height:0.6595em;",[21,3934,3798],{"className":3935},[42,46]," with key\n",[21,3938,3940],{"className":3939},[24],[21,3941,3943],{"className":3942,"ariaHidden":29},[28],[21,3944,3946,3949],{"className":3945},[33],[21,3947],{"className":3948,"style":1287},[37],[21,3950,257],{"className":3951,"style":256},[42,46],". Divide by ",[21,3954,3956],{"className":3955},[24],[21,3957,3959],{"className":3958,"ariaHidden":29},[28],[21,3960,3962,3966],{"className":3961},[33],[21,3963],{"className":3964,"style":3965},[37],"height:1.04em;vertical-align:-0.1828em;",[21,3967,3969],{"className":3968},[42,3321],[21,3970,3972,4044],{"className":3971},[55,56],[21,3973,3975,4041],{"className":3974},[60],[21,3976,3979,4028],{"className":3977,"style":3978},[64],"height:0.8572em;",[21,3980,3982,3985],{"className":3981,"style":2593},[2612],[21,3983],{"className":3984,"style":1678},[72],[21,3986,3988],{"className":3987,"style":3341},[42],[21,3989,3991,3994],{"className":3990},[42],[21,3992,804],{"className":3993},[42,46],[21,3995,3997],{"className":3996},[51],[21,3998,4000,4020],{"className":3999},[55,56],[21,4001,4003,4017],{"className":4002},[60],[21,4004,4006],{"className":4005,"style":2424},[64],[21,4007,4008,4011],{"style":68},[21,4009],{"className":4010,"style":73},[72],[21,4012,4014],{"className":4013},[77,78,79,80],[21,4015,230],{"className":4016,"style":229},[42,46,80],[21,4018,89],{"className":4019},[88],[21,4021,4023],{"className":4022},[60],[21,4024,4026],{"className":4025,"style":96},[64],[21,4027],{},[21,4029,4031,4034],{"style":4030},"top:-2.8172em;",[21,4032],{"className":4033,"style":1678},[72],[21,4035,4037],{"className":4036,"style":3392},[3391],[2628,4038,4039],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,4040],{"d":3399},[21,4042,89],{"className":4043},[88],[21,4045,4047],{"className":4046},[60],[21,4048,4051],{"className":4049,"style":4050},[64],"height:0.1828em;",[21,4052],{},", run a softmax along each row so that row ",[21,4055,4057],{"className":4056},[24],[21,4058,4060],{"className":4059,"ariaHidden":29},[28],[21,4061,4063,4066],{"className":4062},[33],[21,4064],{"className":4065,"style":3932},[37],[21,4067,3798],{"className":4068},[42,46]," becomes\na probability distribution over the ",[21,4071,4073],{"className":4072},[24],[21,4074,4076],{"className":4075,"ariaHidden":29},[28],[21,4077,4079,4082],{"className":4078},[33],[21,4080],{"className":4081,"style":130},[37],[21,4083,2186],{"className":4084},[42,46]," keys, and right-multiply by ",[21,4087,4089],{"className":4088},[24],[21,4090,4092],{"className":4091,"ariaHidden":29},[28],[21,4093,4095,4098],{"className":4094},[33],[21,4096],{"className":4097,"style":2200},[37],[21,4099,1088],{"className":4100,"style":1087},[42,46],". The\ninner dimension ",[21,4103,4105],{"className":4104},[24],[21,4106,4108],{"className":4107,"ariaHidden":29},[28],[21,4109,4111,4114],{"className":4110},[33],[21,4112],{"className":4113,"style":130},[37],[21,4115,2186],{"className":4116},[42,46]," cancels in ",[21,4119,4121],{"className":4120},[24],[21,4122,4124],{"className":4123,"ariaHidden":29},[28],[21,4125,4127,4130,4133],{"className":4126},[33],[21,4128],{"className":4129,"style":2200},[37],[21,4131,3644],{"className":4132},[42,46],[21,4134,1088],{"className":4135,"style":1087},[42,46],", so an ",[21,4138,4140],{"className":4139},[24],[21,4141,4143,4162],{"className":4142,"ariaHidden":29},[28],[21,4144,4146,4150,4153,4156,4159],{"className":4145},[33],[21,4147],{"className":4148,"style":4149},[37],"height:0.6667em;vertical-align:-0.0833em;",[21,4151,134],{"className":4152},[42,46],[21,4154],{"className":4155,"style":1087},[401],[21,4157,2550],{"className":4158},[1218],[21,4160],{"className":4161,"style":1087},[401],[21,4163,4165,4168],{"className":4164},[33],[21,4166],{"className":4167,"style":130},[37],[21,4169,2186],{"className":4170},[42,46]," weight matrix times an ",[21,4173,4175],{"className":4174},[24],[21,4176,4178,4196],{"className":4177,"ariaHidden":29},[28],[21,4179,4181,4184,4187,4190,4193],{"className":4180},[33],[21,4182],{"className":4183,"style":4149},[37],[21,4185,2186],{"className":4186},[42,46],[21,4188],{"className":4189,"style":1087},[401],[21,4191,2550],{"className":4192},[1218],[21,4194],{"className":4195,"style":1087},[401],[21,4197,4199,4202],{"className":4198},[33],[21,4200],{"className":4201,"style":38},[37],[21,4203,4205,4208],{"className":4204},[42],[21,4206,804],{"className":4207},[42,46],[21,4209,4211],{"className":4210},[51],[21,4212,4214,4234],{"className":4213},[55,56],[21,4215,4217,4231],{"className":4216},[60],[21,4218,4220],{"className":4219,"style":2478},[64],[21,4221,4222,4225],{"style":68},[21,4223],{"className":4224,"style":73},[72],[21,4226,4228],{"className":4227},[77,78,79,80],[21,4229,297],{"className":4230,"style":202},[42,46,80],[21,4232,89],{"className":4233},[88],[21,4235,4237],{"className":4236},[60],[21,4238,4240],{"className":4239,"style":96},[64],[21,4241],{}," value matrix returns an ",[21,4244,4246],{"className":4245},[24],[21,4247,4249,4267],{"className":4248,"ariaHidden":29},[28],[21,4250,4252,4255,4258,4261,4264],{"className":4251},[33],[21,4253],{"className":4254,"style":4149},[37],[21,4256,134],{"className":4257},[42,46],[21,4259],{"className":4260,"style":1087},[401],[21,4262,2550],{"className":4263},[1218],[21,4265],{"className":4266,"style":1087},[401],[21,4268,4270,4273],{"className":4269},[33],[21,4271],{"className":4272,"style":38},[37],[21,4274,4276,4279],{"className":4275},[42],[21,4277,804],{"className":4278},[42,46],[21,4280,4282],{"className":4281},[51],[21,4283,4285,4305],{"className":4284},[55,56],[21,4286,4288,4302],{"className":4287},[60],[21,4289,4291],{"className":4290,"style":2478},[64],[21,4292,4293,4296],{"style":68},[21,4294],{"className":4295,"style":73},[72],[21,4297,4299],{"className":4298},[77,78,79,80],[21,4300,297],{"className":4301,"style":202},[42,46,80],[21,4303,89],{"className":4304},[88],[21,4306,4308],{"className":4307},[60],[21,4309,4311],{"className":4310,"style":96},[64],[21,4312],{}," output — one ",[21,4315,4317],{"className":4316},[24],[21,4318,4320],{"className":4319,"ariaHidden":29},[28],[21,4321,4323,4326],{"className":4322},[33],[21,4324],{"className":4325,"style":38},[37],[21,4327,4329,4332],{"className":4328},[42],[21,4330,804],{"className":4331},[42,46],[21,4333,4335],{"className":4334},[51],[21,4336,4338,4358],{"className":4337},[55,56],[21,4339,4341,4355],{"className":4340},[60],[21,4342,4344],{"className":4343,"style":2478},[64],[21,4345,4346,4349],{"style":68},[21,4347],{"className":4348,"style":73},[72],[21,4350,4352],{"className":4351},[77,78,79,80],[21,4353,297],{"className":4354,"style":202},[42,46,80],[21,4356,89],{"className":4357},[88],[21,4359,4361],{"className":4360},[60],[21,4362,4364],{"className":4363,"style":96},[64],[21,4365],{},"-vector per\nquery, exactly the input query count. The whole operation collapses to one\nexpression.",[177,4368,4369,4693,5065],{"type":179},[11,4370,4371,4374,4375,4480,4481,4587,4588,1378],{},[158,4372,4373],{},"Definition (Scaled dot-product attention)."," For queries ",[21,4376,4378],{"className":4377},[24],[21,4379,4381,4399],{"className":4380,"ariaHidden":29},[28],[21,4382,4384,4387,4390,4393,4396],{"className":4383},[33],[21,4385],{"className":4386,"style":2166},[37],[21,4388,739],{"className":4389},[42,46],[21,4391],{"className":4392,"style":686},[401],[21,4394,3730],{"className":4395},[690],[21,4397],{"className":4398,"style":686},[401],[21,4400,4402,4405],{"className":4401},[33],[21,4403],{"className":4404,"style":765},[37],[21,4406,4408,4411],{"className":4407},[42],[21,4409,773],{"className":4410},[42,772],[21,4412,4414],{"className":4413},[51],[21,4415,4417],{"className":4416},[55],[21,4418,4420],{"className":4419},[60],[21,4421,4423],{"className":4422,"style":765},[64],[21,4424,4425,4428],{"style":788},[21,4426],{"className":4427,"style":73},[72],[21,4429,4431],{"className":4430},[77,78,79,80],[21,4432,4434,4437,4440],{"className":4433},[42,80],[21,4435,134],{"className":4436},[42,46,80],[21,4438,2550],{"className":4439},[1218,80],[21,4441,4443,4446],{"className":4442},[42,80],[21,4444,804],{"className":4445},[42,46,80],[21,4447,4449],{"className":4448},[51],[21,4450,4452,4472],{"className":4451},[55,56],[21,4453,4455,4469],{"className":4454},[60],[21,4456,4458],{"className":4457,"style":817},[64],[21,4459,4460,4463],{"style":820},[21,4461],{"className":4462,"style":824},[72],[21,4464,4466],{"className":4465},[77,828,829,80],[21,4467,230],{"className":4468,"style":229},[42,46,80],[21,4470,89],{"className":4471},[88],[21,4473,4475],{"className":4474},[60],[21,4476,4478],{"className":4477,"style":842},[64],[21,4479],{},", keys ",[21,4482,4484],{"className":4483},[24],[21,4485,4487,4506],{"className":4486,"ariaHidden":29},[28],[21,4488,4490,4494,4497,4500,4503],{"className":4489},[33],[21,4491],{"className":4492,"style":4493},[37],"height:0.7224em;vertical-align:-0.0391em;",[21,4495,920],{"className":4496,"style":919},[42,46],[21,4498],{"className":4499,"style":686},[401],[21,4501,3730],{"className":4502},[690],[21,4504],{"className":4505,"style":686},[401],[21,4507,4509,4512],{"className":4508},[33],[21,4510],{"className":4511,"style":765},[37],[21,4513,4515,4518],{"className":4514},[42],[21,4516,773],{"className":4517},[42,772],[21,4519,4521],{"className":4520},[51],[21,4522,4524],{"className":4523},[55],[21,4525,4527],{"className":4526},[60],[21,4528,4530],{"className":4529,"style":765},[64],[21,4531,4532,4535],{"style":788},[21,4533],{"className":4534,"style":73},[72],[21,4536,4538],{"className":4537},[77,78,79,80],[21,4539,4541,4544,4547],{"className":4540},[42,80],[21,4542,2186],{"className":4543},[42,46,80],[21,4545,2550],{"className":4546},[1218,80],[21,4548,4550,4553],{"className":4549},[42,80],[21,4551,804],{"className":4552},[42,46,80],[21,4554,4556],{"className":4555},[51],[21,4557,4559,4579],{"className":4558},[55,56],[21,4560,4562,4576],{"className":4561},[60],[21,4563,4565],{"className":4564,"style":817},[64],[21,4566,4567,4570],{"style":820},[21,4568],{"className":4569,"style":824},[72],[21,4571,4573],{"className":4572},[77,828,829,80],[21,4574,230],{"className":4575,"style":229},[42,46,80],[21,4577,89],{"className":4578},[88],[21,4580,4582],{"className":4581},[60],[21,4583,4585],{"className":4584,"style":842},[64],[21,4586],{},", and values ",[21,4589,4591],{"className":4590},[24],[21,4592,4594,4612],{"className":4593,"ariaHidden":29},[28],[21,4595,4597,4600,4603,4606,4609],{"className":4596},[33],[21,4598],{"className":4599,"style":4493},[37],[21,4601,1088],{"className":4602,"style":1087},[42,46],[21,4604],{"className":4605,"style":686},[401],[21,4607,3730],{"className":4608},[690],[21,4610],{"className":4611,"style":686},[401],[21,4613,4615,4618],{"className":4614},[33],[21,4616],{"className":4617,"style":765},[37],[21,4619,4621,4624],{"className":4620},[42],[21,4622,773],{"className":4623},[42,772],[21,4625,4627],{"className":4626},[51],[21,4628,4630],{"className":4629},[55],[21,4631,4633],{"className":4632},[60],[21,4634,4636],{"className":4635,"style":765},[64],[21,4637,4638,4641],{"style":788},[21,4639],{"className":4640,"style":73},[72],[21,4642,4644],{"className":4643},[77,78,79,80],[21,4645,4647,4650,4653],{"className":4646},[42,80],[21,4648,2186],{"className":4649},[42,46,80],[21,4651,2550],{"className":4652},[1218,80],[21,4654,4656,4659],{"className":4655},[42,80],[21,4657,804],{"className":4658},[42,46,80],[21,4660,4662],{"className":4661},[51],[21,4663,4665,4685],{"className":4664},[55,56],[21,4666,4668,4682],{"className":4667},[60],[21,4669,4671],{"className":4670,"style":1161},[64],[21,4672,4673,4676],{"style":1164},[21,4674],{"className":4675,"style":824},[72],[21,4677,4679],{"className":4678},[77,828,829,80],[21,4680,297],{"className":4681,"style":202},[42,46,80],[21,4683,89],{"className":4684},[88],[21,4686,4688],{"className":4687},[60],[21,4689,4691],{"className":4690,"style":1183},[64],[21,4692],{},[21,4694,4696],{"className":4695},[1582],[21,4697,4699],{"className":4698},[24],[21,4700,4702,4751,4981],{"className":4701,"ariaHidden":29},[28],[21,4703,4705,4708,4715,4718,4721,4724,4727,4730,4733,4736,4739,4742,4745,4748],{"className":4704},[33],[21,4706],{"className":4707,"style":1365},[37],[21,4709,4711],{"className":4710},[353],[21,4712,4714],{"className":4713},[42,3265],"Attn",[21,4716,1369],{"className":4717},[221],[21,4719,739],{"className":4720},[42,46],[21,4722,1378],{"className":4723},[1377],[21,4725],{"className":4726,"style":402},[401],[21,4728,920],{"className":4729,"style":919},[42,46],[21,4731,1378],{"className":4732},[1377],[21,4734],{"className":4735,"style":402},[401],[21,4737,1088],{"className":4738,"style":1087},[42,46],[21,4740,1388],{"className":4741},[273],[21,4743],{"className":4744,"style":686},[401],[21,4746,691],{"className":4747},[690],[21,4749],{"className":4750,"style":686},[401],[21,4752,4754,4758,4764,4767,4770,4960,4963,4966,4969,4972,4975,4978],{"className":4753},[33],[21,4755],{"className":4756,"style":4757},[37],"height:2.4761em;vertical-align:-0.95em;",[21,4759,4761],{"className":4760},[353],[21,4762,3266],{"className":4763},[42,3265],[21,4765],{"className":4766,"style":3270},[401],[21,4768],{"className":4769,"style":402},[401],[21,4771,4773,4779,4954],{"className":4772},[2521],[21,4774,4776],{"className":4775,"style":3281},[221,3280],[21,4777,1369],{"className":4778},[3285,79],[21,4780,4782,4785,4951],{"className":4781},[42],[21,4783],{"className":4784},[221,1657],[21,4786,4788],{"className":4787},[1661],[21,4789,4791,4942],{"className":4790},[55,56],[21,4792,4794,4939],{"className":4793},[60],[21,4795,4798,4891,4899],{"className":4796,"style":4797},[64],"height:1.5261em;",[21,4799,4801,4804],{"style":4800},"top:-2.2528em;",[21,4802],{"className":4803,"style":1678},[72],[21,4805,4807],{"className":4806},[42],[21,4808,4810],{"className":4809},[42,3321],[21,4811,4813,4883],{"className":4812},[55,56],[21,4814,4816,4880],{"className":4815},[60],[21,4817,4819,4868],{"className":4818,"style":3978},[64],[21,4820,4822,4825],{"className":4821,"style":2593},[2612],[21,4823],{"className":4824,"style":1678},[72],[21,4826,4828],{"className":4827,"style":3341},[42],[21,4829,4831,4834],{"className":4830},[42],[21,4832,804],{"className":4833},[42,46],[21,4835,4837],{"className":4836},[51],[21,4838,4840,4860],{"className":4839},[55,56],[21,4841,4843,4857],{"className":4842},[60],[21,4844,4846],{"className":4845,"style":2424},[64],[21,4847,4848,4851],{"style":68},[21,4849],{"className":4850,"style":73},[72],[21,4852,4854],{"className":4853},[77,78,79,80],[21,4855,230],{"className":4856,"style":229},[42,46,80],[21,4858,89],{"className":4859},[88],[21,4861,4863],{"className":4862},[60],[21,4864,4866],{"className":4865,"style":96},[64],[21,4867],{},[21,4869,4870,4873],{"style":4030},[21,4871],{"className":4872,"style":1678},[72],[21,4874,4876],{"className":4875,"style":3392},[3391],[2628,4877,4878],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,4879],{"d":3399},[21,4881,89],{"className":4882},[88],[21,4884,4886],{"className":4885},[60],[21,4887,4889],{"className":4888,"style":4050},[64],[21,4890],{},[21,4892,4893,4896],{"style":1858},[21,4894],{"className":4895,"style":1678},[72],[21,4897],{"className":4898,"style":1866},[1865],[21,4900,4901,4904],{"style":1869},[21,4902],{"className":4903,"style":1678},[72],[21,4905,4907,4910],{"className":4906},[42],[21,4908,739],{"className":4909},[42,46],[21,4911,4913,4916],{"className":4912},[42],[21,4914,920],{"className":4915,"style":919},[42,46],[21,4917,4919],{"className":4918},[51],[21,4920,4922],{"className":4921},[55],[21,4923,4925],{"className":4924},[60],[21,4926,4928],{"className":4927,"style":765},[64],[21,4929,4930,4933],{"style":788},[21,4931],{"className":4932,"style":73},[72],[21,4934,4936],{"className":4935},[77,78,79,80],[21,4937,3137],{"className":4938},[42,80],[21,4940,89],{"className":4941},[88],[21,4943,4945],{"className":4944},[60],[21,4946,4949],{"className":4947,"style":4948},[64],"height:0.93em;",[21,4950],{},[21,4952],{"className":4953},[273,1657],[21,4955,4957],{"className":4956,"style":3281},[273,3280],[21,4958,1388],{"className":4959},[3285,79],[21,4961],{"className":4962,"style":402},[401],[21,4964,1088],{"className":4965,"style":1087},[42,46],[21,4967],{"className":4968,"style":686},[401],[21,4970],{"className":4971,"style":686},[401],[21,4973,3730],{"className":4974},[690],[21,4976],{"className":4977,"style":686},[401],[21,4979],{"className":4980,"style":686},[401],[21,4982,4984,4987,5062],{"className":4983},[33],[21,4985],{"className":4986,"style":3027},[37],[21,4988,4990,4993],{"className":4989},[42],[21,4991,773],{"className":4992},[42,772],[21,4994,4996],{"className":4995},[51],[21,4997,4999],{"className":4998},[55],[21,5000,5002],{"className":5001},[60],[21,5003,5005],{"className":5004,"style":3027},[64],[21,5006,5007,5010],{"style":3127},[21,5008],{"className":5009,"style":73},[72],[21,5011,5013],{"className":5012},[77,78,79,80],[21,5014,5016,5019,5022],{"className":5015},[42,80],[21,5017,134],{"className":5018},[42,46,80],[21,5020,2550],{"className":5021},[1218,80],[21,5023,5025,5028],{"className":5024},[42,80],[21,5026,804],{"className":5027},[42,46,80],[21,5029,5031],{"className":5030},[51],[21,5032,5034,5054],{"className":5033},[55,56],[21,5035,5037,5051],{"className":5036},[60],[21,5038,5040],{"className":5039,"style":1161},[64],[21,5041,5042,5045],{"style":1164},[21,5043],{"className":5044,"style":824},[72],[21,5046,5048],{"className":5047},[77,828,829,80],[21,5049,297],{"className":5050,"style":202},[42,46,80],[21,5052,89],{"className":5053},[88],[21,5055,5057],{"className":5056},[60],[21,5058,5060],{"className":5059,"style":1183},[64],[21,5061],{},[21,5063,2126],{"className":5064},[42],[11,5066,5067,5068,5083,5084,5099,5100,5115],{},"The softmax is applied independently to each row, so row ",[21,5069,5071],{"className":5070},[24],[21,5072,5074],{"className":5073,"ariaHidden":29},[28],[21,5075,5077,5080],{"className":5076},[33],[21,5078],{"className":5079,"style":3932},[37],[21,5081,3798],{"className":5082},[42,46]," of the output is a\nconvex combination of the ",[21,5085,5087],{"className":5086},[24],[21,5088,5090],{"className":5089,"ariaHidden":29},[28],[21,5091,5093,5096],{"className":5092},[33],[21,5094],{"className":5095,"style":130},[37],[21,5097,2186],{"className":5098},[42,46]," value vectors, weighted by how well query ",[21,5101,5103],{"className":5102},[24],[21,5104,5106],{"className":5105,"ariaHidden":29},[28],[21,5107,5109,5112],{"className":5108},[33],[21,5110],{"className":5111,"style":3932},[37],[21,5113,3798],{"className":5114},[42,46],"\nmatches every key.",[11,5117,5118,5119,5220,5221],{},"The only unexplained piece is the ",[21,5120,5122],{"className":5121},[24],[21,5123,5125],{"className":5124,"ariaHidden":29},[28],[21,5126,5128,5132,5136],{"className":5127},[33],[21,5129],{"className":5130,"style":5131},[37],"height:1.1072em;vertical-align:-0.25em;",[21,5133,5135],{"className":5134},[42],"1\u002F",[21,5137,5139],{"className":5138},[42,3321],[21,5140,5142,5212],{"className":5141},[55,56],[21,5143,5145,5209],{"className":5144},[60],[21,5146,5148,5197],{"className":5147,"style":3978},[64],[21,5149,5151,5154],{"className":5150,"style":2593},[2612],[21,5152],{"className":5153,"style":1678},[72],[21,5155,5157],{"className":5156,"style":3341},[42],[21,5158,5160,5163],{"className":5159},[42],[21,5161,804],{"className":5162},[42,46],[21,5164,5166],{"className":5165},[51],[21,5167,5169,5189],{"className":5168},[55,56],[21,5170,5172,5186],{"className":5171},[60],[21,5173,5175],{"className":5174,"style":2424},[64],[21,5176,5177,5180],{"style":68},[21,5178],{"className":5179,"style":73},[72],[21,5181,5183],{"className":5182},[77,78,79,80],[21,5184,230],{"className":5185,"style":229},[42,46,80],[21,5187,89],{"className":5188},[88],[21,5190,5192],{"className":5191},[60],[21,5193,5195],{"className":5194,"style":96},[64],[21,5196],{},[21,5198,5199,5202],{"style":4030},[21,5200],{"className":5201,"style":1678},[72],[21,5203,5205],{"className":5204,"style":3392},[3391],[2628,5206,5207],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,5208],{"d":3399},[21,5210,89],{"className":5211},[88],[21,5213,5215],{"className":5214},[60],[21,5216,5218],{"className":5217,"style":4050},[64],[21,5219],{}," factor. It is essential:\nwithout it the softmax saturates and gradients vanish.",[168,5222,5223],{},[15,5224,5228],{"href":5225,"ariaDescribedBy":5226,"dataFootnoteRef":6,"id":5227},"#user-content-fn-chollet-attn",[174],"user-content-fnref-chollet-attn","2",[5230,5231,5233,5234,5330],"h3",{"id":5232},"why-the-dk-scaling","Why the ",[21,5235,5237],{"className":5236},[24],[21,5238,5240],{"className":5239,"ariaHidden":29},[28],[21,5241,5243,5246],{"className":5242},[33],[21,5244],{"className":5245,"style":3965},[37],[21,5247,5249],{"className":5248},[42,3321],[21,5250,5252,5322],{"className":5251},[55,56],[21,5253,5255,5319],{"className":5254},[60],[21,5256,5258,5307],{"className":5257,"style":3978},[64],[21,5259,5261,5264],{"className":5260,"style":2593},[2612],[21,5262],{"className":5263,"style":1678},[72],[21,5265,5267],{"className":5266,"style":3341},[42],[21,5268,5270,5273],{"className":5269},[42],[21,5271,804],{"className":5272},[42,46],[21,5274,5276],{"className":5275},[51],[21,5277,5279,5299],{"className":5278},[55,56],[21,5280,5282,5296],{"className":5281},[60],[21,5283,5285],{"className":5284,"style":2424},[64],[21,5286,5287,5290],{"style":68},[21,5288],{"className":5289,"style":73},[72],[21,5291,5293],{"className":5292},[77,78,79,80],[21,5294,230],{"className":5295,"style":229},[42,46,80],[21,5297,89],{"className":5298},[88],[21,5300,5302],{"className":5301},[60],[21,5303,5305],{"className":5304,"style":96},[64],[21,5306],{},[21,5308,5309,5312],{"style":4030},[21,5310],{"className":5311,"style":1678},[72],[21,5313,5315],{"className":5314,"style":3392},[3391],[2628,5316,5317],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,5318],{"d":3399},[21,5320,89],{"className":5321},[88],[21,5323,5325],{"className":5324},[60],[21,5326,5328],{"className":5327,"style":4050},[64],[21,5329],{}," scaling",[11,5332,5333,5334,5349,5350,5365,5366,5381,5382,5397,5398,5450],{},"Treat the entries of a query ",[21,5335,5337],{"className":5336},[24],[21,5338,5340],{"className":5339,"ariaHidden":29},[28],[21,5341,5343,5346],{"className":5342},[33],[21,5344],{"className":5345,"style":198},[37],[21,5347,203],{"className":5348,"style":202},[42,46]," and a key ",[21,5351,5353],{"className":5352},[24],[21,5354,5356],{"className":5355,"ariaHidden":29},[28],[21,5357,5359,5362],{"className":5358},[33],[21,5360],{"className":5361,"style":867},[37],[21,5363,230],{"className":5364,"style":229},[42,46]," as independent random variables\nwith mean ",[21,5367,5369],{"className":5368},[24],[21,5370,5372],{"className":5371,"ariaHidden":29},[28],[21,5373,5375,5378],{"className":5374},[33],[21,5376],{"className":5377,"style":112},[37],[21,5379,1373],{"className":5380},[42]," and variance ",[21,5383,5385],{"className":5384},[24],[21,5386,5388],{"className":5387,"ariaHidden":29},[28],[21,5389,5391,5394],{"className":5390},[33],[21,5392],{"className":5393,"style":112},[37],[21,5395,116],{"className":5396},[42],". The dot product is a sum of ",[21,5399,5401],{"className":5400},[24],[21,5402,5404],{"className":5403,"ariaHidden":29},[28],[21,5405,5407,5410],{"className":5406},[33],[21,5408],{"className":5409,"style":38},[37],[21,5411,5413,5416],{"className":5412},[42],[21,5414,804],{"className":5415},[42,46],[21,5417,5419],{"className":5418},[51],[21,5420,5422,5442],{"className":5421},[55,56],[21,5423,5425,5439],{"className":5424},[60],[21,5426,5428],{"className":5427,"style":2424},[64],[21,5429,5430,5433],{"style":68},[21,5431],{"className":5432,"style":73},[72],[21,5434,5436],{"className":5435},[77,78,79,80],[21,5437,230],{"className":5438,"style":229},[42,46,80],[21,5440,89],{"className":5441},[88],[21,5443,5445],{"className":5444},[60],[21,5446,5448],{"className":5447,"style":96},[64],[21,5449],{}," such products,\nand we can compute the variance of that sum exactly.",[177,5452,5454],{"type":5453},"lemma",[11,5455,5456,5459,5460,5614,5615,5766,5767,5614,5913,6063,6064,6299,6300,5381,6315,2126],{},[158,5457,5458],{},"Lemma (Variance of the dot product)."," If ",[21,5461,5463],{"className":5462},[24],[21,5464,5466],{"className":5465,"ariaHidden":29},[28],[21,5467,5469,5473,5514,5517,5520,5524,5527,5530,5533],{"className":5468},[33],[21,5470],{"className":5471,"style":5472},[37],"height:0.6864em;vertical-align:-0.2559em;",[21,5474,5476,5479],{"className":5475},[42],[21,5477,203],{"className":5478,"style":202},[42,46],[21,5480,5482],{"className":5481},[51],[21,5483,5485,5506],{"className":5484},[55,56],[21,5486,5488,5503],{"className":5487},[60],[21,5489,5492],{"className":5490,"style":5491},[64],"height:0.3011em;",[21,5493,5494,5497],{"style":312},[21,5495],{"className":5496,"style":73},[72],[21,5498,5500],{"className":5499},[77,78,79,80],[21,5501,116],{"className":5502},[42,80],[21,5504,89],{"className":5505},[88],[21,5507,5509],{"className":5508},[60],[21,5510,5512],{"className":5511,"style":96},[64],[21,5513],{},[21,5515,1378],{"className":5516},[1377],[21,5518],{"className":5519,"style":402},[401],[21,5521,5523],{"className":5522},[2521],"…",[21,5525],{"className":5526,"style":402},[401],[21,5528,1378],{"className":5529},[1377],[21,5531],{"className":5532,"style":402},[401],[21,5534,5536,5539],{"className":5535},[42],[21,5537,203],{"className":5538,"style":202},[42,46],[21,5540,5542],{"className":5541},[51],[21,5543,5545,5605],{"className":5544},[55,56],[21,5546,5548,5602],{"className":5547},[60],[21,5549,5551],{"className":5550,"style":2424},[64],[21,5552,5553,5556],{"style":312},[21,5554],{"className":5555,"style":73},[72],[21,5557,5559],{"className":5558},[77,78,79,80],[21,5560,5562],{"className":5561},[42,80],[21,5563,5565,5568],{"className":5564},[42,80],[21,5566,804],{"className":5567},[42,46,80],[21,5569,5571],{"className":5570},[51],[21,5572,5574,5594],{"className":5573},[55,56],[21,5575,5577,5591],{"className":5576},[60],[21,5578,5580],{"className":5579,"style":817},[64],[21,5581,5582,5585],{"style":820},[21,5583],{"className":5584,"style":824},[72],[21,5586,5588],{"className":5587},[77,828,829,80],[21,5589,230],{"className":5590,"style":229},[42,46,80],[21,5592,89],{"className":5593},[88],[21,5595,5597],{"className":5596},[60],[21,5598,5600],{"className":5599,"style":842},[64],[21,5601],{},[21,5603,89],{"className":5604},[88],[21,5606,5608],{"className":5607},[60],[21,5609,5612],{"className":5610,"style":5611},[64],"height:0.2559em;",[21,5613],{}," and ",[21,5616,5618],{"className":5617},[24],[21,5619,5621],{"className":5620,"ariaHidden":29},[28],[21,5622,5624,5628,5668,5671,5674,5677,5680,5683,5686],{"className":5623},[33],[21,5625],{"className":5626,"style":5627},[37],"height:0.9503em;vertical-align:-0.2559em;",[21,5629,5631,5634],{"className":5630},[42],[21,5632,230],{"className":5633,"style":229},[42,46],[21,5635,5637],{"className":5636},[51],[21,5638,5640,5660],{"className":5639},[55,56],[21,5641,5643,5657],{"className":5642},[60],[21,5644,5646],{"className":5645,"style":5491},[64],[21,5647,5648,5651],{"style":246},[21,5649],{"className":5650,"style":73},[72],[21,5652,5654],{"className":5653},[77,78,79,80],[21,5655,116],{"className":5656},[42,80],[21,5658,89],{"className":5659},[88],[21,5661,5663],{"className":5662},[60],[21,5664,5666],{"className":5665,"style":96},[64],[21,5667],{},[21,5669,1378],{"className":5670},[1377],[21,5672],{"className":5673,"style":402},[401],[21,5675,5523],{"className":5676},[2521],[21,5678],{"className":5679,"style":402},[401],[21,5681,1378],{"className":5682},[1377],[21,5684],{"className":5685,"style":402},[401],[21,5687,5689,5692],{"className":5688},[42],[21,5690,230],{"className":5691,"style":229},[42,46],[21,5693,5695],{"className":5694},[51],[21,5696,5698,5758],{"className":5697},[55,56],[21,5699,5701,5755],{"className":5700},[60],[21,5702,5704],{"className":5703,"style":2424},[64],[21,5705,5706,5709],{"style":246},[21,5707],{"className":5708,"style":73},[72],[21,5710,5712],{"className":5711},[77,78,79,80],[21,5713,5715],{"className":5714},[42,80],[21,5716,5718,5721],{"className":5717},[42,80],[21,5719,804],{"className":5720},[42,46,80],[21,5722,5724],{"className":5723},[51],[21,5725,5727,5747],{"className":5726},[55,56],[21,5728,5730,5744],{"className":5729},[60],[21,5731,5733],{"className":5732,"style":817},[64],[21,5734,5735,5738],{"style":820},[21,5736],{"className":5737,"style":824},[72],[21,5739,5741],{"className":5740},[77,828,829,80],[21,5742,230],{"className":5743,"style":229},[42,46,80],[21,5745,89],{"className":5746},[88],[21,5748,5750],{"className":5749},[60],[21,5751,5753],{"className":5752,"style":842},[64],[21,5754],{},[21,5756,89],{"className":5757},[88],[21,5759,5761],{"className":5760},[60],[21,5762,5764],{"className":5763,"style":5611},[64],[21,5765],{}," are mutually independent with ",[21,5768,5770],{"className":5769},[24],[21,5771,5773,5840,5904],{"className":5772,"ariaHidden":29},[28],[21,5774,5776,5779,5783,5787,5827,5831,5834,5837],{"className":5775},[33],[21,5777],{"className":5778,"style":1365},[37],[21,5780,5782],{"className":5781},[42,772],"E",[21,5784,5786],{"className":5785},[221],"[",[21,5788,5790,5793],{"className":5789},[42],[21,5791,203],{"className":5792,"style":202},[42,46],[21,5794,5796],{"className":5795},[51],[21,5797,5799,5819],{"className":5798},[55,56],[21,5800,5802,5816],{"className":5801},[60],[21,5803,5805],{"className":5804,"style":243},[64],[21,5806,5807,5810],{"style":312},[21,5808],{"className":5809,"style":73},[72],[21,5811,5813],{"className":5812},[77,78,79,80],[21,5814,3798],{"className":5815},[42,46,80],[21,5817,89],{"className":5818},[88],[21,5820,5822],{"className":5821},[60],[21,5823,5825],{"className":5824,"style":96},[64],[21,5826],{},[21,5828,5830],{"className":5829},[273],"]",[21,5832],{"className":5833,"style":686},[401],[21,5835,691],{"className":5836},[690],[21,5838],{"className":5839,"style":686},[401],[21,5841,5843,5846,5849,5852,5892,5895,5898,5901],{"className":5842},[33],[21,5844],{"className":5845,"style":1365},[37],[21,5847,5782],{"className":5848},[42,772],[21,5850,5786],{"className":5851},[221],[21,5853,5855,5858],{"className":5854},[42],[21,5856,230],{"className":5857,"style":229},[42,46],[21,5859,5861],{"className":5860},[51],[21,5862,5864,5884],{"className":5863},[55,56],[21,5865,5867,5881],{"className":5866},[60],[21,5868,5870],{"className":5869,"style":243},[64],[21,5871,5872,5875],{"style":246},[21,5873],{"className":5874,"style":73},[72],[21,5876,5878],{"className":5877},[77,78,79,80],[21,5879,3798],{"className":5880},[42,46,80],[21,5882,89],{"className":5883},[88],[21,5885,5887],{"className":5886},[60],[21,5888,5890],{"className":5889,"style":96},[64],[21,5891],{},[21,5893,5830],{"className":5894},[273],[21,5896],{"className":5897,"style":686},[401],[21,5899,691],{"className":5900},[690],[21,5902],{"className":5903,"style":686},[401],[21,5905,5907,5910],{"className":5906},[33],[21,5908],{"className":5909,"style":112},[37],[21,5911,1373],{"className":5912},[42],[21,5914,5916],{"className":5915},[24],[21,5917,5919,5987,6054],{"className":5918,"ariaHidden":29},[28],[21,5920,5922,5925,5932,5935,5975,5978,5981,5984],{"className":5921},[33],[21,5923],{"className":5924,"style":1365},[37],[21,5926,5928],{"className":5927},[353],[21,5929,5931],{"className":5930},[42,3265],"Var",[21,5933,1369],{"className":5934},[221],[21,5936,5938,5941],{"className":5937},[42],[21,5939,203],{"className":5940,"style":202},[42,46],[21,5942,5944],{"className":5943},[51],[21,5945,5947,5967],{"className":5946},[55,56],[21,5948,5950,5964],{"className":5949},[60],[21,5951,5953],{"className":5952,"style":243},[64],[21,5954,5955,5958],{"style":312},[21,5956],{"className":5957,"style":73},[72],[21,5959,5961],{"className":5960},[77,78,79,80],[21,5962,3798],{"className":5963},[42,46,80],[21,5965,89],{"className":5966},[88],[21,5968,5970],{"className":5969},[60],[21,5971,5973],{"className":5972,"style":96},[64],[21,5974],{},[21,5976,1388],{"className":5977},[273],[21,5979],{"className":5980,"style":686},[401],[21,5982,691],{"className":5983},[690],[21,5985],{"className":5986,"style":686},[401],[21,5988,5990,5993,5999,6002,6042,6045,6048,6051],{"className":5989},[33],[21,5991],{"className":5992,"style":1365},[37],[21,5994,5996],{"className":5995},[353],[21,5997,5931],{"className":5998},[42,3265],[21,6000,1369],{"className":6001},[221],[21,6003,6005,6008],{"className":6004},[42],[21,6006,230],{"className":6007,"style":229},[42,46],[21,6009,6011],{"className":6010},[51],[21,6012,6014,6034],{"className":6013},[55,56],[21,6015,6017,6031],{"className":6016},[60],[21,6018,6020],{"className":6019,"style":243},[64],[21,6021,6022,6025],{"style":246},[21,6023],{"className":6024,"style":73},[72],[21,6026,6028],{"className":6027},[77,78,79,80],[21,6029,3798],{"className":6030},[42,46,80],[21,6032,89],{"className":6033},[88],[21,6035,6037],{"className":6036},[60],[21,6038,6040],{"className":6039,"style":96},[64],[21,6041],{},[21,6043,1388],{"className":6044},[273],[21,6046],{"className":6047,"style":686},[401],[21,6049,691],{"className":6050},[690],[21,6052],{"className":6053,"style":686},[401],[21,6055,6057,6060],{"className":6056},[33],[21,6058],{"className":6059,"style":112},[37],[21,6061,116],{"className":6062},[42],", then the dot\nproduct ",[21,6065,6067],{"className":6066},[24],[21,6068,6070,6088,6106],{"className":6069,"ariaHidden":29},[28],[21,6071,6073,6076,6079,6082,6085],{"className":6072},[33],[21,6074],{"className":6075,"style":1208},[37],[21,6077,203],{"className":6078,"style":202},[42,46],[21,6080],{"className":6081,"style":1087},[401],[21,6083,1219],{"className":6084},[1218],[21,6086],{"className":6087,"style":1087},[401],[21,6089,6091,6094,6097,6100,6103],{"className":6090},[33],[21,6092],{"className":6093,"style":867},[37],[21,6095,230],{"className":6096,"style":229},[42,46],[21,6098],{"className":6099,"style":686},[401],[21,6101,691],{"className":6102},[690],[21,6104],{"className":6105,"style":686},[401],[21,6107,6109,6113,6216,6219,6259],{"className":6108},[33],[21,6110],{"className":6111,"style":6112},[37],"height:1.2887em;vertical-align:-0.2997em;",[21,6114,6116,6119],{"className":6115},[353],[21,6117,360],{"className":6118,"style":359},[353,357,358],[21,6120,6122],{"className":6121},[51],[21,6123,6125,6207],{"className":6124},[55,56],[21,6126,6128,6204],{"className":6127},[60],[21,6129,6132,6152],{"className":6130,"style":6131},[64],"height:0.989em;",[21,6133,6134,6137],{"style":376},[21,6135],{"className":6136,"style":73},[72],[21,6138,6140],{"className":6139},[77,78,79,80],[21,6141,6143,6146,6149],{"className":6142},[42,80],[21,6144,3798],{"className":6145},[42,46,80],[21,6147,691],{"className":6148},[690,80],[21,6150,116],{"className":6151},[42,80],[21,6153,6155,6158],{"style":6154},"top:-3.2029em;margin-right:0.05em;",[21,6156],{"className":6157,"style":73},[72],[21,6159,6161],{"className":6160},[77,78,79,80],[21,6162,6164],{"className":6163},[42,80],[21,6165,6167,6170],{"className":6166},[42,80],[21,6168,804],{"className":6169},[42,46,80],[21,6171,6173],{"className":6172},[51],[21,6174,6176,6196],{"className":6175},[55,56],[21,6177,6179,6193],{"className":6178},[60],[21,6180,6182],{"className":6181,"style":817},[64],[21,6183,6184,6187],{"style":820},[21,6185],{"className":6186,"style":824},[72],[21,6188,6190],{"className":6189},[77,828,829,80],[21,6191,230],{"className":6192,"style":229},[42,46,80],[21,6194,89],{"className":6195},[88],[21,6197,6199],{"className":6198},[60],[21,6200,6202],{"className":6201,"style":842},[64],[21,6203],{},[21,6205,89],{"className":6206},[88],[21,6208,6210],{"className":6209},[60],[21,6211,6214],{"className":6212,"style":6213},[64],"height:0.2997em;",[21,6215],{},[21,6217],{"className":6218,"style":402},[401],[21,6220,6222,6225],{"className":6221},[42],[21,6223,203],{"className":6224,"style":202},[42,46],[21,6226,6228],{"className":6227},[51],[21,6229,6231,6251],{"className":6230},[55,56],[21,6232,6234,6248],{"className":6233},[60],[21,6235,6237],{"className":6236,"style":243},[64],[21,6238,6239,6242],{"style":312},[21,6240],{"className":6241,"style":73},[72],[21,6243,6245],{"className":6244},[77,78,79,80],[21,6246,3798],{"className":6247},[42,46,80],[21,6249,89],{"className":6250},[88],[21,6252,6254],{"className":6253},[60],[21,6255,6257],{"className":6256,"style":96},[64],[21,6258],{},[21,6260,6262,6265],{"className":6261},[42],[21,6263,230],{"className":6264,"style":229},[42,46],[21,6266,6268],{"className":6267},[51],[21,6269,6271,6291],{"className":6270},[55,56],[21,6272,6274,6288],{"className":6273},[60],[21,6275,6277],{"className":6276,"style":243},[64],[21,6278,6279,6282],{"style":246},[21,6280],{"className":6281,"style":73},[72],[21,6283,6285],{"className":6284},[77,78,79,80],[21,6286,3798],{"className":6287},[42,46,80],[21,6289,89],{"className":6290},[88],[21,6292,6294],{"className":6293},[60],[21,6295,6297],{"className":6296,"style":96},[64],[21,6298],{}," has mean ",[21,6301,6303],{"className":6302},[24],[21,6304,6306],{"className":6305,"ariaHidden":29},[28],[21,6307,6309,6312],{"className":6308},[33],[21,6310],{"className":6311,"style":112},[37],[21,6313,1373],{"className":6314},[42],[21,6316,6318],{"className":6317},[24],[21,6319,6321],{"className":6320,"ariaHidden":29},[28],[21,6322,6324,6327],{"className":6323},[33],[21,6325],{"className":6326,"style":38},[37],[21,6328,6330,6333],{"className":6329},[42],[21,6331,804],{"className":6332},[42,46],[21,6334,6336],{"className":6335},[51],[21,6337,6339,6359],{"className":6338},[55,56],[21,6340,6342,6356],{"className":6341},[60],[21,6343,6345],{"className":6344,"style":2424},[64],[21,6346,6347,6350],{"style":68},[21,6348],{"className":6349,"style":73},[72],[21,6351,6353],{"className":6352},[77,78,79,80],[21,6354,230],{"className":6355,"style":229},[42,46,80],[21,6357,89],{"className":6358},[88],[21,6360,6362],{"className":6361},[60],[21,6363,6365],{"className":6364,"style":96},[64],[21,6366],{},[177,6368,6370,6781,7050],{"type":6369},"proof",[11,6371,6372,6375,6376,6299,6468,6703,6704,6764,6765,6780],{},[158,6373,6374],{},"Proof."," Each term ",[21,6377,6379],{"className":6378},[24],[21,6380,6382],{"className":6381,"ariaHidden":29},[28],[21,6383,6385,6388,6428],{"className":6384},[33],[21,6386],{"className":6387,"style":2331},[37],[21,6389,6391,6394],{"className":6390},[42],[21,6392,203],{"className":6393,"style":202},[42,46],[21,6395,6397],{"className":6396},[51],[21,6398,6400,6420],{"className":6399},[55,56],[21,6401,6403,6417],{"className":6402},[60],[21,6404,6406],{"className":6405,"style":243},[64],[21,6407,6408,6411],{"style":312},[21,6409],{"className":6410,"style":73},[72],[21,6412,6414],{"className":6413},[77,78,79,80],[21,6415,3798],{"className":6416},[42,46,80],[21,6418,89],{"className":6419},[88],[21,6421,6423],{"className":6422},[60],[21,6424,6426],{"className":6425,"style":96},[64],[21,6427],{},[21,6429,6431,6434],{"className":6430},[42],[21,6432,230],{"className":6433,"style":229},[42,46],[21,6435,6437],{"className":6436},[51],[21,6438,6440,6460],{"className":6439},[55,56],[21,6441,6443,6457],{"className":6442},[60],[21,6444,6446],{"className":6445,"style":243},[64],[21,6447,6448,6451],{"style":246},[21,6449],{"className":6450,"style":73},[72],[21,6452,6454],{"className":6453},[77,78,79,80],[21,6455,3798],{"className":6456},[42,46,80],[21,6458,89],{"className":6459},[88],[21,6461,6463],{"className":6462},[60],[21,6464,6466],{"className":6465,"style":96},[64],[21,6467],{},[21,6469,6471],{"className":6470},[24],[21,6472,6474,6578,6694],{"className":6473,"ariaHidden":29},[28],[21,6475,6477,6480,6483,6486,6526,6566,6569,6572,6575],{"className":6476},[33],[21,6478],{"className":6479,"style":1365},[37],[21,6481,5782],{"className":6482},[42,772],[21,6484,5786],{"className":6485},[221],[21,6487,6489,6492],{"className":6488},[42],[21,6490,203],{"className":6491,"style":202},[42,46],[21,6493,6495],{"className":6494},[51],[21,6496,6498,6518],{"className":6497},[55,56],[21,6499,6501,6515],{"className":6500},[60],[21,6502,6504],{"className":6503,"style":243},[64],[21,6505,6506,6509],{"style":312},[21,6507],{"className":6508,"style":73},[72],[21,6510,6512],{"className":6511},[77,78,79,80],[21,6513,3798],{"className":6514},[42,46,80],[21,6516,89],{"className":6517},[88],[21,6519,6521],{"className":6520},[60],[21,6522,6524],{"className":6523,"style":96},[64],[21,6525],{},[21,6527,6529,6532],{"className":6528},[42],[21,6530,230],{"className":6531,"style":229},[42,46],[21,6533,6535],{"className":6534},[51],[21,6536,6538,6558],{"className":6537},[55,56],[21,6539,6541,6555],{"className":6540},[60],[21,6542,6544],{"className":6543,"style":243},[64],[21,6545,6546,6549],{"style":246},[21,6547],{"className":6548,"style":73},[72],[21,6550,6552],{"className":6551},[77,78,79,80],[21,6553,3798],{"className":6554},[42,46,80],[21,6556,89],{"className":6557},[88],[21,6559,6561],{"className":6560},[60],[21,6562,6564],{"className":6563,"style":96},[64],[21,6565],{},[21,6567,5830],{"className":6568},[273],[21,6570],{"className":6571,"style":686},[401],[21,6573,691],{"className":6574},[690],[21,6576],{"className":6577,"style":686},[401],[21,6579,6581,6584,6587,6590,6630,6633,6636,6639,6642,6682,6685,6688,6691],{"className":6580},[33],[21,6582],{"className":6583,"style":1365},[37],[21,6585,5782],{"className":6586},[42,772],[21,6588,5786],{"className":6589},[221],[21,6591,6593,6596],{"className":6592},[42],[21,6594,203],{"className":6595,"style":202},[42,46],[21,6597,6599],{"className":6598},[51],[21,6600,6602,6622],{"className":6601},[55,56],[21,6603,6605,6619],{"className":6604},[60],[21,6606,6608],{"className":6607,"style":243},[64],[21,6609,6610,6613],{"style":312},[21,6611],{"className":6612,"style":73},[72],[21,6614,6616],{"className":6615},[77,78,79,80],[21,6617,3798],{"className":6618},[42,46,80],[21,6620,89],{"className":6621},[88],[21,6623,6625],{"className":6624},[60],[21,6626,6628],{"className":6627,"style":96},[64],[21,6629],{},[21,6631,5830],{"className":6632},[273],[21,6634],{"className":6635,"style":402},[401],[21,6637,5782],{"className":6638},[42,772],[21,6640,5786],{"className":6641},[221],[21,6643,6645,6648],{"className":6644},[42],[21,6646,230],{"className":6647,"style":229},[42,46],[21,6649,6651],{"className":6650},[51],[21,6652,6654,6674],{"className":6653},[55,56],[21,6655,6657,6671],{"className":6656},[60],[21,6658,6660],{"className":6659,"style":243},[64],[21,6661,6662,6665],{"style":246},[21,6663],{"className":6664,"style":73},[72],[21,6666,6668],{"className":6667},[77,78,79,80],[21,6669,3798],{"className":6670},[42,46,80],[21,6672,89],{"className":6673},[88],[21,6675,6677],{"className":6676},[60],[21,6678,6680],{"className":6679,"style":96},[64],[21,6681],{},[21,6683,5830],{"className":6684},[273],[21,6686],{"className":6687,"style":686},[401],[21,6689,691],{"className":6690},[690],[21,6692],{"className":6693,"style":686},[401],[21,6695,6697,6700],{"className":6696},[33],[21,6698],{"className":6699,"style":112},[37],[21,6701,1373],{"className":6702},[42]," by independence, so ",[21,6705,6707],{"className":6706},[24],[21,6708,6710,6734,6755],{"className":6709,"ariaHidden":29},[28],[21,6711,6713,6716,6719,6722,6725,6728,6731],{"className":6712},[33],[21,6714],{"className":6715,"style":1365},[37],[21,6717,5782],{"className":6718},[42,772],[21,6720,5786],{"className":6721},[221],[21,6723,203],{"className":6724,"style":202},[42,46],[21,6726],{"className":6727,"style":1087},[401],[21,6729,1219],{"className":6730},[1218],[21,6732],{"className":6733,"style":1087},[401],[21,6735,6737,6740,6743,6746,6749,6752],{"className":6736},[33],[21,6738],{"className":6739,"style":1365},[37],[21,6741,230],{"className":6742,"style":229},[42,46],[21,6744,5830],{"className":6745},[273],[21,6747],{"className":6748,"style":686},[401],[21,6750,691],{"className":6751},[690],[21,6753],{"className":6754,"style":686},[401],[21,6756,6758,6761],{"className":6757},[33],[21,6759],{"className":6760,"style":112},[37],[21,6762,1373],{"className":6763},[42],". For the\nvariance, independence across ",[21,6766,6768],{"className":6767},[24],[21,6769,6771],{"className":6770,"ariaHidden":29},[28],[21,6772,6774,6777],{"className":6773},[33],[21,6775],{"className":6776,"style":3932},[37],[21,6778,3798],{"className":6779},[42,46]," makes the terms uncorrelated, so variances\nadd:",[21,6782,6784],{"className":6783},[1582],[21,6785,6787],{"className":6786},[24],[21,6788,6790,6817,6838],{"className":6789,"ariaHidden":29},[28],[21,6791,6793,6796,6802,6805,6808,6811,6814],{"className":6792},[33],[21,6794],{"className":6795,"style":1365},[37],[21,6797,6799],{"className":6798},[353],[21,6800,5931],{"className":6801},[42,3265],[21,6803,1369],{"className":6804},[221],[21,6806,203],{"className":6807,"style":202},[42,46],[21,6809],{"className":6810,"style":1087},[401],[21,6812,1219],{"className":6813},[1218],[21,6815],{"className":6816,"style":1087},[401],[21,6818,6820,6823,6826,6829,6832,6835],{"className":6819},[33],[21,6821],{"className":6822,"style":1365},[37],[21,6824,230],{"className":6825,"style":229},[42,46],[21,6827,1388],{"className":6828},[273],[21,6830],{"className":6831,"style":686},[401],[21,6833,691],{"className":6834},[690],[21,6836],{"className":6837,"style":686},[401],[21,6839,6841,6845,6952,6955,6961,6964,7004,7044,7047],{"className":6840},[33],[21,6842],{"className":6843,"style":6844},[37],"height:3.1306em;vertical-align:-1.2777em;",[21,6846,6848],{"className":6847},[353,1989],[21,6849,6851,6943],{"className":6850},[55,56],[21,6852,6854,6940],{"className":6853},[60],[21,6855,6858,6878,6888],{"className":6856,"style":6857},[64],"height:1.853em;",[21,6859,6860,6863],{"style":2002},[21,6861],{"className":6862,"style":2006},[72],[21,6864,6866],{"className":6865},[77,78,79,80],[21,6867,6869,6872,6875],{"className":6868},[42,80],[21,6870,3798],{"className":6871},[42,46,80],[21,6873,691],{"className":6874},[690,80],[21,6876,116],{"className":6877},[42,80],[21,6879,6880,6883],{"style":2015},[21,6881],{"className":6882,"style":2006},[72],[21,6884,6885],{},[21,6886,360],{"className":6887},[353,357,2024],[21,6889,6891,6894],{"style":6890},"top:-4.3169em;margin-left:0em;",[21,6892],{"className":6893,"style":2006},[72],[21,6895,6897],{"className":6896},[77,78,79,80],[21,6898,6900],{"className":6899},[42,80],[21,6901,6903,6906],{"className":6902},[42,80],[21,6904,804],{"className":6905},[42,46,80],[21,6907,6909],{"className":6908},[51],[21,6910,6912,6932],{"className":6911},[55,56],[21,6913,6915,6929],{"className":6914},[60],[21,6916,6918],{"className":6917,"style":817},[64],[21,6919,6920,6923],{"style":820},[21,6921],{"className":6922,"style":824},[72],[21,6924,6926],{"className":6925},[77,828,829,80],[21,6927,230],{"className":6928,"style":229},[42,46,80],[21,6930,89],{"className":6931},[88],[21,6933,6935],{"className":6934},[60],[21,6936,6938],{"className":6937,"style":842},[64],[21,6939],{},[21,6941,89],{"className":6942},[88],[21,6944,6946],{"className":6945},[60],[21,6947,6950],{"className":6948,"style":6949},[64],"height:1.2777em;",[21,6951],{},[21,6953],{"className":6954,"style":402},[401],[21,6956,6958],{"className":6957},[353],[21,6959,5931],{"className":6960},[42,3265],[21,6962,1369],{"className":6963},[221],[21,6965,6967,6970],{"className":6966},[42],[21,6968,203],{"className":6969,"style":202},[42,46],[21,6971,6973],{"className":6972},[51],[21,6974,6976,6996],{"className":6975},[55,56],[21,6977,6979,6993],{"className":6978},[60],[21,6980,6982],{"className":6981,"style":243},[64],[21,6983,6984,6987],{"style":312},[21,6985],{"className":6986,"style":73},[72],[21,6988,6990],{"className":6989},[77,78,79,80],[21,6991,3798],{"className":6992},[42,46,80],[21,6994,89],{"className":6995},[88],[21,6997,6999],{"className":6998},[60],[21,7000,7002],{"className":7001,"style":96},[64],[21,7003],{},[21,7005,7007,7010],{"className":7006},[42],[21,7008,230],{"className":7009,"style":229},[42,46],[21,7011,7013],{"className":7012},[51],[21,7014,7016,7036],{"className":7015},[55,56],[21,7017,7019,7033],{"className":7018},[60],[21,7020,7022],{"className":7021,"style":243},[64],[21,7023,7024,7027],{"style":246},[21,7025],{"className":7026,"style":73},[72],[21,7028,7030],{"className":7029},[77,78,79,80],[21,7031,3798],{"className":7032},[42,46,80],[21,7034,89],{"className":7035},[88],[21,7037,7039],{"className":7038},[60],[21,7040,7042],{"className":7041,"style":96},[64],[21,7043],{},[21,7045,1388],{"className":7046},[273],[21,7048,2126],{"className":7049},[42],[11,7051,7052,7053,7636,7637,7689,7690,7790,7791],{},"For a single term, ",[21,7054,7056],{"className":7055},[24],[21,7057,7059,7166,7298,7435,7573,7591,7609,7627],{"className":7058,"ariaHidden":29},[28],[21,7060,7062,7065,7071,7074,7114,7154,7157,7160,7163],{"className":7061},[33],[21,7063],{"className":7064,"style":1365},[37],[21,7066,7068],{"className":7067},[353],[21,7069,5931],{"className":7070},[42,3265],[21,7072,1369],{"className":7073},[221],[21,7075,7077,7080],{"className":7076},[42],[21,7078,203],{"className":7079,"style":202},[42,46],[21,7081,7083],{"className":7082},[51],[21,7084,7086,7106],{"className":7085},[55,56],[21,7087,7089,7103],{"className":7088},[60],[21,7090,7092],{"className":7091,"style":243},[64],[21,7093,7094,7097],{"style":312},[21,7095],{"className":7096,"style":73},[72],[21,7098,7100],{"className":7099},[77,78,79,80],[21,7101,3798],{"className":7102},[42,46,80],[21,7104,89],{"className":7105},[88],[21,7107,7109],{"className":7108},[60],[21,7110,7112],{"className":7111,"style":96},[64],[21,7113],{},[21,7115,7117,7120],{"className":7116},[42],[21,7118,230],{"className":7119,"style":229},[42,46],[21,7121,7123],{"className":7122},[51],[21,7124,7126,7146],{"className":7125},[55,56],[21,7127,7129,7143],{"className":7128},[60],[21,7130,7132],{"className":7131,"style":243},[64],[21,7133,7134,7137],{"style":246},[21,7135],{"className":7136,"style":73},[72],[21,7138,7140],{"className":7139},[77,78,79,80],[21,7141,3798],{"className":7142},[42,46,80],[21,7144,89],{"className":7145},[88],[21,7147,7149],{"className":7148},[60],[21,7150,7152],{"className":7151,"style":96},[64],[21,7153],{},[21,7155,1388],{"className":7156},[273],[21,7158],{"className":7159,"style":686},[401],[21,7161,691],{"className":7162},[690],[21,7164],{"className":7165,"style":686},[401],[21,7167,7169,7173,7176,7179,7233,7285,7288,7291,7295],{"className":7168},[33],[21,7170],{"className":7171,"style":7172},[37],"height:1.0728em;vertical-align:-0.2587em;",[21,7174,5782],{"className":7175},[42,772],[21,7177,5786],{"className":7178},[221],[21,7180,7182,7185],{"className":7181},[42],[21,7183,203],{"className":7184,"style":202},[42,46],[21,7186,7188],{"className":7187},[51],[21,7189,7191,7224],{"className":7190},[55,56],[21,7192,7194,7221],{"className":7193},[60],[21,7195,7198,7210],{"className":7196,"style":7197},[64],"height:0.8141em;",[21,7199,7201,7204],{"style":7200},"top:-2.4413em;margin-left:-0.0359em;margin-right:0.05em;",[21,7202],{"className":7203,"style":73},[72],[21,7205,7207],{"className":7206},[77,78,79,80],[21,7208,3798],{"className":7209},[42,46,80],[21,7211,7212,7215],{"style":788},[21,7213],{"className":7214,"style":73},[72],[21,7216,7218],{"className":7217},[77,78,79,80],[21,7219,5228],{"className":7220},[42,80],[21,7222,89],{"className":7223},[88],[21,7225,7227],{"className":7226},[60],[21,7228,7231],{"className":7229,"style":7230},[64],"height:0.2587em;",[21,7232],{},[21,7234,7236,7239],{"className":7235},[42],[21,7237,230],{"className":7238,"style":229},[42,46],[21,7240,7242],{"className":7241},[51],[21,7243,7245,7277],{"className":7244},[55,56],[21,7246,7248,7274],{"className":7247},[60],[21,7249,7251,7263],{"className":7250,"style":7197},[64],[21,7252,7254,7257],{"style":7253},"top:-2.4413em;margin-left:-0.0315em;margin-right:0.05em;",[21,7255],{"className":7256,"style":73},[72],[21,7258,7260],{"className":7259},[77,78,79,80],[21,7261,3798],{"className":7262},[42,46,80],[21,7264,7265,7268],{"style":788},[21,7266],{"className":7267,"style":73},[72],[21,7269,7271],{"className":7270},[77,78,79,80],[21,7272,5228],{"className":7273},[42,80],[21,7275,89],{"className":7276},[88],[21,7278,7280],{"className":7279},[60],[21,7281,7283],{"className":7282,"style":7230},[64],[21,7284],{},[21,7286,5830],{"className":7287},[273],[21,7289],{"className":7290,"style":1087},[401],[21,7292,7294],{"className":7293},[1218],"−",[21,7296],{"className":7297,"style":1087},[401],[21,7299,7301,7305,7308,7311,7314,7354,7394,7397,7426,7429,7432],{"className":7300},[33],[21,7302],{"className":7303,"style":7304},[37],"height:1.0641em;vertical-align:-0.25em;",[21,7306,1369],{"className":7307},[221],[21,7309,5782],{"className":7310},[42,772],[21,7312,5786],{"className":7313},[221],[21,7315,7317,7320],{"className":7316},[42],[21,7318,203],{"className":7319,"style":202},[42,46],[21,7321,7323],{"className":7322},[51],[21,7324,7326,7346],{"className":7325},[55,56],[21,7327,7329,7343],{"className":7328},[60],[21,7330,7332],{"className":7331,"style":243},[64],[21,7333,7334,7337],{"style":312},[21,7335],{"className":7336,"style":73},[72],[21,7338,7340],{"className":7339},[77,78,79,80],[21,7341,3798],{"className":7342},[42,46,80],[21,7344,89],{"className":7345},[88],[21,7347,7349],{"className":7348},[60],[21,7350,7352],{"className":7351,"style":96},[64],[21,7353],{},[21,7355,7357,7360],{"className":7356},[42],[21,7358,230],{"className":7359,"style":229},[42,46],[21,7361,7363],{"className":7362},[51],[21,7364,7366,7386],{"className":7365},[55,56],[21,7367,7369,7383],{"className":7368},[60],[21,7370,7372],{"className":7371,"style":243},[64],[21,7373,7374,7377],{"style":246},[21,7375],{"className":7376,"style":73},[72],[21,7378,7380],{"className":7379},[77,78,79,80],[21,7381,3798],{"className":7382},[42,46,80],[21,7384,89],{"className":7385},[88],[21,7387,7389],{"className":7388},[60],[21,7390,7392],{"className":7391,"style":96},[64],[21,7393],{},[21,7395,5830],{"className":7396},[273],[21,7398,7400,7403],{"className":7399},[273],[21,7401,1388],{"className":7402},[273],[21,7404,7406],{"className":7405},[51],[21,7407,7409],{"className":7408},[55],[21,7410,7412],{"className":7411},[60],[21,7413,7415],{"className":7414,"style":7197},[64],[21,7416,7417,7420],{"style":788},[21,7418],{"className":7419,"style":73},[72],[21,7421,7423],{"className":7422},[77,78,79,80],[21,7424,5228],{"className":7425},[42,80],[21,7427],{"className":7428,"style":686},[401],[21,7430,691],{"className":7431},[690],[21,7433],{"className":7434,"style":686},[401],[21,7436,7438,7441,7444,7447,7498,7501,7504,7507,7510,7561,7564,7567,7570],{"className":7437},[33],[21,7439],{"className":7440,"style":7172},[37],[21,7442,5782],{"className":7443},[42,772],[21,7445,5786],{"className":7446},[221],[21,7448,7450,7453],{"className":7449},[42],[21,7451,203],{"className":7452,"style":202},[42,46],[21,7454,7456],{"className":7455},[51],[21,7457,7459,7490],{"className":7458},[55,56],[21,7460,7462,7487],{"className":7461},[60],[21,7463,7465,7476],{"className":7464,"style":7197},[64],[21,7466,7467,7470],{"style":7200},[21,7468],{"className":7469,"style":73},[72],[21,7471,7473],{"className":7472},[77,78,79,80],[21,7474,3798],{"className":7475},[42,46,80],[21,7477,7478,7481],{"style":788},[21,7479],{"className":7480,"style":73},[72],[21,7482,7484],{"className":7483},[77,78,79,80],[21,7485,5228],{"className":7486},[42,80],[21,7488,89],{"className":7489},[88],[21,7491,7493],{"className":7492},[60],[21,7494,7496],{"className":7495,"style":7230},[64],[21,7497],{},[21,7499,5830],{"className":7500},[273],[21,7502],{"className":7503,"style":402},[401],[21,7505,5782],{"className":7506},[42,772],[21,7508,5786],{"className":7509},[221],[21,7511,7513,7516],{"className":7512},[42],[21,7514,230],{"className":7515,"style":229},[42,46],[21,7517,7519],{"className":7518},[51],[21,7520,7522,7553],{"className":7521},[55,56],[21,7523,7525,7550],{"className":7524},[60],[21,7526,7528,7539],{"className":7527,"style":7197},[64],[21,7529,7530,7533],{"style":7253},[21,7531],{"className":7532,"style":73},[72],[21,7534,7536],{"className":7535},[77,78,79,80],[21,7537,3798],{"className":7538},[42,46,80],[21,7540,7541,7544],{"style":788},[21,7542],{"className":7543,"style":73},[72],[21,7545,7547],{"className":7546},[77,78,79,80],[21,7548,5228],{"className":7549},[42,80],[21,7551,89],{"className":7552},[88],[21,7554,7556],{"className":7555},[60],[21,7557,7559],{"className":7558,"style":7230},[64],[21,7560],{},[21,7562,5830],{"className":7563},[273],[21,7565],{"className":7566,"style":1087},[401],[21,7568,7294],{"className":7569},[1218],[21,7571],{"className":7572,"style":1087},[401],[21,7574,7576,7579,7582,7585,7588],{"className":7575},[33],[21,7577],{"className":7578,"style":112},[37],[21,7580,1373],{"className":7581},[42],[21,7583],{"className":7584,"style":686},[401],[21,7586,691],{"className":7587},[690],[21,7589],{"className":7590,"style":686},[401],[21,7592,7594,7597,7600,7603,7606],{"className":7593},[33],[21,7595],{"className":7596,"style":112},[37],[21,7598,116],{"className":7599},[42],[21,7601],{"className":7602,"style":1087},[401],[21,7604,1219],{"className":7605},[1218],[21,7607],{"className":7608,"style":1087},[401],[21,7610,7612,7615,7618,7621,7624],{"className":7611},[33],[21,7613],{"className":7614,"style":112},[37],[21,7616,116],{"className":7617},[42],[21,7619],{"className":7620,"style":686},[401],[21,7622,691],{"className":7623},[690],[21,7625],{"className":7626,"style":686},[401],[21,7628,7630,7633],{"className":7629},[33],[21,7631],{"className":7632,"style":112},[37],[21,7634,116],{"className":7635},[42],", again using independence. Summing ",[21,7638,7640],{"className":7639},[24],[21,7641,7643],{"className":7642,"ariaHidden":29},[28],[21,7644,7646,7649],{"className":7645},[33],[21,7647],{"className":7648,"style":38},[37],[21,7650,7652,7655],{"className":7651},[42],[21,7653,804],{"className":7654},[42,46],[21,7656,7658],{"className":7657},[51],[21,7659,7661,7681],{"className":7660},[55,56],[21,7662,7664,7678],{"className":7663},[60],[21,7665,7667],{"className":7666,"style":2424},[64],[21,7668,7669,7672],{"style":68},[21,7670],{"className":7671,"style":73},[72],[21,7673,7675],{"className":7674},[77,78,79,80],[21,7676,230],{"className":7677,"style":229},[42,46,80],[21,7679,89],{"className":7680},[88],[21,7682,7684],{"className":7683},[60],[21,7685,7687],{"className":7686,"style":96},[64],[21,7688],{}," unit-variance terms gives\n",[21,7691,7693],{"className":7692},[24],[21,7694,7696,7723,7744],{"className":7695,"ariaHidden":29},[28],[21,7697,7699,7702,7708,7711,7714,7717,7720],{"className":7698},[33],[21,7700],{"className":7701,"style":1365},[37],[21,7703,7705],{"className":7704},[353],[21,7706,5931],{"className":7707},[42,3265],[21,7709,1369],{"className":7710},[221],[21,7712,203],{"className":7713,"style":202},[42,46],[21,7715],{"className":7716,"style":1087},[401],[21,7718,1219],{"className":7719},[1218],[21,7721],{"className":7722,"style":1087},[401],[21,7724,7726,7729,7732,7735,7738,7741],{"className":7725},[33],[21,7727],{"className":7728,"style":1365},[37],[21,7730,230],{"className":7731,"style":229},[42,46],[21,7733,1388],{"className":7734},[273],[21,7736],{"className":7737,"style":686},[401],[21,7739,691],{"className":7740},[690],[21,7742],{"className":7743,"style":686},[401],[21,7745,7747,7750],{"className":7746},[33],[21,7748],{"className":7749,"style":38},[37],[21,7751,7753,7756],{"className":7752},[42],[21,7754,804],{"className":7755},[42,46],[21,7757,7759],{"className":7758},[51],[21,7760,7762,7782],{"className":7761},[55,56],[21,7763,7765,7779],{"className":7764},[60],[21,7766,7768],{"className":7767,"style":2424},[64],[21,7769,7770,7773],{"style":68},[21,7771],{"className":7772,"style":73},[72],[21,7774,7776],{"className":7775},[77,78,79,80],[21,7777,230],{"className":7778,"style":229},[42,46,80],[21,7780,89],{"className":7781},[88],[21,7783,7785],{"className":7784},[60],[21,7786,7788],{"className":7787,"style":96},[64],[21,7789],{},". ",[21,7792,7794],{"className":7793},[24],[21,7795,7797],{"className":7796,"ariaHidden":29},[28],[21,7798,7800,7804],{"className":7799},[33],[21,7801],{"className":7802,"style":7803},[37],"height:0.675em;",[21,7805,7809],{"className":7806},[7807,7808],"enclosing","qed",[21,7810,7813],{"className":7811},[42,7812],"amsrm","□",[11,7815,7816,7817,7913,7914,7917,7918,8031,8032,8084,8085,8181,8182,8349],{},"So the raw scores have standard deviation ",[21,7818,7820],{"className":7819},[24],[21,7821,7823],{"className":7822,"ariaHidden":29},[28],[21,7824,7826,7829],{"className":7825},[33],[21,7827],{"className":7828,"style":3965},[37],[21,7830,7832],{"className":7831},[42,3321],[21,7833,7835,7905],{"className":7834},[55,56],[21,7836,7838,7902],{"className":7837},[60],[21,7839,7841,7890],{"className":7840,"style":3978},[64],[21,7842,7844,7847],{"className":7843,"style":2593},[2612],[21,7845],{"className":7846,"style":1678},[72],[21,7848,7850],{"className":7849,"style":3341},[42],[21,7851,7853,7856],{"className":7852},[42],[21,7854,804],{"className":7855},[42,46],[21,7857,7859],{"className":7858},[51],[21,7860,7862,7882],{"className":7861},[55,56],[21,7863,7865,7879],{"className":7864},[60],[21,7866,7868],{"className":7867,"style":2424},[64],[21,7869,7870,7873],{"style":68},[21,7871],{"className":7872,"style":73},[72],[21,7874,7876],{"className":7875},[77,78,79,80],[21,7877,230],{"className":7878,"style":229},[42,46,80],[21,7880,89],{"className":7881},[88],[21,7883,7885],{"className":7884},[60],[21,7886,7888],{"className":7887,"style":96},[64],[21,7889],{},[21,7891,7892,7895],{"style":4030},[21,7893],{"className":7894,"style":1678},[72],[21,7896,7898],{"className":7897,"style":3392},[3391],[2628,7899,7900],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,7901],{"d":3399},[21,7903,89],{"className":7904},[88],[21,7906,7908],{"className":7907},[60],[21,7909,7911],{"className":7910,"style":4050},[64],[21,7912],{},", which ",[163,7915,7916],{},"grows"," with the\nmodel dimension. Feed scores of magnitude ",[21,7919,7921],{"className":7920},[24],[21,7922,7924,7941],{"className":7923,"ariaHidden":29},[28],[21,7925,7927,7931,7935,7938],{"className":7926},[33],[21,7928],{"className":7929,"style":7930},[37],"height:0.3669em;",[21,7932,7934],{"className":7933},[690],"∼",[21,7936],{"className":7937,"style":3270},[401],[21,7939],{"className":7940,"style":686},[401],[21,7942,7944,7947],{"className":7943},[33],[21,7945],{"className":7946,"style":3965},[37],[21,7948,7950],{"className":7949},[42,3321],[21,7951,7953,8023],{"className":7952},[55,56],[21,7954,7956,8020],{"className":7955},[60],[21,7957,7959,8008],{"className":7958,"style":3978},[64],[21,7960,7962,7965],{"className":7961,"style":2593},[2612],[21,7963],{"className":7964,"style":1678},[72],[21,7966,7968],{"className":7967,"style":3341},[42],[21,7969,7971,7974],{"className":7970},[42],[21,7972,804],{"className":7973},[42,46],[21,7975,7977],{"className":7976},[51],[21,7978,7980,8000],{"className":7979},[55,56],[21,7981,7983,7997],{"className":7982},[60],[21,7984,7986],{"className":7985,"style":2424},[64],[21,7987,7988,7991],{"style":68},[21,7989],{"className":7990,"style":73},[72],[21,7992,7994],{"className":7993},[77,78,79,80],[21,7995,230],{"className":7996,"style":229},[42,46,80],[21,7998,89],{"className":7999},[88],[21,8001,8003],{"className":8002},[60],[21,8004,8006],{"className":8005,"style":96},[64],[21,8007],{},[21,8009,8010,8013],{"style":4030},[21,8011],{"className":8012,"style":1678},[72],[21,8014,8016],{"className":8015,"style":3392},[3391],[2628,8017,8018],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,8019],{"d":3399},[21,8021,89],{"className":8022},[88],[21,8024,8026],{"className":8025},[60],[21,8027,8029],{"className":8028,"style":4050},[64],[21,8030],{}," into a softmax and,\nfor ",[21,8033,8035],{"className":8034},[24],[21,8036,8038],{"className":8037,"ariaHidden":29},[28],[21,8039,8041,8044],{"className":8040},[33],[21,8042],{"className":8043,"style":38},[37],[21,8045,8047,8050],{"className":8046},[42],[21,8048,804],{"className":8049},[42,46],[21,8051,8053],{"className":8052},[51],[21,8054,8056,8076],{"className":8055},[55,56],[21,8057,8059,8073],{"className":8058},[60],[21,8060,8062],{"className":8061,"style":2424},[64],[21,8063,8064,8067],{"style":68},[21,8065],{"className":8066,"style":73},[72],[21,8068,8070],{"className":8069},[77,78,79,80],[21,8071,230],{"className":8072,"style":229},[42,46,80],[21,8074,89],{"className":8075},[88],[21,8077,8079],{"className":8078},[60],[21,8080,8082],{"className":8081,"style":96},[64],[21,8083],{}," in the hundreds, the largest entry dwarfs the rest: the softmax\nsaturates onto one position, approaching a one-hot vector. In that regime its\nJacobian is nearly zero, and almost no gradient flows back to the queries and keys.\nDividing by ",[21,8086,8088],{"className":8087},[24],[21,8089,8091],{"className":8090,"ariaHidden":29},[28],[21,8092,8094,8097],{"className":8093},[33],[21,8095],{"className":8096,"style":3965},[37],[21,8098,8100],{"className":8099},[42,3321],[21,8101,8103,8173],{"className":8102},[55,56],[21,8104,8106,8170],{"className":8105},[60],[21,8107,8109,8158],{"className":8108,"style":3978},[64],[21,8110,8112,8115],{"className":8111,"style":2593},[2612],[21,8113],{"className":8114,"style":1678},[72],[21,8116,8118],{"className":8117,"style":3341},[42],[21,8119,8121,8124],{"className":8120},[42],[21,8122,804],{"className":8123},[42,46],[21,8125,8127],{"className":8126},[51],[21,8128,8130,8150],{"className":8129},[55,56],[21,8131,8133,8147],{"className":8132},[60],[21,8134,8136],{"className":8135,"style":2424},[64],[21,8137,8138,8141],{"style":68},[21,8139],{"className":8140,"style":73},[72],[21,8142,8144],{"className":8143},[77,78,79,80],[21,8145,230],{"className":8146,"style":229},[42,46,80],[21,8148,89],{"className":8149},[88],[21,8151,8153],{"className":8152},[60],[21,8154,8156],{"className":8155,"style":96},[64],[21,8157],{},[21,8159,8160,8163],{"style":4030},[21,8161],{"className":8162,"style":1678},[72],[21,8164,8166],{"className":8165,"style":3392},[3391],[2628,8167,8168],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,8169],{"d":3399},[21,8171,89],{"className":8172},[88],[21,8174,8176],{"className":8175},[60],[21,8177,8179],{"className":8178,"style":4050},[64],[21,8180],{}," rescales the scores to unit variance, ",[21,8183,8185],{"className":8184},[24],[21,8186,8188,8340],{"className":8187,"ariaHidden":29},[28],[21,8189,8191,8195,8201,8204,8207,8331,8334,8337],{"className":8190},[33],[21,8192],{"className":8193,"style":8194},[37],"height:1.2072em;vertical-align:-0.35em;",[21,8196,8198],{"className":8197},[353],[21,8199,5931],{"className":8200},[42,3265],[21,8202],{"className":8203,"style":3270},[401],[21,8205],{"className":8206,"style":402},[401],[21,8208,8210,8216,8219,8222,8225,8228,8231,8234,8237,8241,8325],{"className":8209},[2521],[21,8211,8213],{"className":8212,"style":3281},[221,3280],[21,8214,1369],{"className":8215},[3285,829],[21,8217,1369],{"className":8218},[221],[21,8220,203],{"className":8221,"style":202},[42,46],[21,8223],{"className":8224,"style":1087},[401],[21,8226,1219],{"className":8227},[1218],[21,8229],{"className":8230,"style":1087},[401],[21,8232,230],{"className":8233,"style":229},[42,46],[21,8235,1388],{"className":8236},[273],[21,8238,8240],{"className":8239},[42],"\u002F",[21,8242,8244],{"className":8243},[42,3321],[21,8245,8247,8317],{"className":8246},[55,56],[21,8248,8250,8314],{"className":8249},[60],[21,8251,8253,8302],{"className":8252,"style":3978},[64],[21,8254,8256,8259],{"className":8255,"style":2593},[2612],[21,8257],{"className":8258,"style":1678},[72],[21,8260,8262],{"className":8261,"style":3341},[42],[21,8263,8265,8268],{"className":8264},[42],[21,8266,804],{"className":8267},[42,46],[21,8269,8271],{"className":8270},[51],[21,8272,8274,8294],{"className":8273},[55,56],[21,8275,8277,8291],{"className":8276},[60],[21,8278,8280],{"className":8279,"style":2424},[64],[21,8281,8282,8285],{"style":68},[21,8283],{"className":8284,"style":73},[72],[21,8286,8288],{"className":8287},[77,78,79,80],[21,8289,230],{"className":8290,"style":229},[42,46,80],[21,8292,89],{"className":8293},[88],[21,8295,8297],{"className":8296},[60],[21,8298,8300],{"className":8299,"style":96},[64],[21,8301],{},[21,8303,8304,8307],{"style":4030},[21,8305],{"className":8306,"style":1678},[72],[21,8308,8310],{"className":8309,"style":3392},[3391],[2628,8311,8312],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,8313],{"d":3399},[21,8315,89],{"className":8316},[88],[21,8318,8320],{"className":8319},[60],[21,8321,8323],{"className":8322,"style":4050},[64],[21,8324],{},[21,8326,8328],{"className":8327,"style":3281},[273,3280],[21,8329,1388],{"className":8330},[3285,829],[21,8332],{"className":8333,"style":686},[401],[21,8335,691],{"className":8336},[690],[21,8338],{"className":8339,"style":686},[401],[21,8341,8343,8346],{"className":8342},[33],[21,8344],{"className":8345,"style":112},[37],[21,8347,116],{"className":8348},[42],", keeping the softmax in its sensitive,\nhigh-gradient region regardless of dimension.",[2128,8351],{"hash":8352},"98babe1adf74ccc313bc8d01040b34a1d80d5de4bd89ed7550429f54c7bdd188",[8354,8355,8359],"pre",{"className":8356,"code":8357,"language":8358,"meta":6,"style":6},"language-algorithm shiki shiki-themes Vesper Light - Orange Boost (Quick Open Adjusted) vesper","caption: $\\textsc{ScaledDotProductAttention}(Q, K, V)$ — soft content-based lookup\n$S \\gets Q K^{T}$ \u002F\u002F $n \\times n$ raw scores\n$S \\gets S \u002F \\sqrt{d_k}$ \u002F\u002F unit-variance correction\nif masked then \u002F\u002F decoder self-attention\n  for each $i, j$ with $j > i$ do\n    $S_{ij} \\gets -\\infty$ \u002F\u002F forbid attending to the future\nfor each row $i$ do\n  $A_{i} \\gets \\softmax(S_{i})$ \u002F\u002F weights sum to $1$\nreturn $A V$ \u002F\u002F blend the values\n","algorithm",[8360,8361,8362,8369,8375,8381,8387,8393,8399,8405,8411],"code",{"__ignoreMap":6},[21,8363,8366],{"class":8364,"line":8365},"line",1,[21,8367,8368],{},"caption: $\\textsc{ScaledDotProductAttention}(Q, K, V)$ — soft content-based lookup\n",[21,8370,8372],{"class":8364,"line":8371},2,[21,8373,8374],{},"$S \\gets Q K^{T}$ \u002F\u002F $n \\times n$ raw scores\n",[21,8376,8378],{"class":8364,"line":8377},3,[21,8379,8380],{},"$S \\gets S \u002F \\sqrt{d_k}$ \u002F\u002F unit-variance correction\n",[21,8382,8384],{"class":8364,"line":8383},4,[21,8385,8386],{},"if masked then \u002F\u002F decoder self-attention\n",[21,8388,8390],{"class":8364,"line":8389},5,[21,8391,8392],{},"  for each $i, j$ with $j > i$ do\n",[21,8394,8396],{"class":8364,"line":8395},6,[21,8397,8398],{},"    $S_{ij} \\gets -\\infty$ \u002F\u002F forbid attending to the future\n",[21,8400,8402],{"class":8364,"line":8401},7,[21,8403,8404],{},"for each row $i$ do\n",[21,8406,8408],{"class":8364,"line":8407},8,[21,8409,8410],{},"  $A_{i} \\gets \\softmax(S_{i})$ \u002F\u002F weights sum to $1$\n",[21,8412,8414],{"class":8364,"line":8413},9,[21,8415,8416],{},"return $A V$ \u002F\u002F blend the values\n",[5230,8418,8420],{"id":8419},"a-worked-example","A worked example",[11,8422,8423,8424,8549],{},"Push numbers through the formula. Take three\ntokens with ",[21,8425,8427],{"className":8426},[24],[21,8428,8430,8485,8540],{"className":8429,"ariaHidden":29},[28],[21,8431,8433,8436,8476,8479,8482],{"className":8432},[33],[21,8434],{"className":8435,"style":38},[37],[21,8437,8439,8442],{"className":8438},[42],[21,8440,804],{"className":8441},[42,46],[21,8443,8445],{"className":8444},[51],[21,8446,8448,8468],{"className":8447},[55,56],[21,8449,8451,8465],{"className":8450},[60],[21,8452,8454],{"className":8453,"style":2424},[64],[21,8455,8456,8459],{"style":68},[21,8457],{"className":8458,"style":73},[72],[21,8460,8462],{"className":8461},[77,78,79,80],[21,8463,230],{"className":8464,"style":229},[42,46,80],[21,8466,89],{"className":8467},[88],[21,8469,8471],{"className":8470},[60],[21,8472,8474],{"className":8473,"style":96},[64],[21,8475],{},[21,8477],{"className":8478,"style":686},[401],[21,8480,691],{"className":8481},[690],[21,8483],{"className":8484,"style":686},[401],[21,8486,8488,8491,8531,8534,8537],{"className":8487},[33],[21,8489],{"className":8490,"style":38},[37],[21,8492,8494,8497],{"className":8493},[42],[21,8495,804],{"className":8496},[42,46],[21,8498,8500],{"className":8499},[51],[21,8501,8503,8523],{"className":8502},[55,56],[21,8504,8506,8520],{"className":8505},[60],[21,8507,8509],{"className":8508,"style":2478},[64],[21,8510,8511,8514],{"style":68},[21,8512],{"className":8513,"style":73},[72],[21,8515,8517],{"className":8516},[77,78,79,80],[21,8518,297],{"className":8519,"style":202},[42,46,80],[21,8521,89],{"className":8522},[88],[21,8524,8526],{"className":8525},[60],[21,8527,8529],{"className":8528,"style":96},[64],[21,8530],{},[21,8532],{"className":8533,"style":686},[401],[21,8535,691],{"className":8536},[690],[21,8538],{"className":8539,"style":686},[401],[21,8541,8543,8546],{"className":8542},[33],[21,8544],{"className":8545,"style":112},[37],[21,8547,5228],{"className":8548},[42],", and, to keep the arithmetic legible, let the\nqueries and keys coincide, so this is self-attention:",[21,8551,8553],{"className":8552},[1582],[21,8554,8556],{"className":8555},[24],[21,8557,8559,8577,8595,8836,8904,9127],{"className":8558,"ariaHidden":29},[28],[21,8560,8562,8565,8568,8571,8574],{"className":8561},[33],[21,8563],{"className":8564,"style":2166},[37],[21,8566,739],{"className":8567},[42,46],[21,8569],{"className":8570,"style":686},[401],[21,8572,691],{"className":8573},[690],[21,8575],{"className":8576,"style":686},[401],[21,8578,8580,8583,8586,8589,8592],{"className":8579},[33],[21,8581],{"className":8582,"style":2200},[37],[21,8584,920],{"className":8585,"style":919},[42,46],[21,8587],{"className":8588,"style":686},[401],[21,8590,691],{"className":8591},[690],[21,8593],{"className":8594,"style":686},[401],[21,8596,8598,8602,8821,8824,8827,8830,8833],{"className":8597},[33],[21,8599],{"className":8600,"style":8601},[37],"height:3.6em;vertical-align:-1.55em;",[21,8603,8605,8652,8783],{"className":8604},[2521],[21,8606,8608],{"className":8607},[221],[21,8609,8612],{"className":8610},[3285,8611],"mult",[21,8613,8615,8643],{"className":8614},[55,56],[21,8616,8618,8640],{"className":8617},[60],[21,8619,8622],{"className":8620,"style":8621},[64],"height:2.05em;",[21,8623,8625,8629],{"style":8624},"top:-4.05em;",[21,8626],{"className":8627,"style":8628},[72],"height:5.6em;",[21,8630,8632],{"style":8631},"width:0.667em;height:3.6em;",[2628,8633,8637],{"xmlns":2630,"width":8634,"height":8635,"viewBox":8636},"0.667em","3.6em","0 0 667 3600",[2636,8638],{"d":8639},"M403 1759 V84 H666 V0 H319 V1759 v0 v1759 v84 h347 v-84\nH403z M403 1759 V0 H319 V1759 v0 v1759 v84 h84z",[21,8641,89],{"className":8642},[88],[21,8644,8646],{"className":8645},[60],[21,8647,8650],{"className":8648,"style":8649},[64],"height:1.55em;",[21,8651],{},[21,8653,8655],{"className":8654},[42],[21,8656,8659,8719,8724,8727],{"className":8657},[8658],"mtable",[21,8660,8663],{"className":8661},[8662],"col-align-c",[21,8664,8666,8711],{"className":8665},[55,56],[21,8667,8669,8708],{"className":8668},[60],[21,8670,8672,8684,8696],{"className":8671,"style":8621},[64],[21,8673,8675,8678],{"style":8674},"top:-4.21em;",[21,8676],{"className":8677,"style":1678},[72],[21,8679,8681],{"className":8680},[42],[21,8682,116],{"className":8683},[42],[21,8685,8687,8690],{"style":8686},"top:-3.01em;",[21,8688],{"className":8689,"style":1678},[72],[21,8691,8693],{"className":8692},[42],[21,8694,1373],{"className":8695},[42],[21,8697,8699,8702],{"style":8698},"top:-1.81em;",[21,8700],{"className":8701,"style":1678},[72],[21,8703,8705],{"className":8704},[42],[21,8706,116],{"className":8707},[42],[21,8709,89],{"className":8710},[88],[21,8712,8714],{"className":8713},[60],[21,8715,8717],{"className":8716,"style":8649},[64],[21,8718],{},[21,8720],{"className":8721,"style":8723},[8722],"arraycolsep","width:0.5em;",[21,8725],{"className":8726,"style":8723},[8722],[21,8728,8730],{"className":8729},[8662],[21,8731,8733,8775],{"className":8732},[55,56],[21,8734,8736,8772],{"className":8735},[60],[21,8737,8739,8750,8761],{"className":8738,"style":8621},[64],[21,8740,8741,8744],{"style":8674},[21,8742],{"className":8743,"style":1678},[72],[21,8745,8747],{"className":8746},[42],[21,8748,1373],{"className":8749},[42],[21,8751,8752,8755],{"style":8686},[21,8753],{"className":8754,"style":1678},[72],[21,8756,8758],{"className":8757},[42],[21,8759,116],{"className":8760},[42],[21,8762,8763,8766],{"style":8698},[21,8764],{"className":8765,"style":1678},[72],[21,8767,8769],{"className":8768},[42],[21,8770,116],{"className":8771},[42],[21,8773,89],{"className":8774},[88],[21,8776,8778],{"className":8777},[60],[21,8779,8781],{"className":8780,"style":8649},[64],[21,8782],{},[21,8784,8786],{"className":8785},[273],[21,8787,8789],{"className":8788},[3285,8611],[21,8790,8792,8813],{"className":8791},[55,56],[21,8793,8795,8810],{"className":8794},[60],[21,8796,8798],{"className":8797,"style":8621},[64],[21,8799,8800,8803],{"style":8624},[21,8801],{"className":8802,"style":8628},[72],[21,8804,8805],{"style":8631},[2628,8806,8807],{"xmlns":2630,"width":8634,"height":8635,"viewBox":8636},[2636,8808],{"d":8809},"M347 1759 V0 H0 V84 H263 V1759 v0 v1759 H0 v84 H347z\nM347 1759 V0 H263 V1759 v0 v1759 h84z",[21,8811,89],{"className":8812},[88],[21,8814,8816],{"className":8815},[60],[21,8817,8819],{"className":8818,"style":8649},[64],[21,8820],{},[21,8822],{"className":8823,"style":686},[401],[21,8825],{"className":8826,"style":686},[401],[21,8828,3730],{"className":8829},[690],[21,8831],{"className":8832,"style":686},[401],[21,8834],{"className":8835,"style":686},[401],[21,8837,8839,8843,8883,8886,8889,8892,8895,8898,8901],{"className":8838},[33],[21,8840],{"className":8841,"style":8842},[37],"height:1.0585em;vertical-align:-0.1944em;",[21,8844,8846,8849],{"className":8845},[42],[21,8847,773],{"className":8848},[42,772],[21,8850,8852],{"className":8851},[51],[21,8853,8855],{"className":8854},[55],[21,8856,8858],{"className":8857},[60],[21,8859,8862],{"className":8860,"style":8861},[64],"height:0.8641em;",[21,8863,8864,8867],{"style":3127},[21,8865],{"className":8866,"style":73},[72],[21,8868,8870],{"className":8869},[77,78,79,80],[21,8871,8873,8877,8880],{"className":8872},[42,80],[21,8874,8876],{"className":8875},[42,80],"3",[21,8878,2550],{"className":8879},[1218,80],[21,8881,5228],{"className":8882},[42,80],[21,8884,1378],{"className":8885},[1377],[21,8887],{"className":8888,"style":1958},[401],[21,8890],{"className":8891,"style":402},[401],[21,8893,1088],{"className":8894,"style":1087},[42,46],[21,8896],{"className":8897,"style":686},[401],[21,8899,691],{"className":8900},[690],[21,8902],{"className":8903,"style":686},[401],[21,8905,8907,8910,9112,9115,9118,9121,9124],{"className":8906},[33],[21,8908],{"className":8909,"style":8601},[37],[21,8911,8913,8950,9075],{"className":8912},[2521],[21,8914,8916],{"className":8915},[221],[21,8917,8919],{"className":8918},[3285,8611],[21,8920,8922,8942],{"className":8921},[55,56],[21,8923,8925,8939],{"className":8924},[60],[21,8926,8928],{"className":8927,"style":8621},[64],[21,8929,8930,8933],{"style":8624},[21,8931],{"className":8932,"style":8628},[72],[21,8934,8935],{"style":8631},[2628,8936,8937],{"xmlns":2630,"width":8634,"height":8635,"viewBox":8636},[2636,8938],{"d":8639},[21,8940,89],{"className":8941},[88],[21,8943,8945],{"className":8944},[60],[21,8946,8948],{"className":8947,"style":8649},[64],[21,8949],{},[21,8951,8953],{"className":8952},[42],[21,8954,8956,9013,9016,9019],{"className":8955},[8658],[21,8957,8959],{"className":8958},[8662],[21,8960,8962,9005],{"className":8961},[55,56],[21,8963,8965,9002],{"className":8964},[60],[21,8966,8968,8979,8990],{"className":8967,"style":8621},[64],[21,8969,8970,8973],{"style":8674},[21,8971],{"className":8972,"style":1678},[72],[21,8974,8976],{"className":8975},[42],[21,8977,116],{"className":8978},[42],[21,8980,8981,8984],{"style":8686},[21,8982],{"className":8983,"style":1678},[72],[21,8985,8987],{"className":8986},[42],[21,8988,1373],{"className":8989},[42],[21,8991,8992,8995],{"style":8698},[21,8993],{"className":8994,"style":1678},[72],[21,8996,8998],{"className":8997},[42],[21,8999,9001],{"className":9000},[42],"0.5",[21,9003,89],{"className":9004},[88],[21,9006,9008],{"className":9007},[60],[21,9009,9011],{"className":9010,"style":8649},[64],[21,9012],{},[21,9014],{"className":9015,"style":8723},[8722],[21,9017],{"className":9018,"style":8723},[8722],[21,9020,9022],{"className":9021},[8662],[21,9023,9025,9067],{"className":9024},[55,56],[21,9026,9028,9064],{"className":9027},[60],[21,9029,9031,9042,9053],{"className":9030,"style":8621},[64],[21,9032,9033,9036],{"style":8674},[21,9034],{"className":9035,"style":1678},[72],[21,9037,9039],{"className":9038},[42],[21,9040,1373],{"className":9041},[42],[21,9043,9044,9047],{"style":8686},[21,9045],{"className":9046,"style":1678},[72],[21,9048,9050],{"className":9049},[42],[21,9051,116],{"className":9052},[42],[21,9054,9055,9058],{"style":8698},[21,9056],{"className":9057,"style":1678},[72],[21,9059,9061],{"className":9060},[42],[21,9062,9001],{"className":9063},[42],[21,9065,89],{"className":9066},[88],[21,9068,9070],{"className":9069},[60],[21,9071,9073],{"className":9072,"style":8649},[64],[21,9074],{},[21,9076,9078],{"className":9077},[273],[21,9079,9081],{"className":9080},[3285,8611],[21,9082,9084,9104],{"className":9083},[55,56],[21,9085,9087,9101],{"className":9086},[60],[21,9088,9090],{"className":9089,"style":8621},[64],[21,9091,9092,9095],{"style":8624},[21,9093],{"className":9094,"style":8628},[72],[21,9096,9097],{"style":8631},[2628,9098,9099],{"xmlns":2630,"width":8634,"height":8635,"viewBox":8636},[2636,9100],{"d":8809},[21,9102,89],{"className":9103},[88],[21,9105,9107],{"className":9106},[60],[21,9108,9110],{"className":9109,"style":8649},[64],[21,9111],{},[21,9113],{"className":9114,"style":686},[401],[21,9116],{"className":9117,"style":686},[401],[21,9119,3730],{"className":9120},[690],[21,9122],{"className":9123,"style":686},[401],[21,9125],{"className":9126,"style":686},[401],[21,9128,9130,9133,9171],{"className":9129},[33],[21,9131],{"className":9132,"style":8861},[37],[21,9134,9136,9139],{"className":9135},[42],[21,9137,773],{"className":9138},[42,772],[21,9140,9142],{"className":9141},[51],[21,9143,9145],{"className":9144},[55],[21,9146,9148],{"className":9147},[60],[21,9149,9151],{"className":9150,"style":8861},[64],[21,9152,9153,9156],{"style":3127},[21,9154],{"className":9155,"style":73},[72],[21,9157,9159],{"className":9158},[77,78,79,80],[21,9160,9162,9165,9168],{"className":9161},[42,80],[21,9163,8876],{"className":9164},[42,80],[21,9166,2550],{"className":9167},[1218,80],[21,9169,5228],{"className":9170},[42,80],[21,9172,2126],{"className":9173},[42],[11,9175,9176,9243,9244,9351,9352,9367,9368,9398,9399,9480,9481,9562,9563,9644],{},[158,9177,9178,9179,2126],{},"Step 1 — raw scores ",[21,9180,9182],{"className":9181},[24],[21,9183,9185,9205],{"className":9184,"ariaHidden":29},[28],[21,9186,9188,9191,9196,9199,9202],{"className":9187},[33],[21,9189],{"className":9190,"style":2200},[37],[21,9192,9195],{"className":9193,"style":9194},[42,46],"margin-right:0.0576em;","S",[21,9197],{"className":9198,"style":686},[401],[21,9200,691],{"className":9201},[690],[21,9203],{"className":9204,"style":686},[401],[21,9206,9208,9211,9214],{"className":9207},[33],[21,9209],{"className":9210,"style":3691},[37],[21,9212,739],{"className":9213},[42,46],[21,9215,9217,9220],{"className":9216},[42],[21,9218,920],{"className":9219,"style":919},[42,46],[21,9221,9223],{"className":9222},[51],[21,9224,9226],{"className":9225},[55],[21,9227,9229],{"className":9228},[60],[21,9230,9232],{"className":9231,"style":765},[64],[21,9233,9234,9237],{"style":788},[21,9235],{"className":9236,"style":73},[72],[21,9238,9240],{"className":9239},[77,78,79,80],[21,9241,3137],{"className":9242},[42,80]," Each entry is ",[21,9245,9247],{"className":9246},[24],[21,9248,9250,9305],{"className":9249,"ariaHidden":29},[28],[21,9251,9253,9256,9296,9299,9302],{"className":9252},[33],[21,9254],{"className":9255,"style":1208},[37],[21,9257,9259,9262],{"className":9258},[42],[21,9260,203],{"className":9261,"style":202},[42,46],[21,9263,9265],{"className":9264},[51],[21,9266,9268,9288],{"className":9267},[55,56],[21,9269,9271,9285],{"className":9270},[60],[21,9272,9274],{"className":9273,"style":243},[64],[21,9275,9276,9279],{"style":312},[21,9277],{"className":9278,"style":73},[72],[21,9280,9282],{"className":9281},[77,78,79,80],[21,9283,3798],{"className":9284},[42,46,80],[21,9286,89],{"className":9287},[88],[21,9289,9291],{"className":9290},[60],[21,9292,9294],{"className":9293,"style":96},[64],[21,9295],{},[21,9297],{"className":9298,"style":1087},[401],[21,9300,1219],{"className":9301},[1218],[21,9303],{"className":9304,"style":1087},[401],[21,9306,9308,9311],{"className":9307},[33],[21,9309],{"className":9310,"style":569},[37],[21,9312,9314,9317],{"className":9313},[42],[21,9315,230],{"className":9316,"style":229},[42,46],[21,9318,9320],{"className":9319},[51],[21,9321,9323,9343],{"className":9322},[55,56],[21,9324,9326,9340],{"className":9325},[60],[21,9327,9329],{"className":9328,"style":243},[64],[21,9330,9331,9334],{"style":246},[21,9332],{"className":9333,"style":73},[72],[21,9335,9337],{"className":9336},[77,78,79,80],[21,9338,257],{"className":9339,"style":256},[42,46,80],[21,9341,89],{"className":9342},[88],[21,9344,9346],{"className":9345},[60],[21,9347,9349],{"className":9348,"style":267},[64],[21,9350],{},". Row ",[21,9353,9355],{"className":9354},[24],[21,9356,9358],{"className":9357,"ariaHidden":29},[28],[21,9359,9361,9364],{"className":9360},[33],[21,9362],{"className":9363,"style":112},[37],[21,9365,116],{"className":9366},[42]," is\nquery ",[21,9369,9371],{"className":9370},[24],[21,9372,9374],{"className":9373,"ariaHidden":29},[28],[21,9375,9377,9380,9383,9386,9389,9392,9395],{"className":9376},[33],[21,9378],{"className":9379,"style":1365},[37],[21,9381,1369],{"className":9382},[221],[21,9384,116],{"className":9385},[42],[21,9387,1378],{"className":9388},[1377],[21,9390],{"className":9391,"style":402},[401],[21,9393,1373],{"className":9394},[42],[21,9396,1388],{"className":9397},[273]," dotted against the three keys: ",[21,9400,9402],{"className":9401},[24],[21,9403,9405,9438,9471],{"className":9404,"ariaHidden":29},[28],[21,9406,9408,9411,9414,9417,9420,9423,9426,9429,9432,9435],{"className":9407},[33],[21,9409],{"className":9410,"style":1365},[37],[21,9412,1369],{"className":9413},[221],[21,9415,116],{"className":9416},[42],[21,9418,1378],{"className":9419},[1377],[21,9421],{"className":9422,"style":402},[401],[21,9424,1373],{"className":9425},[42],[21,9427,1388],{"className":9428},[273],[21,9430],{"className":9431,"style":1087},[401],[21,9433,1219],{"className":9434},[1218],[21,9436],{"className":9437,"style":1087},[401],[21,9439,9441,9444,9447,9450,9453,9456,9459,9462,9465,9468],{"className":9440},[33],[21,9442],{"className":9443,"style":1365},[37],[21,9445,1369],{"className":9446},[221],[21,9448,116],{"className":9449},[42],[21,9451,1378],{"className":9452},[1377],[21,9454],{"className":9455,"style":402},[401],[21,9457,1373],{"className":9458},[42],[21,9460,1388],{"className":9461},[273],[21,9463],{"className":9464,"style":686},[401],[21,9466,691],{"className":9467},[690],[21,9469],{"className":9470,"style":686},[401],[21,9472,9474,9477],{"className":9473},[33],[21,9475],{"className":9476,"style":112},[37],[21,9478,116],{"className":9479},[42],",\n",[21,9482,9484],{"className":9483},[24],[21,9485,9487,9520,9553],{"className":9486,"ariaHidden":29},[28],[21,9488,9490,9493,9496,9499,9502,9505,9508,9511,9514,9517],{"className":9489},[33],[21,9491],{"className":9492,"style":1365},[37],[21,9494,1369],{"className":9495},[221],[21,9497,116],{"className":9498},[42],[21,9500,1378],{"className":9501},[1377],[21,9503],{"className":9504,"style":402},[401],[21,9506,1373],{"className":9507},[42],[21,9509,1388],{"className":9510},[273],[21,9512],{"className":9513,"style":1087},[401],[21,9515,1219],{"className":9516},[1218],[21,9518],{"className":9519,"style":1087},[401],[21,9521,9523,9526,9529,9532,9535,9538,9541,9544,9547,9550],{"className":9522},[33],[21,9524],{"className":9525,"style":1365},[37],[21,9527,1369],{"className":9528},[221],[21,9530,1373],{"className":9531},[42],[21,9533,1378],{"className":9534},[1377],[21,9536],{"className":9537,"style":402},[401],[21,9539,116],{"className":9540},[42],[21,9542,1388],{"className":9543},[273],[21,9545],{"className":9546,"style":686},[401],[21,9548,691],{"className":9549},[690],[21,9551],{"className":9552,"style":686},[401],[21,9554,9556,9559],{"className":9555},[33],[21,9557],{"className":9558,"style":112},[37],[21,9560,1373],{"className":9561},[42],", ",[21,9564,9566],{"className":9565},[24],[21,9567,9569,9602,9635],{"className":9568,"ariaHidden":29},[28],[21,9570,9572,9575,9578,9581,9584,9587,9590,9593,9596,9599],{"className":9571},[33],[21,9573],{"className":9574,"style":1365},[37],[21,9576,1369],{"className":9577},[221],[21,9579,116],{"className":9580},[42],[21,9582,1378],{"className":9583},[1377],[21,9585],{"className":9586,"style":402},[401],[21,9588,1373],{"className":9589},[42],[21,9591,1388],{"className":9592},[273],[21,9594],{"className":9595,"style":1087},[401],[21,9597,1219],{"className":9598},[1218],[21,9600],{"className":9601,"style":1087},[401],[21,9603,9605,9608,9611,9614,9617,9620,9623,9626,9629,9632],{"className":9604},[33],[21,9606],{"className":9607,"style":1365},[37],[21,9609,1369],{"className":9610},[221],[21,9612,116],{"className":9613},[42],[21,9615,1378],{"className":9616},[1377],[21,9618],{"className":9619,"style":402},[401],[21,9621,116],{"className":9622},[42],[21,9624,1388],{"className":9625},[273],[21,9627],{"className":9628,"style":686},[401],[21,9630,691],{"className":9631},[690],[21,9633],{"className":9634,"style":686},[401],[21,9636,9638,9641],{"className":9637},[33],[21,9639],{"className":9640,"style":112},[37],[21,9642,116],{"className":9643},[42],". Filling in all nine,",[21,9646,9648],{"className":9647},[1582],[21,9649,9651],{"className":9650},[24],[21,9652,9654,9672,9720,10004],{"className":9653,"ariaHidden":29},[28],[21,9655,9657,9660,9663,9666,9669],{"className":9656},[33],[21,9658],{"className":9659,"style":2200},[37],[21,9661,9195],{"className":9662,"style":9194},[42,46],[21,9664],{"className":9665,"style":686},[401],[21,9667,691],{"className":9668},[690],[21,9670],{"className":9671,"style":686},[401],[21,9673,9675,9679,9682,9711,9714,9717],{"className":9674},[33],[21,9676],{"className":9677,"style":9678},[37],"height:1.0935em;vertical-align:-0.1944em;",[21,9680,739],{"className":9681},[42,46],[21,9683,9685,9688],{"className":9684},[42],[21,9686,920],{"className":9687,"style":919},[42,46],[21,9689,9691],{"className":9690},[51],[21,9692,9694],{"className":9693},[55],[21,9695,9697],{"className":9696},[60],[21,9698,9700],{"className":9699,"style":3027},[64],[21,9701,9702,9705],{"style":3127},[21,9703],{"className":9704,"style":73},[72],[21,9706,9708],{"className":9707},[77,78,79,80],[21,9709,3137],{"className":9710},[42,80],[21,9712],{"className":9713,"style":686},[401],[21,9715,691],{"className":9716},[690],[21,9718],{"className":9719,"style":686},[401],[21,9721,9723,9726,9989,9992,9995,9998,10001],{"className":9722},[33],[21,9724],{"className":9725,"style":8601},[37],[21,9727,9729,9766,9952],{"className":9728},[2521],[21,9730,9732],{"className":9731},[221],[21,9733,9735],{"className":9734},[3285,8611],[21,9736,9738,9758],{"className":9737},[55,56],[21,9739,9741,9755],{"className":9740},[60],[21,9742,9744],{"className":9743,"style":8621},[64],[21,9745,9746,9749],{"style":8624},[21,9747],{"className":9748,"style":8628},[72],[21,9750,9751],{"style":8631},[2628,9752,9753],{"xmlns":2630,"width":8634,"height":8635,"viewBox":8636},[2636,9754],{"d":8639},[21,9756,89],{"className":9757},[88],[21,9759,9761],{"className":9760},[60],[21,9762,9764],{"className":9763,"style":8649},[64],[21,9765],{},[21,9767,9769],{"className":9768},[42],[21,9770,9772,9828,9831,9834,9890,9893,9896],{"className":9771},[8658],[21,9773,9775],{"className":9774},[8662],[21,9776,9778,9820],{"className":9777},[55,56],[21,9779,9781,9817],{"className":9780},[60],[21,9782,9784,9795,9806],{"className":9783,"style":8621},[64],[21,9785,9786,9789],{"style":8674},[21,9787],{"className":9788,"style":1678},[72],[21,9790,9792],{"className":9791},[42],[21,9793,116],{"className":9794},[42],[21,9796,9797,9800],{"style":8686},[21,9798],{"className":9799,"style":1678},[72],[21,9801,9803],{"className":9802},[42],[21,9804,1373],{"className":9805},[42],[21,9807,9808,9811],{"style":8698},[21,9809],{"className":9810,"style":1678},[72],[21,9812,9814],{"className":9813},[42],[21,9815,116],{"className":9816},[42],[21,9818,89],{"className":9819},[88],[21,9821,9823],{"className":9822},[60],[21,9824,9826],{"className":9825,"style":8649},[64],[21,9827],{},[21,9829],{"className":9830,"style":8723},[8722],[21,9832],{"className":9833,"style":8723},[8722],[21,9835,9837],{"className":9836},[8662],[21,9838,9840,9882],{"className":9839},[55,56],[21,9841,9843,9879],{"className":9842},[60],[21,9844,9846,9857,9868],{"className":9845,"style":8621},[64],[21,9847,9848,9851],{"style":8674},[21,9849],{"className":9850,"style":1678},[72],[21,9852,9854],{"className":9853},[42],[21,9855,1373],{"className":9856},[42],[21,9858,9859,9862],{"style":8686},[21,9860],{"className":9861,"style":1678},[72],[21,9863,9865],{"className":9864},[42],[21,9866,116],{"className":9867},[42],[21,9869,9870,9873],{"style":8698},[21,9871],{"className":9872,"style":1678},[72],[21,9874,9876],{"className":9875},[42],[21,9877,116],{"className":9878},[42],[21,9880,89],{"className":9881},[88],[21,9883,9885],{"className":9884},[60],[21,9886,9888],{"className":9887,"style":8649},[64],[21,9889],{},[21,9891],{"className":9892,"style":8723},[8722],[21,9894],{"className":9895,"style":8723},[8722],[21,9897,9899],{"className":9898},[8662],[21,9900,9902,9944],{"className":9901},[55,56],[21,9903,9905,9941],{"className":9904},[60],[21,9906,9908,9919,9930],{"className":9907,"style":8621},[64],[21,9909,9910,9913],{"style":8674},[21,9911],{"className":9912,"style":1678},[72],[21,9914,9916],{"className":9915},[42],[21,9917,116],{"className":9918},[42],[21,9920,9921,9924],{"style":8686},[21,9922],{"className":9923,"style":1678},[72],[21,9925,9927],{"className":9926},[42],[21,9928,116],{"className":9929},[42],[21,9931,9932,9935],{"style":8698},[21,9933],{"className":9934,"style":1678},[72],[21,9936,9938],{"className":9937},[42],[21,9939,5228],{"className":9940},[42],[21,9942,89],{"className":9943},[88],[21,9945,9947],{"className":9946},[60],[21,9948,9950],{"className":9949,"style":8649},[64],[21,9951],{},[21,9953,9955],{"className":9954},[273],[21,9956,9958],{"className":9957},[3285,8611],[21,9959,9961,9981],{"className":9960},[55,56],[21,9962,9964,9978],{"className":9963},[60],[21,9965,9967],{"className":9966,"style":8621},[64],[21,9968,9969,9972],{"style":8624},[21,9970],{"className":9971,"style":8628},[72],[21,9973,9974],{"style":8631},[2628,9975,9976],{"xmlns":2630,"width":8634,"height":8635,"viewBox":8636},[2636,9977],{"d":8809},[21,9979,89],{"className":9980},[88],[21,9982,9984],{"className":9983},[60],[21,9985,9987],{"className":9986,"style":8649},[64],[21,9988],{},[21,9990],{"className":9991,"style":686},[401],[21,9993],{"className":9994,"style":686},[401],[21,9996,3730],{"className":9997},[690],[21,9999],{"className":10000,"style":686},[401],[21,10002],{"className":10003,"style":686},[401],[21,10005,10007,10010,10048],{"className":10006},[33],[21,10008],{"className":10009,"style":8861},[37],[21,10011,10013,10016],{"className":10012},[42],[21,10014,773],{"className":10015},[42,772],[21,10017,10019],{"className":10018},[51],[21,10020,10022],{"className":10021},[55],[21,10023,10025],{"className":10024},[60],[21,10026,10028],{"className":10027,"style":8861},[64],[21,10029,10030,10033],{"style":3127},[21,10031],{"className":10032,"style":73},[72],[21,10034,10036],{"className":10035},[77,78,79,80],[21,10037,10039,10042,10045],{"className":10038},[42,80],[21,10040,8876],{"className":10041},[42,80],[21,10043,2550],{"className":10044},[1218,80],[21,10046,8876],{"className":10047},[42,80],[21,10049,2126],{"className":10050},[42],[11,10052,10053,10235],{},[158,10054,10055,10056,2126],{},"Step 2 — scale by ",[21,10057,10059],{"className":10058},[24],[21,10060,10062,10161,10225],{"className":10061,"ariaHidden":29},[28],[21,10063,10065,10068,10152,10155,10158],{"className":10064},[33],[21,10066],{"className":10067,"style":3965},[37],[21,10069,10071],{"className":10070},[42,3321],[21,10072,10074,10144],{"className":10073},[55,56],[21,10075,10077,10141],{"className":10076},[60],[21,10078,10080,10129],{"className":10079,"style":3978},[64],[21,10081,10083,10086],{"className":10082,"style":2593},[2612],[21,10084],{"className":10085,"style":1678},[72],[21,10087,10089],{"className":10088,"style":3341},[42],[21,10090,10092,10095],{"className":10091},[42],[21,10093,804],{"className":10094},[42,46],[21,10096,10098],{"className":10097},[51],[21,10099,10101,10121],{"className":10100},[55,56],[21,10102,10104,10118],{"className":10103},[60],[21,10105,10107],{"className":10106,"style":2424},[64],[21,10108,10109,10112],{"style":68},[21,10110],{"className":10111,"style":73},[72],[21,10113,10115],{"className":10114},[77,78,79,80],[21,10116,230],{"className":10117,"style":229},[42,46,80],[21,10119,89],{"className":10120},[88],[21,10122,10124],{"className":10123},[60],[21,10125,10127],{"className":10126,"style":96},[64],[21,10128],{},[21,10130,10131,10134],{"style":4030},[21,10132],{"className":10133,"style":1678},[72],[21,10135,10137],{"className":10136,"style":3392},[3391],[2628,10138,10139],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,10140],{"d":3399},[21,10142,89],{"className":10143},[88],[21,10145,10147],{"className":10146},[60],[21,10148,10150],{"className":10149,"style":4050},[64],[21,10151],{},[21,10153],{"className":10154,"style":686},[401],[21,10156,691],{"className":10157},[690],[21,10159],{"className":10160,"style":686},[401],[21,10162,10164,10168,10215,10218,10222],{"className":10163},[33],[21,10165],{"className":10166,"style":10167},[37],"height:1.04em;vertical-align:-0.1328em;",[21,10169,10171],{"className":10170},[42,3321],[21,10172,10174,10206],{"className":10173},[55,56],[21,10175,10177,10203],{"className":10176},[60],[21,10178,10181,10190],{"className":10179,"style":10180},[64],"height:0.9072em;",[21,10182,10184,10187],{"className":10183,"style":2593},[2612],[21,10185],{"className":10186,"style":1678},[72],[21,10188,5228],{"className":10189,"style":3341},[42],[21,10191,10193,10196],{"style":10192},"top:-2.8672em;",[21,10194],{"className":10195,"style":1678},[72],[21,10197,10199],{"className":10198,"style":3392},[3391],[2628,10200,10201],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,10202],{"d":3399},[21,10204,89],{"className":10205},[88],[21,10207,10209],{"className":10208},[60],[21,10210,10213],{"className":10211,"style":10212},[64],"height:0.1328em;",[21,10214],{},[21,10216],{"className":10217,"style":686},[401],[21,10219,10221],{"className":10220},[690],"≈",[21,10223],{"className":10224,"style":686},[401],[21,10226,10228,10231],{"className":10227},[33],[21,10229],{"className":10230,"style":112},[37],[21,10232,10234],{"className":10233},[42],"1.414"," Every score shrinks by\nthe same factor, pulling the logits toward zero before the softmax:",[21,10237,10239],{"className":10238},[1582],[21,10240,10242],{"className":10241},[24],[21,10243,10245,10366],{"className":10244,"ariaHidden":29},[28],[21,10246,10248,10252,10357,10360,10363],{"className":10247},[33],[21,10249],{"className":10250,"style":10251},[37],"height:2.2903em;vertical-align:-0.93em;",[21,10253,10255,10258,10354],{"className":10254},[42],[21,10256],{"className":10257},[221,1657],[21,10259,10261],{"className":10260},[1661],[21,10262,10264,10346],{"className":10263},[55,56],[21,10265,10267,10343],{"className":10266},[60],[21,10268,10271,10324,10332],{"className":10269,"style":10270},[64],"height:1.3603em;",[21,10272,10274,10277],{"style":10273},"top:-2.2028em;",[21,10275],{"className":10276,"style":1678},[72],[21,10278,10280],{"className":10279},[42],[21,10281,10283],{"className":10282},[42,3321],[21,10284,10286,10316],{"className":10285},[55,56],[21,10287,10289,10313],{"className":10288},[60],[21,10290,10292,10301],{"className":10291,"style":10180},[64],[21,10293,10295,10298],{"className":10294,"style":2593},[2612],[21,10296],{"className":10297,"style":1678},[72],[21,10299,5228],{"className":10300,"style":3341},[42],[21,10302,10303,10306],{"style":10192},[21,10304],{"className":10305,"style":1678},[72],[21,10307,10309],{"className":10308,"style":3392},[3391],[2628,10310,10311],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,10312],{"d":3399},[21,10314,89],{"className":10315},[88],[21,10317,10319],{"className":10318},[60],[21,10320,10322],{"className":10321,"style":10212},[64],[21,10323],{},[21,10325,10326,10329],{"style":1858},[21,10327],{"className":10328,"style":1678},[72],[21,10330],{"className":10331,"style":1866},[1865],[21,10333,10334,10337],{"style":1869},[21,10335],{"className":10336,"style":1678},[72],[21,10338,10340],{"className":10339},[42],[21,10341,9195],{"className":10342,"style":9194},[42,46],[21,10344,89],{"className":10345},[88],[21,10347,10349],{"className":10348},[60],[21,10350,10352],{"className":10351,"style":4948},[64],[21,10353],{},[21,10355],{"className":10356},[273,1657],[21,10358],{"className":10359,"style":686},[401],[21,10361,691],{"className":10362},[690],[21,10364],{"className":10365,"style":686},[401],[21,10367,10369,10372,10636,10639],{"className":10368},[33],[21,10370],{"className":10371,"style":8601},[37],[21,10373,10375,10412,10599],{"className":10374},[2521],[21,10376,10378],{"className":10377},[221],[21,10379,10381],{"className":10380},[3285,8611],[21,10382,10384,10404],{"className":10383},[55,56],[21,10385,10387,10401],{"className":10386},[60],[21,10388,10390],{"className":10389,"style":8621},[64],[21,10391,10392,10395],{"style":8624},[21,10393],{"className":10394,"style":8628},[72],[21,10396,10397],{"style":8631},[2628,10398,10399],{"xmlns":2630,"width":8634,"height":8635,"viewBox":8636},[2636,10400],{"d":8639},[21,10402,89],{"className":10403},[88],[21,10405,10407],{"className":10406},[60],[21,10408,10410],{"className":10409,"style":8649},[64],[21,10411],{},[21,10413,10415],{"className":10414},[42],[21,10416,10418,10475,10478,10481,10537,10540,10543],{"className":10417},[8658],[21,10419,10421],{"className":10420},[8662],[21,10422,10424,10467],{"className":10423},[55,56],[21,10425,10427,10464],{"className":10426},[60],[21,10428,10430,10442,10453],{"className":10429,"style":8621},[64],[21,10431,10432,10435],{"style":8674},[21,10433],{"className":10434,"style":1678},[72],[21,10436,10438],{"className":10437},[42],[21,10439,10441],{"className":10440},[42],"0.707",[21,10443,10444,10447],{"style":8686},[21,10445],{"className":10446,"style":1678},[72],[21,10448,10450],{"className":10449},[42],[21,10451,1373],{"className":10452},[42],[21,10454,10455,10458],{"style":8698},[21,10456],{"className":10457,"style":1678},[72],[21,10459,10461],{"className":10460},[42],[21,10462,10441],{"className":10463},[42],[21,10465,89],{"className":10466},[88],[21,10468,10470],{"className":10469},[60],[21,10471,10473],{"className":10472,"style":8649},[64],[21,10474],{},[21,10476],{"className":10477,"style":8723},[8722],[21,10479],{"className":10480,"style":8723},[8722],[21,10482,10484],{"className":10483},[8662],[21,10485,10487,10529],{"className":10486},[55,56],[21,10488,10490,10526],{"className":10489},[60],[21,10491,10493,10504,10515],{"className":10492,"style":8621},[64],[21,10494,10495,10498],{"style":8674},[21,10496],{"className":10497,"style":1678},[72],[21,10499,10501],{"className":10500},[42],[21,10502,1373],{"className":10503},[42],[21,10505,10506,10509],{"style":8686},[21,10507],{"className":10508,"style":1678},[72],[21,10510,10512],{"className":10511},[42],[21,10513,10441],{"className":10514},[42],[21,10516,10517,10520],{"style":8698},[21,10518],{"className":10519,"style":1678},[72],[21,10521,10523],{"className":10522},[42],[21,10524,10441],{"className":10525},[42],[21,10527,89],{"className":10528},[88],[21,10530,10532],{"className":10531},[60],[21,10533,10535],{"className":10534,"style":8649},[64],[21,10536],{},[21,10538],{"className":10539,"style":8723},[8722],[21,10541],{"className":10542,"style":8723},[8722],[21,10544,10546],{"className":10545},[8662],[21,10547,10549,10591],{"className":10548},[55,56],[21,10550,10552,10588],{"className":10551},[60],[21,10553,10555,10566,10577],{"className":10554,"style":8621},[64],[21,10556,10557,10560],{"style":8674},[21,10558],{"className":10559,"style":1678},[72],[21,10561,10563],{"className":10562},[42],[21,10564,10441],{"className":10565},[42],[21,10567,10568,10571],{"style":8686},[21,10569],{"className":10570,"style":1678},[72],[21,10572,10574],{"className":10573},[42],[21,10575,10441],{"className":10576},[42],[21,10578,10579,10582],{"style":8698},[21,10580],{"className":10581,"style":1678},[72],[21,10583,10585],{"className":10584},[42],[21,10586,10234],{"className":10587},[42],[21,10589,89],{"className":10590},[88],[21,10592,10594],{"className":10593},[60],[21,10595,10597],{"className":10596,"style":8649},[64],[21,10598],{},[21,10600,10602],{"className":10601},[273],[21,10603,10605],{"className":10604},[3285,8611],[21,10606,10608,10628],{"className":10607},[55,56],[21,10609,10611,10625],{"className":10610},[60],[21,10612,10614],{"className":10613,"style":8621},[64],[21,10615,10616,10619],{"style":8624},[21,10617],{"className":10618,"style":8628},[72],[21,10620,10621],{"style":8631},[2628,10622,10623],{"xmlns":2630,"width":8634,"height":8635,"viewBox":8636},[2636,10624],{"d":8809},[21,10626,89],{"className":10627},[88],[21,10629,10631],{"className":10630},[60],[21,10632,10634],{"className":10633,"style":8649},[64],[21,10635],{},[21,10637],{"className":10638,"style":402},[401],[21,10640,2126],{"className":10641},[42],[11,10643,10644,10647,10648,10663,10664,9562,10728,9562,10790,10852,10853,10869,10870,10905,10906,10941],{},[158,10645,10646],{},"Step 3 — row-wise softmax."," Work row ",[21,10649,10651],{"className":10650},[24],[21,10652,10654],{"className":10653,"ariaHidden":29},[28],[21,10655,10657,10660],{"className":10656},[33],[21,10658],{"className":10659,"style":112},[37],[21,10661,116],{"className":10662},[42]," explicitly. Exponentiate: ",[21,10665,10667],{"className":10666},[24],[21,10668,10670,10718],{"className":10669,"ariaHidden":29},[28],[21,10671,10673,10676,10709,10712,10715],{"className":10672},[33],[21,10674],{"className":10675,"style":7197},[37],[21,10677,10679,10683],{"className":10678},[42],[21,10680,10682],{"className":10681},[42,46],"e",[21,10684,10686],{"className":10685},[51],[21,10687,10689],{"className":10688},[55],[21,10690,10692],{"className":10691},[60],[21,10693,10695],{"className":10694,"style":7197},[64],[21,10696,10697,10700],{"style":788},[21,10698],{"className":10699,"style":73},[72],[21,10701,10703],{"className":10702},[77,78,79,80],[21,10704,10706],{"className":10705},[42,80],[21,10707,10441],{"className":10708},[42,80],[21,10710],{"className":10711,"style":686},[401],[21,10713,10221],{"className":10714},[690],[21,10716],{"className":10717,"style":686},[401],[21,10719,10721,10724],{"className":10720},[33],[21,10722],{"className":10723,"style":112},[37],[21,10725,10727],{"className":10726},[42],"2.028",[21,10729,10731],{"className":10730},[24],[21,10732,10734,10781],{"className":10733,"ariaHidden":29},[28],[21,10735,10737,10740,10772,10775,10778],{"className":10736},[33],[21,10738],{"className":10739,"style":7197},[37],[21,10741,10743,10746],{"className":10742},[42],[21,10744,10682],{"className":10745},[42,46],[21,10747,10749],{"className":10748},[51],[21,10750,10752],{"className":10751},[55],[21,10753,10755],{"className":10754},[60],[21,10756,10758],{"className":10757,"style":7197},[64],[21,10759,10760,10763],{"style":788},[21,10761],{"className":10762,"style":73},[72],[21,10764,10766],{"className":10765},[77,78,79,80],[21,10767,10769],{"className":10768},[42,80],[21,10770,1373],{"className":10771},[42,80],[21,10773],{"className":10774,"style":686},[401],[21,10776,691],{"className":10777},[690],[21,10779],{"className":10780,"style":686},[401],[21,10782,10784,10787],{"className":10783},[33],[21,10785],{"className":10786,"style":112},[37],[21,10788,116],{"className":10789},[42],[21,10791,10793],{"className":10792},[24],[21,10794,10796,10843],{"className":10795,"ariaHidden":29},[28],[21,10797,10799,10802,10834,10837,10840],{"className":10798},[33],[21,10800],{"className":10801,"style":7197},[37],[21,10803,10805,10808],{"className":10804},[42],[21,10806,10682],{"className":10807},[42,46],[21,10809,10811],{"className":10810},[51],[21,10812,10814],{"className":10813},[55],[21,10815,10817],{"className":10816},[60],[21,10818,10820],{"className":10819,"style":7197},[64],[21,10821,10822,10825],{"style":788},[21,10823],{"className":10824,"style":73},[72],[21,10826,10828],{"className":10827},[77,78,79,80],[21,10829,10831],{"className":10830},[42,80],[21,10832,10441],{"className":10833},[42,80],[21,10835],{"className":10836,"style":686},[401],[21,10838,10221],{"className":10839},[690],[21,10841],{"className":10842,"style":686},[401],[21,10844,10846,10849],{"className":10845},[33],[21,10847],{"className":10848,"style":112},[37],[21,10850,10727],{"className":10851},[42],", summing to ",[21,10854,10856],{"className":10855},[24],[21,10857,10859],{"className":10858,"ariaHidden":29},[28],[21,10860,10862,10865],{"className":10861},[33],[21,10863],{"className":10864,"style":112},[37],[21,10866,10868],{"className":10867},[42],"5.056",".\nDividing gives ",[21,10871,10873],{"className":10872},[24],[21,10874,10876,10895],{"className":10875,"ariaHidden":29},[28],[21,10877,10879,10882,10886,10889,10892],{"className":10878},[33],[21,10880],{"className":10881,"style":1365},[37],[21,10883,10885],{"className":10884},[42],"2.028\u002F5.056",[21,10887],{"className":10888,"style":686},[401],[21,10890,10221],{"className":10891},[690],[21,10893],{"className":10894,"style":686},[401],[21,10896,10898,10901],{"className":10897},[33],[21,10899],{"className":10900,"style":112},[37],[21,10902,10904],{"className":10903},[42],"0.401"," for the first and third keys and\n",[21,10907,10909],{"className":10908},[24],[21,10910,10912,10931],{"className":10911,"ariaHidden":29},[28],[21,10913,10915,10918,10922,10925,10928],{"className":10914},[33],[21,10916],{"className":10917,"style":1365},[37],[21,10919,10921],{"className":10920},[42],"1\u002F5.056",[21,10923],{"className":10924,"style":686},[401],[21,10926,10221],{"className":10927},[690],[21,10929],{"className":10930,"style":686},[401],[21,10932,10934,10937],{"className":10933},[33],[21,10935],{"className":10936,"style":112},[37],[21,10938,10940],{"className":10939},[42],"0.198"," for the second. Doing the same to every row,",[21,10943,10945],{"className":10944},[1582],[21,10946,10948],{"className":10947},[24],[21,10949,10951,10969,11115],{"className":10950,"ariaHidden":29},[28],[21,10952,10954,10957,10960,10963,10966],{"className":10953},[33],[21,10955],{"className":10956,"style":2200},[37],[21,10958,3644],{"className":10959},[42,46],[21,10961],{"className":10962,"style":686},[401],[21,10964,691],{"className":10965},[690],[21,10967],{"className":10968,"style":686},[401],[21,10970,10972,10976,10982,10985,10988,11106,11109,11112],{"className":10971},[33],[21,10973],{"className":10974,"style":10975},[37],"height:2.4em;vertical-align:-0.95em;",[21,10977,10979],{"className":10978},[353],[21,10980,3266],{"className":10981},[42,3265],[21,10983],{"className":10984,"style":3270},[401],[21,10986],{"className":10987,"style":402},[401],[21,10989,10991,10997,11100],{"className":10990},[2521],[21,10992,10994],{"className":10993,"style":3281},[221,3280],[21,10995,1369],{"className":10996},[3285,79],[21,10998,11000,11003,11097],{"className":10999},[42],[21,11001],{"className":11002},[221,1657],[21,11004,11006],{"className":11005},[1661],[21,11007,11009,11089],{"className":11008},[55,56],[21,11010,11012,11086],{"className":11011},[60],[21,11013,11015,11067,11075],{"className":11014,"style":10270},[64],[21,11016,11017,11020],{"style":10273},[21,11018],{"className":11019,"style":1678},[72],[21,11021,11023],{"className":11022},[42],[21,11024,11026],{"className":11025},[42,3321],[21,11027,11029,11059],{"className":11028},[55,56],[21,11030,11032,11056],{"className":11031},[60],[21,11033,11035,11044],{"className":11034,"style":10180},[64],[21,11036,11038,11041],{"className":11037,"style":2593},[2612],[21,11039],{"className":11040,"style":1678},[72],[21,11042,5228],{"className":11043,"style":3341},[42],[21,11045,11046,11049],{"style":10192},[21,11047],{"className":11048,"style":1678},[72],[21,11050,11052],{"className":11051,"style":3392},[3391],[2628,11053,11054],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,11055],{"d":3399},[21,11057,89],{"className":11058},[88],[21,11060,11062],{"className":11061},[60],[21,11063,11065],{"className":11064,"style":10212},[64],[21,11066],{},[21,11068,11069,11072],{"style":1858},[21,11070],{"className":11071,"style":1678},[72],[21,11073],{"className":11074,"style":1866},[1865],[21,11076,11077,11080],{"style":1869},[21,11078],{"className":11079,"style":1678},[72],[21,11081,11083],{"className":11082},[42],[21,11084,9195],{"className":11085,"style":9194},[42,46],[21,11087,89],{"className":11088},[88],[21,11090,11092],{"className":11091},[60],[21,11093,11095],{"className":11094,"style":4948},[64],[21,11096],{},[21,11098],{"className":11099},[273,1657],[21,11101,11103],{"className":11102,"style":3281},[273,3280],[21,11104,1388],{"className":11105},[3285,79],[21,11107],{"className":11108,"style":686},[401],[21,11110,691],{"className":11111},[690],[21,11113],{"className":11114,"style":686},[401],[21,11116,11118,11121,11386,11389],{"className":11117},[33],[21,11119],{"className":11120,"style":8601},[37],[21,11122,11124,11161,11349],{"className":11123},[2521],[21,11125,11127],{"className":11126},[221],[21,11128,11130],{"className":11129},[3285,8611],[21,11131,11133,11153],{"className":11132},[55,56],[21,11134,11136,11150],{"className":11135},[60],[21,11137,11139],{"className":11138,"style":8621},[64],[21,11140,11141,11144],{"style":8624},[21,11142],{"className":11143,"style":8628},[72],[21,11145,11146],{"style":8631},[2628,11147,11148],{"xmlns":2630,"width":8634,"height":8635,"viewBox":8636},[2636,11149],{"d":8639},[21,11151,89],{"className":11152},[88],[21,11154,11156],{"className":11155},[60],[21,11157,11159],{"className":11158,"style":8649},[64],[21,11160],{},[21,11162,11164],{"className":11163},[42],[21,11165,11167,11224,11227,11230,11286,11289,11292],{"className":11166},[8658],[21,11168,11170],{"className":11169},[8662],[21,11171,11173,11216],{"className":11172},[55,56],[21,11174,11176,11213],{"className":11175},[60],[21,11177,11179,11190,11201],{"className":11178,"style":8621},[64],[21,11180,11181,11184],{"style":8674},[21,11182],{"className":11183,"style":1678},[72],[21,11185,11187],{"className":11186},[42],[21,11188,10904],{"className":11189},[42],[21,11191,11192,11195],{"style":8686},[21,11193],{"className":11194,"style":1678},[72],[21,11196,11198],{"className":11197},[42],[21,11199,10940],{"className":11200},[42],[21,11202,11203,11206],{"style":8698},[21,11204],{"className":11205,"style":1678},[72],[21,11207,11209],{"className":11208},[42],[21,11210,11212],{"className":11211},[42],"0.248",[21,11214,89],{"className":11215},[88],[21,11217,11219],{"className":11218},[60],[21,11220,11222],{"className":11221,"style":8649},[64],[21,11223],{},[21,11225],{"className":11226,"style":8723},[8722],[21,11228],{"className":11229,"style":8723},[8722],[21,11231,11233],{"className":11232},[8662],[21,11234,11236,11278],{"className":11235},[55,56],[21,11237,11239,11275],{"className":11238},[60],[21,11240,11242,11253,11264],{"className":11241,"style":8621},[64],[21,11243,11244,11247],{"style":8674},[21,11245],{"className":11246,"style":1678},[72],[21,11248,11250],{"className":11249},[42],[21,11251,10940],{"className":11252},[42],[21,11254,11255,11258],{"style":8686},[21,11256],{"className":11257,"style":1678},[72],[21,11259,11261],{"className":11260},[42],[21,11262,10904],{"className":11263},[42],[21,11265,11266,11269],{"style":8698},[21,11267],{"className":11268,"style":1678},[72],[21,11270,11272],{"className":11271},[42],[21,11273,11212],{"className":11274},[42],[21,11276,89],{"className":11277},[88],[21,11279,11281],{"className":11280},[60],[21,11282,11284],{"className":11283,"style":8649},[64],[21,11285],{},[21,11287],{"className":11288,"style":8723},[8722],[21,11290],{"className":11291,"style":8723},[8722],[21,11293,11295],{"className":11294},[8662],[21,11296,11298,11341],{"className":11297},[55,56],[21,11299,11301,11338],{"className":11300},[60],[21,11302,11304,11315,11326],{"className":11303,"style":8621},[64],[21,11305,11306,11309],{"style":8674},[21,11307],{"className":11308,"style":1678},[72],[21,11310,11312],{"className":11311},[42],[21,11313,10904],{"className":11314},[42],[21,11316,11317,11320],{"style":8686},[21,11318],{"className":11319,"style":1678},[72],[21,11321,11323],{"className":11322},[42],[21,11324,10904],{"className":11325},[42],[21,11327,11328,11331],{"style":8698},[21,11329],{"className":11330,"style":1678},[72],[21,11332,11334],{"className":11333},[42],[21,11335,11337],{"className":11336},[42],"0.503",[21,11339,89],{"className":11340},[88],[21,11342,11344],{"className":11343},[60],[21,11345,11347],{"className":11346,"style":8649},[64],[21,11348],{},[21,11350,11352],{"className":11351},[273],[21,11353,11355],{"className":11354},[3285,8611],[21,11356,11358,11378],{"className":11357},[55,56],[21,11359,11361,11375],{"className":11360},[60],[21,11362,11364],{"className":11363,"style":8621},[64],[21,11365,11366,11369],{"style":8624},[21,11367],{"className":11368,"style":8628},[72],[21,11370,11371],{"style":8631},[2628,11372,11373],{"xmlns":2630,"width":8634,"height":8635,"viewBox":8636},[2636,11374],{"d":8809},[21,11376,89],{"className":11377},[88],[21,11379,11381],{"className":11380},[60],[21,11382,11384],{"className":11383,"style":8649},[64],[21,11385],{},[21,11387],{"className":11388,"style":402},[401],[21,11390,1378],{"className":11391},[1377],[11,11393,11394,11395,11410,11411,11441,11442,11457,11458,11473,11474,11555],{},"and each row sums to ",[21,11396,11398],{"className":11397},[24],[21,11399,11401],{"className":11400,"ariaHidden":29},[28],[21,11402,11404,11407],{"className":11403},[33],[21,11405],{"className":11406,"style":112},[37],[21,11408,116],{"className":11409},[42],", as a distribution over the three keys must. Notice the\nthird token — whose query ",[21,11412,11414],{"className":11413},[24],[21,11415,11417],{"className":11416,"ariaHidden":29},[28],[21,11418,11420,11423,11426,11429,11432,11435,11438],{"className":11419},[33],[21,11421],{"className":11422,"style":1365},[37],[21,11424,1369],{"className":11425},[221],[21,11427,116],{"className":11428},[42],[21,11430,1378],{"className":11431},[1377],[21,11433],{"className":11434,"style":402},[401],[21,11436,116],{"className":11437},[42],[21,11439,1388],{"className":11440},[273]," aligns with every key — spreads its weight most\nevenly but still favours itself (key ",[21,11443,11445],{"className":11444},[24],[21,11446,11448],{"className":11447,"ariaHidden":29},[28],[21,11449,11451,11454],{"className":11450},[33],[21,11452],{"className":11453,"style":112},[37],[21,11455,8876],{"className":11456},[42],", weight ",[21,11459,11461],{"className":11460},[24],[21,11462,11464],{"className":11463,"ariaHidden":29},[28],[21,11465,11467,11470],{"className":11466},[33],[21,11468],{"className":11469,"style":112},[37],[21,11471,11337],{"className":11472},[42],"), because ",[21,11475,11477],{"className":11476},[24],[21,11478,11480,11513,11546],{"className":11479,"ariaHidden":29},[28],[21,11481,11483,11486,11489,11492,11495,11498,11501,11504,11507,11510],{"className":11482},[33],[21,11484],{"className":11485,"style":1365},[37],[21,11487,1369],{"className":11488},[221],[21,11490,116],{"className":11491},[42],[21,11493,1378],{"className":11494},[1377],[21,11496],{"className":11497,"style":402},[401],[21,11499,116],{"className":11500},[42],[21,11502,1388],{"className":11503},[273],[21,11505],{"className":11506,"style":1087},[401],[21,11508,1219],{"className":11509},[1218],[21,11511],{"className":11512,"style":1087},[401],[21,11514,11516,11519,11522,11525,11528,11531,11534,11537,11540,11543],{"className":11515},[33],[21,11517],{"className":11518,"style":1365},[37],[21,11520,1369],{"className":11521},[221],[21,11523,116],{"className":11524},[42],[21,11526,1378],{"className":11527},[1377],[21,11529],{"className":11530,"style":402},[401],[21,11532,116],{"className":11533},[42],[21,11535,1388],{"className":11536},[273],[21,11538],{"className":11539,"style":686},[401],[21,11541,691],{"className":11542},[690],[21,11544],{"className":11545,"style":686},[401],[21,11547,11549,11552],{"className":11548},[33],[21,11550],{"className":11551,"style":112},[37],[21,11553,5228],{"className":11554},[42]," is the largest score in that row.",[11,11557,11558,11599,11600,11615,11616,11941],{},[158,11559,11560,11561,2126],{},"Step 4 — blend the values, ",[21,11562,11564],{"className":11563},[24],[21,11565,11567,11587],{"className":11566,"ariaHidden":29},[28],[21,11568,11570,11573,11578,11581,11584],{"className":11569},[33],[21,11571],{"className":11572,"style":2200},[37],[21,11574,11577],{"className":11575,"style":11576},[42,46],"margin-right:0.0278em;","O",[21,11579],{"className":11580,"style":686},[401],[21,11582,691],{"className":11583},[690],[21,11585],{"className":11586,"style":686},[401],[21,11588,11590,11593,11596],{"className":11589},[33],[21,11591],{"className":11592,"style":2200},[37],[21,11594,3644],{"className":11595},[42,46],[21,11597,1088],{"className":11598,"style":1087},[42,46]," Row ",[21,11601,11603],{"className":11602},[24],[21,11604,11606],{"className":11605,"ariaHidden":29},[28],[21,11607,11609,11612],{"className":11608},[33],[21,11610],{"className":11611,"style":112},[37],[21,11613,116],{"className":11614},[42]," of the output is ",[21,11617,11619],{"className":11618},[24],[21,11620,11622,11685,11746,11807,11843,11879,11915],{"className":11621,"ariaHidden":29},[28],[21,11623,11625,11629,11632,11635,11675,11678,11682],{"className":11624},[33],[21,11626],{"className":11627,"style":11628},[37],"height:0.7944em;vertical-align:-0.15em;",[21,11630,10904],{"className":11631},[42],[21,11633],{"className":11634,"style":402},[401],[21,11636,11638,11641],{"className":11637},[42],[21,11639,297],{"className":11640,"style":202},[42,46],[21,11642,11644],{"className":11643},[51],[21,11645,11647,11667],{"className":11646},[55,56],[21,11648,11650,11664],{"className":11649},[60],[21,11651,11653],{"className":11652,"style":5491},[64],[21,11654,11655,11658],{"style":312},[21,11656],{"className":11657,"style":73},[72],[21,11659,11661],{"className":11660},[77,78,79,80],[21,11662,116],{"className":11663},[42,80],[21,11665,89],{"className":11666},[88],[21,11668,11670],{"className":11669},[60],[21,11671,11673],{"className":11672,"style":96},[64],[21,11674],{},[21,11676],{"className":11677,"style":1087},[401],[21,11679,11681],{"className":11680},[1218],"+",[21,11683],{"className":11684,"style":1087},[401],[21,11686,11688,11691,11694,11697,11737,11740,11743],{"className":11687},[33],[21,11689],{"className":11690,"style":11628},[37],[21,11692,10940],{"className":11693},[42],[21,11695],{"className":11696,"style":402},[401],[21,11698,11700,11703],{"className":11699},[42],[21,11701,297],{"className":11702,"style":202},[42,46],[21,11704,11706],{"className":11705},[51],[21,11707,11709,11729],{"className":11708},[55,56],[21,11710,11712,11726],{"className":11711},[60],[21,11713,11715],{"className":11714,"style":5491},[64],[21,11716,11717,11720],{"style":312},[21,11718],{"className":11719,"style":73},[72],[21,11721,11723],{"className":11722},[77,78,79,80],[21,11724,5228],{"className":11725},[42,80],[21,11727,89],{"className":11728},[88],[21,11730,11732],{"className":11731},[60],[21,11733,11735],{"className":11734,"style":96},[64],[21,11736],{},[21,11738],{"className":11739,"style":1087},[401],[21,11741,11681],{"className":11742},[1218],[21,11744],{"className":11745,"style":1087},[401],[21,11747,11749,11752,11755,11758,11798,11801,11804],{"className":11748},[33],[21,11750],{"className":11751,"style":11628},[37],[21,11753,10904],{"className":11754},[42],[21,11756],{"className":11757,"style":402},[401],[21,11759,11761,11764],{"className":11760},[42],[21,11762,297],{"className":11763,"style":202},[42,46],[21,11765,11767],{"className":11766},[51],[21,11768,11770,11790],{"className":11769},[55,56],[21,11771,11773,11787],{"className":11772},[60],[21,11774,11776],{"className":11775,"style":5491},[64],[21,11777,11778,11781],{"style":312},[21,11779],{"className":11780,"style":73},[72],[21,11782,11784],{"className":11783},[77,78,79,80],[21,11785,8876],{"className":11786},[42,80],[21,11788,89],{"className":11789},[88],[21,11791,11793],{"className":11792},[60],[21,11794,11796],{"className":11795,"style":96},[64],[21,11797],{},[21,11799],{"className":11800,"style":686},[401],[21,11802,691],{"className":11803},[690],[21,11805],{"className":11806,"style":686},[401],[21,11808,11810,11813,11816,11819,11822,11825,11828,11831,11834,11837,11840],{"className":11809},[33],[21,11811],{"className":11812,"style":1365},[37],[21,11814,10904],{"className":11815},[42],[21,11817,1369],{"className":11818},[221],[21,11820,116],{"className":11821},[42],[21,11823,1378],{"className":11824},[1377],[21,11826],{"className":11827,"style":402},[401],[21,11829,1373],{"className":11830},[42],[21,11832,1388],{"className":11833},[273],[21,11835],{"className":11836,"style":1087},[401],[21,11838,11681],{"className":11839},[1218],[21,11841],{"className":11842,"style":1087},[401],[21,11844,11846,11849,11852,11855,11858,11861,11864,11867,11870,11873,11876],{"className":11845},[33],[21,11847],{"className":11848,"style":1365},[37],[21,11850,10940],{"className":11851},[42],[21,11853,1369],{"className":11854},[221],[21,11856,1373],{"className":11857},[42],[21,11859,1378],{"className":11860},[1377],[21,11862],{"className":11863,"style":402},[401],[21,11865,116],{"className":11866},[42],[21,11868,1388],{"className":11869},[273],[21,11871],{"className":11872,"style":1087},[401],[21,11874,11681],{"className":11875},[1218],[21,11877],{"className":11878,"style":1087},[401],[21,11880,11882,11885,11888,11891,11894,11897,11900,11903,11906,11909,11912],{"className":11881},[33],[21,11883],{"className":11884,"style":1365},[37],[21,11886,10904],{"className":11887},[42],[21,11889,1369],{"className":11890},[221],[21,11892,9001],{"className":11893},[42],[21,11895,1378],{"className":11896},[1377],[21,11898],{"className":11899,"style":402},[401],[21,11901,9001],{"className":11902},[42],[21,11904,1388],{"className":11905},[273],[21,11907],{"className":11908,"style":686},[401],[21,11910,691],{"className":11911},[690],[21,11913],{"className":11914,"style":686},[401],[21,11916,11918,11921,11924,11928,11931,11934,11938],{"className":11917},[33],[21,11919],{"className":11920,"style":1365},[37],[21,11922,1369],{"className":11923},[221],[21,11925,11927],{"className":11926},[42],"0.602",[21,11929,1378],{"className":11930},[1377],[21,11932],{"className":11933,"style":402},[401],[21,11935,11937],{"className":11936},[42],"0.398",[21,11939,1388],{"className":11940},[273],". All three outputs:",[21,11943,11945],{"className":11944},[1582],[21,11946,11948],{"className":11947},[24],[21,11949,11951,11969,11990,12213],{"className":11950,"ariaHidden":29},[28],[21,11952,11954,11957,11960,11963,11966],{"className":11953},[33],[21,11955],{"className":11956,"style":2200},[37],[21,11958,11577],{"className":11959,"style":11576},[42,46],[21,11961],{"className":11962,"style":686},[401],[21,11964,691],{"className":11965},[690],[21,11967],{"className":11968,"style":686},[401],[21,11970,11972,11975,11978,11981,11984,11987],{"className":11971},[33],[21,11973],{"className":11974,"style":2200},[37],[21,11976,3644],{"className":11977},[42,46],[21,11979,1088],{"className":11980,"style":1087},[42,46],[21,11982],{"className":11983,"style":686},[401],[21,11985,691],{"className":11986},[690],[21,11988],{"className":11989,"style":686},[401],[21,11991,11993,11996,12198,12201,12204,12207,12210],{"className":11992},[33],[21,11994],{"className":11995,"style":8601},[37],[21,11997,11999,12036,12161],{"className":11998},[2521],[21,12000,12002],{"className":12001},[221],[21,12003,12005],{"className":12004},[3285,8611],[21,12006,12008,12028],{"className":12007},[55,56],[21,12009,12011,12025],{"className":12010},[60],[21,12012,12014],{"className":12013,"style":8621},[64],[21,12015,12016,12019],{"style":8624},[21,12017],{"className":12018,"style":8628},[72],[21,12020,12021],{"style":8631},[2628,12022,12023],{"xmlns":2630,"width":8634,"height":8635,"viewBox":8636},[2636,12024],{"d":8639},[21,12026,89],{"className":12027},[88],[21,12029,12031],{"className":12030},[60],[21,12032,12034],{"className":12033,"style":8649},[64],[21,12035],{},[21,12037,12039],{"className":12038},[42],[21,12040,12042,12099,12102,12105],{"className":12041},[8658],[21,12043,12045],{"className":12044},[8662],[21,12046,12048,12091],{"className":12047},[55,56],[21,12049,12051,12088],{"className":12050},[60],[21,12052,12054,12065,12076],{"className":12053,"style":8621},[64],[21,12055,12056,12059],{"style":8674},[21,12057],{"className":12058,"style":1678},[72],[21,12060,12062],{"className":12061},[42],[21,12063,11927],{"className":12064},[42],[21,12066,12067,12070],{"style":8686},[21,12068],{"className":12069,"style":1678},[72],[21,12071,12073],{"className":12072},[42],[21,12074,11937],{"className":12075},[42],[21,12077,12078,12081],{"style":8698},[21,12079],{"className":12080,"style":1678},[72],[21,12082,12084],{"className":12083},[42],[21,12085,12087],{"className":12086},[42],"0.500",[21,12089,89],{"className":12090},[88],[21,12092,12094],{"className":12093},[60],[21,12095,12097],{"className":12096,"style":8649},[64],[21,12098],{},[21,12100],{"className":12101,"style":8723},[8722],[21,12103],{"className":12104,"style":8723},[8722],[21,12106,12108],{"className":12107},[8662],[21,12109,12111,12153],{"className":12110},[55,56],[21,12112,12114,12150],{"className":12113},[60],[21,12115,12117,12128,12139],{"className":12116,"style":8621},[64],[21,12118,12119,12122],{"style":8674},[21,12120],{"className":12121,"style":1678},[72],[21,12123,12125],{"className":12124},[42],[21,12126,11937],{"className":12127},[42],[21,12129,12130,12133],{"style":8686},[21,12131],{"className":12132,"style":1678},[72],[21,12134,12136],{"className":12135},[42],[21,12137,11927],{"className":12138},[42],[21,12140,12141,12144],{"style":8698},[21,12142],{"className":12143,"style":1678},[72],[21,12145,12147],{"className":12146},[42],[21,12148,12087],{"className":12149},[42],[21,12151,89],{"className":12152},[88],[21,12154,12156],{"className":12155},[60],[21,12157,12159],{"className":12158,"style":8649},[64],[21,12160],{},[21,12162,12164],{"className":12163},[273],[21,12165,12167],{"className":12166},[3285,8611],[21,12168,12170,12190],{"className":12169},[55,56],[21,12171,12173,12187],{"className":12172},[60],[21,12174,12176],{"className":12175,"style":8621},[64],[21,12177,12178,12181],{"style":8624},[21,12179],{"className":12180,"style":8628},[72],[21,12182,12183],{"style":8631},[2628,12184,12185],{"xmlns":2630,"width":8634,"height":8635,"viewBox":8636},[2636,12186],{"d":8809},[21,12188,89],{"className":12189},[88],[21,12191,12193],{"className":12192},[60],[21,12194,12196],{"className":12195,"style":8649},[64],[21,12197],{},[21,12199],{"className":12200,"style":686},[401],[21,12202],{"className":12203,"style":686},[401],[21,12205,3730],{"className":12206},[690],[21,12208],{"className":12209,"style":686},[401],[21,12211],{"className":12212,"style":686},[401],[21,12214,12216,12219,12257],{"className":12215},[33],[21,12217],{"className":12218,"style":8861},[37],[21,12220,12222,12225],{"className":12221},[42],[21,12223,773],{"className":12224},[42,772],[21,12226,12228],{"className":12227},[51],[21,12229,12231],{"className":12230},[55],[21,12232,12234],{"className":12233},[60],[21,12235,12237],{"className":12236,"style":8861},[64],[21,12238,12239,12242],{"style":3127},[21,12240],{"className":12241,"style":73},[72],[21,12243,12245],{"className":12244},[77,78,79,80],[21,12246,12248,12251,12254],{"className":12247},[42,80],[21,12249,8876],{"className":12250},[42,80],[21,12252,2550],{"className":12253},[1218,80],[21,12255,5228],{"className":12256},[42,80],[21,12258,2126],{"className":12259},[42],[11,12261,12262,12263,12266],{},"Each output row is a point ",[163,12264,12265],{},"inside"," the triangle spanned by the three value\nvectors — a convex combination, never an extrapolation beyond them. That is the\ngeometric signature of attention: the output can only ever be a weighted average\nof the values it was given.",[616,12268,12270],{"id":12269},"self-attention","Self-attention",[11,12272,12273,12274,12277,12278,12348,12349,12351],{},"When the queries, keys, and values are all linear projections of the ",[163,12275,12276],{},"same","\nsequence ",[21,12279,12281],{"className":12280},[24],[21,12282,12284,12304],{"className":12283,"ariaHidden":29},[28],[21,12285,12287,12290,12295,12298,12301],{"className":12286},[33],[21,12288],{"className":12289,"style":4493},[37],[21,12291,12294],{"className":12292,"style":12293},[42,46],"margin-right:0.0785em;","X",[21,12296],{"className":12297,"style":686},[401],[21,12299,3730],{"className":12300},[690],[21,12302],{"className":12303,"style":686},[401],[21,12305,12307,12310],{"className":12306},[33],[21,12308],{"className":12309,"style":765},[37],[21,12311,12313,12316],{"className":12312},[42],[21,12314,773],{"className":12315},[42,772],[21,12317,12319],{"className":12318},[51],[21,12320,12322],{"className":12321},[55],[21,12323,12325],{"className":12324},[60],[21,12326,12328],{"className":12327,"style":765},[64],[21,12329,12330,12333],{"style":788},[21,12331],{"className":12332,"style":73},[72],[21,12334,12336],{"className":12335},[77,78,79,80],[21,12337,12339,12342,12345],{"className":12338},[42,80],[21,12340,134],{"className":12341},[42,46,80],[21,12343,2550],{"className":12344},[1218,80],[21,12346,804],{"className":12347},[42,46,80],", attention becomes ",[158,12350,12269],{},":\neach position attends to every other position in its own input, including itself.",[177,12353,12354],{"type":179},[11,12355,12356,12359,12360,9562,12433,9562,12506,12579,12580,12595],{},[158,12357,12358],{},"Definition (Self-attention)."," Self-attention is scaled dot-product attention\nin which ",[21,12361,12363],{"className":12362},[24],[21,12364,12366,12384],{"className":12365,"ariaHidden":29},[28],[21,12367,12369,12372,12375,12378,12381],{"className":12368},[33],[21,12370],{"className":12371,"style":2166},[37],[21,12373,739],{"className":12374},[42,46],[21,12376],{"className":12377,"style":686},[401],[21,12379,691],{"className":12380},[690],[21,12382],{"className":12383,"style":686},[401],[21,12385,12387,12390,12393],{"className":12386},[33],[21,12388],{"className":12389,"style":701},[37],[21,12391,12294],{"className":12392,"style":12293},[42,46],[21,12394,12396,12399],{"className":12395},[42],[21,12397,713],{"className":12398,"style":712},[42,46],[21,12400,12402],{"className":12401},[51],[21,12403,12405,12425],{"className":12404},[55,56],[21,12406,12408,12422],{"className":12407},[60],[21,12409,12411],{"className":12410,"style":726},[64],[21,12412,12413,12416],{"style":729},[21,12414],{"className":12415,"style":73},[72],[21,12417,12419],{"className":12418},[77,78,79,80],[21,12420,739],{"className":12421},[42,46,80],[21,12423,89],{"className":12424},[88],[21,12426,12428],{"className":12427},[60],[21,12429,12431],{"className":12430,"style":267},[64],[21,12432],{},[21,12434,12436],{"className":12435},[24],[21,12437,12439,12457],{"className":12438,"ariaHidden":29},[28],[21,12440,12442,12445,12448,12451,12454],{"className":12441},[33],[21,12443],{"className":12444,"style":2200},[37],[21,12446,920],{"className":12447,"style":919},[42,46],[21,12449],{"className":12450,"style":686},[401],[21,12452,691],{"className":12453},[690],[21,12455],{"className":12456,"style":686},[401],[21,12458,12460,12463,12466],{"className":12459},[33],[21,12461],{"className":12462,"style":886},[37],[21,12464,12294],{"className":12465,"style":12293},[42,46],[21,12467,12469,12472],{"className":12468},[42],[21,12470,713],{"className":12471,"style":712},[42,46],[21,12473,12475],{"className":12474},[51],[21,12476,12478,12498],{"className":12477},[55,56],[21,12479,12481,12495],{"className":12480},[60],[21,12482,12484],{"className":12483,"style":726},[64],[21,12485,12486,12489],{"style":729},[21,12487],{"className":12488,"style":73},[72],[21,12490,12492],{"className":12491},[77,78,79,80],[21,12493,920],{"className":12494,"style":919},[42,46,80],[21,12496,89],{"className":12497},[88],[21,12499,12501],{"className":12500},[60],[21,12502,12504],{"className":12503,"style":96},[64],[21,12505],{},[21,12507,12509],{"className":12508},[24],[21,12510,12512,12530],{"className":12511,"ariaHidden":29},[28],[21,12513,12515,12518,12521,12524,12527],{"className":12514},[33],[21,12516],{"className":12517,"style":2200},[37],[21,12519,1088],{"className":12520,"style":1087},[42,46],[21,12522],{"className":12523,"style":686},[401],[21,12525,691],{"className":12526},[690],[21,12528],{"className":12529,"style":686},[401],[21,12531,12533,12536,12539],{"className":12532},[33],[21,12534],{"className":12535,"style":886},[37],[21,12537,12294],{"className":12538,"style":12293},[42,46],[21,12540,12542,12545],{"className":12541},[42],[21,12543,713],{"className":12544,"style":712},[42,46],[21,12546,12548],{"className":12547},[51],[21,12549,12551,12571],{"className":12550},[55,56],[21,12552,12554,12568],{"className":12553},[60],[21,12555,12557],{"className":12556,"style":726},[64],[21,12558,12559,12562],{"style":729},[21,12560],{"className":12561,"style":73},[72],[21,12563,12565],{"className":12564},[77,78,79,80],[21,12566,1088],{"className":12567,"style":1087},[42,46,80],[21,12569,89],{"className":12570},[88],[21,12572,12574],{"className":12573},[60],[21,12575,12577],{"className":12576,"style":96},[64],[21,12578],{}," all derive from one input\nsequence ",[21,12581,12583],{"className":12582},[24],[21,12584,12586],{"className":12585,"ariaHidden":29},[28],[21,12587,12589,12592],{"className":12588},[33],[21,12590],{"className":12591,"style":2200},[37],[21,12593,12294],{"className":12594,"style":12293},[42,46],". The result re-expresses each position as a content-weighted\nsummary of the whole sequence.",[11,12597,12598,12599],{},"Its central object is the ",[158,12600,12601],{},"attention matrix",[21,12603,12605],{"className":12604},[1582],[21,12606,12608],{"className":12607},[24],[21,12609,12611,12629,12843,12951],{"className":12610,"ariaHidden":29},[28],[21,12612,12614,12617,12620,12623,12626],{"className":12613},[33],[21,12615],{"className":12616,"style":2200},[37],[21,12618,3644],{"className":12619},[42,46],[21,12621],{"className":12622,"style":686},[401],[21,12624,691],{"className":12625},[690],[21,12627],{"className":12628,"style":686},[401],[21,12630,12632,12635,12641,12644,12647,12834,12837,12840],{"className":12631},[33],[21,12633],{"className":12634,"style":4757},[37],[21,12636,12638],{"className":12637},[353],[21,12639,3266],{"className":12640},[42,3265],[21,12642],{"className":12643,"style":3270},[401],[21,12645],{"className":12646,"style":402},[401],[21,12648,12650,12656,12828],{"className":12649},[2521],[21,12651,12653],{"className":12652,"style":3281},[221,3280],[21,12654,1369],{"className":12655},[3285,79],[21,12657,12659,12662,12825],{"className":12658},[42],[21,12660],{"className":12661},[221,1657],[21,12663,12665],{"className":12664},[1661],[21,12666,12668,12817],{"className":12667},[55,56],[21,12669,12671,12814],{"className":12670},[60],[21,12672,12674,12766,12774],{"className":12673,"style":4797},[64],[21,12675,12676,12679],{"style":4800},[21,12677],{"className":12678,"style":1678},[72],[21,12680,12682],{"className":12681},[42],[21,12683,12685],{"className":12684},[42,3321],[21,12686,12688,12758],{"className":12687},[55,56],[21,12689,12691,12755],{"className":12690},[60],[21,12692,12694,12743],{"className":12693,"style":3978},[64],[21,12695,12697,12700],{"className":12696,"style":2593},[2612],[21,12698],{"className":12699,"style":1678},[72],[21,12701,12703],{"className":12702,"style":3341},[42],[21,12704,12706,12709],{"className":12705},[42],[21,12707,804],{"className":12708},[42,46],[21,12710,12712],{"className":12711},[51],[21,12713,12715,12735],{"className":12714},[55,56],[21,12716,12718,12732],{"className":12717},[60],[21,12719,12721],{"className":12720,"style":2424},[64],[21,12722,12723,12726],{"style":68},[21,12724],{"className":12725,"style":73},[72],[21,12727,12729],{"className":12728},[77,78,79,80],[21,12730,230],{"className":12731,"style":229},[42,46,80],[21,12733,89],{"className":12734},[88],[21,12736,12738],{"className":12737},[60],[21,12739,12741],{"className":12740,"style":96},[64],[21,12742],{},[21,12744,12745,12748],{"style":4030},[21,12746],{"className":12747,"style":1678},[72],[21,12749,12751],{"className":12750,"style":3392},[3391],[2628,12752,12753],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,12754],{"d":3399},[21,12756,89],{"className":12757},[88],[21,12759,12761],{"className":12760},[60],[21,12762,12764],{"className":12763,"style":4050},[64],[21,12765],{},[21,12767,12768,12771],{"style":1858},[21,12769],{"className":12770,"style":1678},[72],[21,12772],{"className":12773,"style":1866},[1865],[21,12775,12776,12779],{"style":1869},[21,12777],{"className":12778,"style":1678},[72],[21,12780,12782,12785],{"className":12781},[42],[21,12783,739],{"className":12784},[42,46],[21,12786,12788,12791],{"className":12787},[42],[21,12789,920],{"className":12790,"style":919},[42,46],[21,12792,12794],{"className":12793},[51],[21,12795,12797],{"className":12796},[55],[21,12798,12800],{"className":12799},[60],[21,12801,12803],{"className":12802,"style":765},[64],[21,12804,12805,12808],{"style":788},[21,12806],{"className":12807,"style":73},[72],[21,12809,12811],{"className":12810},[77,78,79,80],[21,12812,3137],{"className":12813},[42,80],[21,12815,89],{"className":12816},[88],[21,12818,12820],{"className":12819},[60],[21,12821,12823],{"className":12822,"style":4948},[64],[21,12824],{},[21,12826],{"className":12827},[273,1657],[21,12829,12831],{"className":12830,"style":3281},[273,3280],[21,12832,1388],{"className":12833},[3285,79],[21,12835],{"className":12836,"style":686},[401],[21,12838,3730],{"className":12839},[690],[21,12841],{"className":12842,"style":686},[401],[21,12844,12846,12850,12889,12892,12895,12898,12942,12945,12948],{"className":12845},[33],[21,12847],{"className":12848,"style":12849},[37],"height:1.1074em;vertical-align:-0.2861em;",[21,12851,12853,12856],{"className":12852},[42],[21,12854,773],{"className":12855},[42,772],[21,12857,12859],{"className":12858},[51],[21,12860,12862],{"className":12861},[55],[21,12863,12865],{"className":12864},[60],[21,12866,12869],{"className":12867,"style":12868},[64],"height:0.8213em;",[21,12870,12871,12874],{"style":3127},[21,12872],{"className":12873,"style":73},[72],[21,12875,12877],{"className":12876},[77,78,79,80],[21,12878,12880,12883,12886],{"className":12879},[42,80],[21,12881,134],{"className":12882},[42,46,80],[21,12884,2550],{"className":12885},[1218,80],[21,12887,134],{"className":12888},[42,46,80],[21,12890,1378],{"className":12891},[1377],[21,12893],{"className":12894,"style":1958},[401],[21,12896],{"className":12897,"style":402},[401],[21,12899,12901,12904],{"className":12900},[42],[21,12902,3644],{"className":12903},[42,46],[21,12905,12907],{"className":12906},[51],[21,12908,12910,12934],{"className":12909},[55,56],[21,12911,12913,12931],{"className":12912},[60],[21,12914,12916],{"className":12915,"style":243},[64],[21,12917,12918,12921],{"style":68},[21,12919],{"className":12920,"style":73},[72],[21,12922,12924],{"className":12923},[77,78,79,80],[21,12925,12927],{"className":12926},[42,80],[21,12928,12930],{"className":12929,"style":256},[42,46,80],"ij",[21,12932,89],{"className":12933},[88],[21,12935,12937],{"className":12936},[60],[21,12938,12940],{"className":12939,"style":267},[64],[21,12941],{},[21,12943],{"className":12944,"style":686},[401],[21,12946,691],{"className":12947},[690],[21,12949],{"className":12950,"style":686},[401],[21,12952,12954,12957,12964,12967,12974,12977],{"className":12953},[33],[21,12955],{"className":12956,"style":2331},[37],[21,12958,12960],{"className":12959},[42,1965],[21,12961,12963],{"className":12962},[42],"how much position ",[21,12965,3798],{"className":12966},[42,46],[21,12968,12970],{"className":12969},[42,1965],[21,12971,12973],{"className":12972},[42]," attends to position ",[21,12975,257],{"className":12976,"style":256},[42,46],[21,12978,1378],{"className":12979},[1377],[11,12981,12982,12983,13099,13100,13103,13104,13119,13120,13135,13136,13139,13140,13143],{},"a row-stochastic matrix (",[21,12984,12986],{"className":12985},[24],[21,12987,12989,13090],{"className":12988,"ariaHidden":29},[28],[21,12990,12992,12995,13035,13038,13081,13084,13087],{"className":12991},[33],[21,12993],{"className":12994,"style":349},[37],[21,12996,12998,13001],{"className":12997},[353],[21,12999,360],{"className":13000,"style":359},[353,357,358],[21,13002,13004],{"className":13003},[51],[21,13005,13007,13027],{"className":13006},[55,56],[21,13008,13010,13024],{"className":13009},[60],[21,13011,13013],{"className":13012,"style":373},[64],[21,13014,13015,13018],{"style":376},[21,13016],{"className":13017,"style":73},[72],[21,13019,13021],{"className":13020},[77,78,79,80],[21,13022,257],{"className":13023,"style":256},[42,46,80],[21,13025,89],{"className":13026},[88],[21,13028,13030],{"className":13029},[60],[21,13031,13033],{"className":13032,"style":395},[64],[21,13034],{},[21,13036],{"className":13037,"style":402},[401],[21,13039,13041,13044],{"className":13040},[42],[21,13042,3644],{"className":13043},[42,46],[21,13045,13047],{"className":13046},[51],[21,13048,13050,13073],{"className":13049},[55,56],[21,13051,13053,13070],{"className":13052},[60],[21,13054,13056],{"className":13055,"style":243},[64],[21,13057,13058,13061],{"style":68},[21,13059],{"className":13060,"style":73},[72],[21,13062,13064],{"className":13063},[77,78,79,80],[21,13065,13067],{"className":13066},[42,80],[21,13068,12930],{"className":13069,"style":256},[42,46,80],[21,13071,89],{"className":13072},[88],[21,13074,13076],{"className":13075},[60],[21,13077,13079],{"className":13078,"style":267},[64],[21,13080],{},[21,13082],{"className":13083,"style":686},[401],[21,13085,691],{"className":13086},[690],[21,13088],{"className":13089,"style":686},[401],[21,13091,13093,13096],{"className":13092},[33],[21,13094],{"className":13095,"style":112},[37],[21,13097,116],{"className":13098},[42],") that is a soft, content-based,\n",[158,13101,13102],{},"all-pairs routing table",": row ",[21,13105,13107],{"className":13106},[24],[21,13108,13110],{"className":13109,"ariaHidden":29},[28],[21,13111,13113,13116],{"className":13112},[33],[21,13114],{"className":13115,"style":3932},[37],[21,13117,3798],{"className":13118},[42,46]," says where token ",[21,13121,13123],{"className":13122},[24],[21,13124,13126],{"className":13125,"ariaHidden":29},[28],[21,13127,13129,13132],{"className":13128},[33],[21,13130],{"className":13131,"style":3932},[37],[21,13133,3798],{"className":13134},[42,46]," gathers its information.\nUnlike a convolution's fixed local stencil or a recurrence's strictly leftward\nchain, every entry is non-zero and ",[163,13137,13138],{},"learned from content",": the word ",[203,13141,13142],{},"it"," can\nroute directly to the noun it refers to, however far back.",[2128,13145],{"hash":13146},"e1ce3af2d701a9d05bee403986619251fea3b30dced65a3e76c4ee9255eb5c28",[5230,13148,13150],{"id":13149},"permutation-equivariance","Permutation equivariance",[11,13152,13153,13154,13169,13170,13173],{},"Self-attention has no notion of order. Permuting the rows of ",[21,13155,13157],{"className":13156},[24],[21,13158,13160],{"className":13159,"ariaHidden":29},[28],[21,13161,13163,13166],{"className":13162},[33],[21,13164],{"className":13165,"style":2200},[37],[21,13167,12294],{"className":13168,"style":12293},[42,46]," permutes the rows\nof the output identically and changes nothing else; the mechanism treats its\ninput as a ",[163,13171,13172],{},"set",", not a sequence.",[177,13175,13177],{"type":13176},"theorem",[11,13178,13179,13182,13183,13199,13200,13233,13234,13430,13431,2126],{},[158,13180,13181],{},"Theorem (Permutation equivariance of self-attention)."," Let ",[21,13184,13186],{"className":13185},[24],[21,13187,13189],{"className":13188,"ariaHidden":29},[28],[21,13190,13192,13195],{"className":13191},[33],[21,13193],{"className":13194,"style":2200},[37],[21,13196,13198],{"className":13197,"style":712},[42,46],"P"," be an ",[21,13201,13203],{"className":13202},[24],[21,13204,13206,13224],{"className":13205,"ariaHidden":29},[28],[21,13207,13209,13212,13215,13218,13221],{"className":13208},[33],[21,13210],{"className":13211,"style":4149},[37],[21,13213,134],{"className":13214},[42,46],[21,13216],{"className":13217,"style":1087},[401],[21,13219,2550],{"className":13220},[1218],[21,13222],{"className":13223,"style":1087},[401],[21,13225,13227,13230],{"className":13226},[33],[21,13228],{"className":13229,"style":130},[37],[21,13231,134],{"className":13232},[42,46]," permutation matrix and let ",[21,13235,13237],{"className":13236},[24],[21,13238,13240,13271],{"className":13239,"ariaHidden":29},[28],[21,13241,13243,13246,13253,13256,13259,13262,13265,13268],{"className":13242},[33],[21,13244],{"className":13245,"style":1365},[37],[21,13247,13249],{"className":13248},[353],[21,13250,13252],{"className":13251},[42,3265],"SA",[21,13254,1369],{"className":13255},[221],[21,13257,12294],{"className":13258,"style":12293},[42,46],[21,13260,1388],{"className":13261},[273],[21,13263],{"className":13264,"style":686},[401],[21,13266,691],{"className":13267},[690],[21,13269],{"className":13270,"style":686},[401],[21,13272,13274,13277,13283,13286,13289,13329,13332,13335,13338,13378,13381,13384,13387,13427],{"className":13273},[33],[21,13275],{"className":13276,"style":217},[37],[21,13278,13280],{"className":13279},[353],[21,13281,4714],{"className":13282},[42,3265],[21,13284,1369],{"className":13285},[221],[21,13287,12294],{"className":13288,"style":12293},[42,46],[21,13290,13292,13295],{"className":13291},[42],[21,13293,713],{"className":13294,"style":712},[42,46],[21,13296,13298],{"className":13297},[51],[21,13299,13301,13321],{"className":13300},[55,56],[21,13302,13304,13318],{"className":13303},[60],[21,13305,13307],{"className":13306,"style":726},[64],[21,13308,13309,13312],{"style":729},[21,13310],{"className":13311,"style":73},[72],[21,13313,13315],{"className":13314},[77,78,79,80],[21,13316,739],{"className":13317},[42,46,80],[21,13319,89],{"className":13320},[88],[21,13322,13324],{"className":13323},[60],[21,13325,13327],{"className":13326,"style":267},[64],[21,13328],{},[21,13330,1378],{"className":13331},[1377],[21,13333],{"className":13334,"style":402},[401],[21,13336,12294],{"className":13337,"style":12293},[42,46],[21,13339,13341,13344],{"className":13340},[42],[21,13342,713],{"className":13343,"style":712},[42,46],[21,13345,13347],{"className":13346},[51],[21,13348,13350,13370],{"className":13349},[55,56],[21,13351,13353,13367],{"className":13352},[60],[21,13354,13356],{"className":13355,"style":726},[64],[21,13357,13358,13361],{"style":729},[21,13359],{"className":13360,"style":73},[72],[21,13362,13364],{"className":13363},[77,78,79,80],[21,13365,920],{"className":13366,"style":919},[42,46,80],[21,13368,89],{"className":13369},[88],[21,13371,13373],{"className":13372},[60],[21,13374,13376],{"className":13375,"style":96},[64],[21,13377],{},[21,13379,1378],{"className":13380},[1377],[21,13382],{"className":13383,"style":402},[401],[21,13385,12294],{"className":13386,"style":12293},[42,46],[21,13388,13390,13393],{"className":13389},[42],[21,13391,713],{"className":13392,"style":712},[42,46],[21,13394,13396],{"className":13395},[51],[21,13397,13399,13419],{"className":13398},[55,56],[21,13400,13402,13416],{"className":13401},[60],[21,13403,13405],{"className":13404,"style":726},[64],[21,13406,13407,13410],{"style":729},[21,13408],{"className":13409,"style":73},[72],[21,13411,13413],{"className":13412},[77,78,79,80],[21,13414,1088],{"className":13415,"style":1087},[42,46,80],[21,13417,89],{"className":13418},[88],[21,13420,13422],{"className":13421},[60],[21,13423,13425],{"className":13424,"style":96},[64],[21,13426],{},[21,13428,1388],{"className":13429},[273],". Then ",[21,13432,13434],{"className":13433},[24],[21,13435,13437,13470],{"className":13436,"ariaHidden":29},[28],[21,13438,13440,13443,13449,13452,13455,13458,13461,13464,13467],{"className":13439},[33],[21,13441],{"className":13442,"style":1365},[37],[21,13444,13446],{"className":13445},[353],[21,13447,13252],{"className":13448},[42,3265],[21,13450,1369],{"className":13451},[221],[21,13453,13198],{"className":13454,"style":712},[42,46],[21,13456,12294],{"className":13457,"style":12293},[42,46],[21,13459,1388],{"className":13460},[273],[21,13462],{"className":13463,"style":686},[401],[21,13465,691],{"className":13466},[690],[21,13468],{"className":13469,"style":686},[401],[21,13471,13473,13476,13479,13482,13485,13491,13494,13497],{"className":13472},[33],[21,13474],{"className":13475,"style":1365},[37],[21,13477,13198],{"className":13478,"style":712},[42,46],[21,13480],{"className":13481,"style":402},[401],[21,13483],{"className":13484,"style":402},[401],[21,13486,13488],{"className":13487},[353],[21,13489,13252],{"className":13490},[42,3265],[21,13492,1369],{"className":13493},[221],[21,13495,12294],{"className":13496,"style":12293},[42,46],[21,13498,1388],{"className":13499},[273],[177,13501,13502,13984,14125],{"type":6369},[11,13503,13504,13506,13507,9562,13544,9562,13581,13617,13618,13633,13634,13788,13789,13830,13831,13846,13847,13983],{},[158,13505,6374],{}," Permuting the inputs sends ",[21,13508,13510],{"className":13509},[24],[21,13511,13513,13532],{"className":13512,"ariaHidden":29},[28],[21,13514,13516,13519,13522,13525,13529],{"className":13515},[33],[21,13517],{"className":13518,"style":2166},[37],[21,13520,739],{"className":13521},[42,46],[21,13523],{"className":13524,"style":686},[401],[21,13526,13528],{"className":13527},[690],"↦",[21,13530],{"className":13531,"style":686},[401],[21,13533,13535,13538,13541],{"className":13534},[33],[21,13536],{"className":13537,"style":2166},[37],[21,13539,13198],{"className":13540,"style":712},[42,46],[21,13542,739],{"className":13543},[42,46],[21,13545,13547],{"className":13546},[24],[21,13548,13550,13569],{"className":13549,"ariaHidden":29},[28],[21,13551,13553,13557,13560,13563,13566],{"className":13552},[33],[21,13554],{"className":13555,"style":13556},[37],"height:0.6943em;vertical-align:-0.011em;",[21,13558,920],{"className":13559,"style":919},[42,46],[21,13561],{"className":13562,"style":686},[401],[21,13564,13528],{"className":13565},[690],[21,13567],{"className":13568,"style":686},[401],[21,13570,13572,13575,13578],{"className":13571},[33],[21,13573],{"className":13574,"style":2200},[37],[21,13576,13198],{"className":13577,"style":712},[42,46],[21,13579,920],{"className":13580,"style":919},[42,46],[21,13582,13584],{"className":13583},[24],[21,13585,13587,13605],{"className":13586,"ariaHidden":29},[28],[21,13588,13590,13593,13596,13599,13602],{"className":13589},[33],[21,13591],{"className":13592,"style":13556},[37],[21,13594,1088],{"className":13595,"style":1087},[42,46],[21,13597],{"className":13598,"style":686},[401],[21,13600,13528],{"className":13601},[690],[21,13603],{"className":13604,"style":686},[401],[21,13606,13608,13611,13614],{"className":13607},[33],[21,13609],{"className":13610,"style":2200},[37],[21,13612,13198],{"className":13613,"style":712},[42,46],[21,13615,1088],{"className":13616,"style":1087},[42,46],", since each is ",[21,13619,13621],{"className":13620},[24],[21,13622,13624],{"className":13623,"ariaHidden":29},[28],[21,13625,13627,13630],{"className":13626},[33],[21,13628],{"className":13629,"style":2200},[37],[21,13631,12294],{"className":13632,"style":12293},[42,46]," times a fixed projection. The scores become ",[21,13635,13637],{"className":13636},[24],[21,13638,13640,13706],{"className":13639,"ariaHidden":29},[28],[21,13641,13643,13647,13650,13653,13656,13659,13662,13665,13668,13697,13700,13703],{"className":13642},[33],[21,13644],{"className":13645,"style":13646},[37],"height:1.0991em;vertical-align:-0.25em;",[21,13648,1369],{"className":13649},[221],[21,13651,13198],{"className":13652,"style":712},[42,46],[21,13654,739],{"className":13655},[42,46],[21,13657,1388],{"className":13658},[273],[21,13660,1369],{"className":13661},[221],[21,13663,13198],{"className":13664,"style":712},[42,46],[21,13666,920],{"className":13667,"style":919},[42,46],[21,13669,13671,13674],{"className":13670},[273],[21,13672,1388],{"className":13673},[273],[21,13675,13677],{"className":13676},[51],[21,13678,13680],{"className":13679},[55],[21,13681,13683],{"className":13682},[60],[21,13684,13686],{"className":13685,"style":765},[64],[21,13687,13688,13691],{"style":788},[21,13689],{"className":13690,"style":73},[72],[21,13692,13694],{"className":13693},[77,78,79,80],[21,13695,3137],{"className":13696},[42,80],[21,13698],{"className":13699,"style":686},[401],[21,13701,691],{"className":13702},[690],[21,13704],{"className":13705,"style":686},[401],[21,13707,13709,13712,13715,13718,13721,13724,13753,13756,13759],{"className":13708},[33],[21,13710],{"className":13711,"style":13646},[37],[21,13713,13198],{"className":13714,"style":712},[42,46],[21,13716],{"className":13717,"style":402},[401],[21,13719,1369],{"className":13720},[221],[21,13722,739],{"className":13723},[42,46],[21,13725,13727,13730],{"className":13726},[42],[21,13728,920],{"className":13729,"style":919},[42,46],[21,13731,13733],{"className":13732},[51],[21,13734,13736],{"className":13735},[55],[21,13737,13739],{"className":13738},[60],[21,13740,13742],{"className":13741,"style":765},[64],[21,13743,13744,13747],{"style":788},[21,13745],{"className":13746,"style":73},[72],[21,13748,13750],{"className":13749},[77,78,79,80],[21,13751,3137],{"className":13752},[42,80],[21,13754,1388],{"className":13755},[273],[21,13757],{"className":13758,"style":402},[401],[21,13760,13762,13765],{"className":13761},[42],[21,13763,13198],{"className":13764,"style":712},[42,46],[21,13766,13768],{"className":13767},[51],[21,13769,13771],{"className":13770},[55],[21,13772,13774],{"className":13773},[60],[21,13775,13777],{"className":13776,"style":765},[64],[21,13778,13779,13782],{"style":788},[21,13780],{"className":13781,"style":73},[72],[21,13783,13785],{"className":13784},[77,78,79,80],[21,13786,3137],{"className":13787},[42,80],". The softmax acts row-wise and ",[21,13790,13792],{"className":13791},[24],[21,13793,13795],{"className":13794,"ariaHidden":29},[28],[21,13796,13798,13801],{"className":13797},[33],[21,13799],{"className":13800,"style":765},[37],[21,13802,13804,13807],{"className":13803},[42],[21,13805,13198],{"className":13806,"style":712},[42,46],[21,13808,13810],{"className":13809},[51],[21,13811,13813],{"className":13812},[55],[21,13814,13816],{"className":13815},[60],[21,13817,13819],{"className":13818,"style":765},[64],[21,13820,13821,13824],{"style":788},[21,13822],{"className":13823,"style":73},[72],[21,13825,13827],{"className":13826},[77,78,79,80],[21,13828,3137],{"className":13829},[42,80]," permutes columns\nconsistently with the row permutation ",[21,13832,13834],{"className":13833},[24],[21,13835,13837],{"className":13836,"ariaHidden":29},[28],[21,13838,13840,13843],{"className":13839},[33],[21,13841],{"className":13842,"style":2200},[37],[21,13844,13198],{"className":13845,"style":712},[42,46],", so ",[21,13848,13850],{"className":13849},[24],[21,13851,13853,13921],{"className":13852,"ariaHidden":29},[28],[21,13854,13856,13859,13865,13868,13871,13874,13877,13880,13909,13912,13915,13918],{"className":13855},[33],[21,13857],{"className":13858,"style":13646},[37],[21,13860,13862],{"className":13861},[353],[21,13863,3266],{"className":13864},[42,3265],[21,13866,1369],{"className":13867},[221],[21,13869,13198],{"className":13870,"style":712},[42,46],[21,13872],{"className":13873,"style":402},[401],[21,13875,9195],{"className":13876,"style":9194},[42,46],[21,13878],{"className":13879,"style":402},[401],[21,13881,13883,13886],{"className":13882},[42],[21,13884,13198],{"className":13885,"style":712},[42,46],[21,13887,13889],{"className":13888},[51],[21,13890,13892],{"className":13891},[55],[21,13893,13895],{"className":13894},[60],[21,13896,13898],{"className":13897,"style":765},[64],[21,13899,13900,13903],{"style":788},[21,13901],{"className":13902,"style":73},[72],[21,13904,13906],{"className":13905},[77,78,79,80],[21,13907,3137],{"className":13908},[42,80],[21,13910,1388],{"className":13911},[273],[21,13913],{"className":13914,"style":686},[401],[21,13916,691],{"className":13917},[690],[21,13919],{"className":13920,"style":686},[401],[21,13922,13924,13927,13930,13933,13936,13942,13945,13948,13951,13954],{"className":13923},[33],[21,13925],{"className":13926,"style":13646},[37],[21,13928,13198],{"className":13929,"style":712},[42,46],[21,13931],{"className":13932,"style":402},[401],[21,13934],{"className":13935,"style":402},[401],[21,13937,13939],{"className":13938},[353],[21,13940,3266],{"className":13941},[42,3265],[21,13943,1369],{"className":13944},[221],[21,13946,9195],{"className":13947,"style":9194},[42,46],[21,13949,1388],{"className":13950},[273],[21,13952],{"className":13953,"style":402},[401],[21,13955,13957,13960],{"className":13956},[42],[21,13958,13198],{"className":13959,"style":712},[42,46],[21,13961,13963],{"className":13962},[51],[21,13964,13966],{"className":13965},[55],[21,13967,13969],{"className":13968},[60],[21,13970,13972],{"className":13971,"style":765},[64],[21,13973,13974,13977],{"style":788},[21,13975],{"className":13976,"style":73},[72],[21,13978,13980],{"className":13979},[77,78,79,80],[21,13981,3137],{"className":13982},[42,80],". Multiplying by the values,",[21,13985,13987],{"className":13986},[1582],[21,13988,13990],{"className":13989},[24],[21,13991,13993,14065,14086],{"className":13992,"ariaHidden":29},[28],[21,13994,13996,14000,14003,14006,14009,14015,14018,14021,14024,14027,14056,14059,14062],{"className":13995},[33],[21,13997],{"className":13998,"style":13999},[37],"height:1.1491em;vertical-align:-0.25em;",[21,14001,13198],{"className":14002,"style":712},[42,46],[21,14004],{"className":14005,"style":402},[401],[21,14007],{"className":14008,"style":402},[401],[21,14010,14012],{"className":14011},[353],[21,14013,3266],{"className":14014},[42,3265],[21,14016,1369],{"className":14017},[221],[21,14019,9195],{"className":14020,"style":9194},[42,46],[21,14022,1388],{"className":14023},[273],[21,14025],{"className":14026,"style":402},[401],[21,14028,14030,14033],{"className":14029},[42],[21,14031,13198],{"className":14032,"style":712},[42,46],[21,14034,14036],{"className":14035},[51],[21,14037,14039],{"className":14038},[55],[21,14040,14042],{"className":14041},[60],[21,14043,14045],{"className":14044,"style":3027},[64],[21,14046,14047,14050],{"style":3127},[21,14048],{"className":14049,"style":73},[72],[21,14051,14053],{"className":14052},[77,78,79,80],[21,14054,3137],{"className":14055},[42,80],[21,14057],{"className":14058,"style":1087},[401],[21,14060,1219],{"className":14061},[1218],[21,14063],{"className":14064,"style":1087},[401],[21,14066,14068,14071,14074,14077,14080,14083],{"className":14067},[33],[21,14069],{"className":14070,"style":2200},[37],[21,14072,13198],{"className":14073,"style":712},[42,46],[21,14075,1088],{"className":14076,"style":1087},[42,46],[21,14078],{"className":14079,"style":686},[401],[21,14081,691],{"className":14082},[690],[21,14084],{"className":14085,"style":686},[401],[21,14087,14089,14092,14095,14098,14101,14107,14110,14113,14116,14119,14122],{"className":14088},[33],[21,14090],{"className":14091,"style":1365},[37],[21,14093,13198],{"className":14094,"style":712},[42,46],[21,14096],{"className":14097,"style":402},[401],[21,14099],{"className":14100,"style":402},[401],[21,14102,14104],{"className":14103},[353],[21,14105,3266],{"className":14106},[42,3265],[21,14108,1369],{"className":14109},[221],[21,14111,9195],{"className":14112,"style":9194},[42,46],[21,14114,1388],{"className":14115},[273],[21,14117],{"className":14118,"style":402},[401],[21,14120,1088],{"className":14121,"style":1087},[42,46],[21,14123,1378],{"className":14124},[1377],[11,14126,14127,14128,14191,14192,7790,14228],{},"because ",[21,14129,14131],{"className":14130},[24],[21,14132,14134,14181],{"className":14133,"ariaHidden":29},[28],[21,14135,14137,14140,14169,14172,14175,14178],{"className":14136},[33],[21,14138],{"className":14139,"style":765},[37],[21,14141,14143,14146],{"className":14142},[42],[21,14144,13198],{"className":14145,"style":712},[42,46],[21,14147,14149],{"className":14148},[51],[21,14150,14152],{"className":14151},[55],[21,14153,14155],{"className":14154},[60],[21,14156,14158],{"className":14157,"style":765},[64],[21,14159,14160,14163],{"style":788},[21,14161],{"className":14162,"style":73},[72],[21,14164,14166],{"className":14165},[77,78,79,80],[21,14167,3137],{"className":14168},[42,80],[21,14170,13198],{"className":14171,"style":712},[42,46],[21,14173],{"className":14174,"style":686},[401],[21,14176,691],{"className":14177},[690],[21,14179],{"className":14180,"style":686},[401],[21,14182,14184,14187],{"className":14183},[33],[21,14185],{"className":14186,"style":2200},[37],[21,14188,14190],{"className":14189,"style":12293},[42,46],"I",". The right side is ",[21,14193,14195],{"className":14194},[24],[21,14196,14198],{"className":14197,"ariaHidden":29},[28],[21,14199,14201,14204,14207,14210,14213,14219,14222,14225],{"className":14200},[33],[21,14202],{"className":14203,"style":1365},[37],[21,14205,13198],{"className":14206,"style":712},[42,46],[21,14208],{"className":14209,"style":402},[401],[21,14211],{"className":14212,"style":402},[401],[21,14214,14216],{"className":14215},[353],[21,14217,13252],{"className":14218},[42,3265],[21,14220,1369],{"className":14221},[221],[21,14223,12294],{"className":14224,"style":12293},[42,46],[21,14226,1388],{"className":14227},[273],[21,14229,14231],{"className":14230},[24],[21,14232,14234],{"className":14233,"ariaHidden":29},[28],[21,14235,14237,14240],{"className":14236},[33],[21,14238],{"className":14239,"style":7803},[37],[21,14241,14243],{"className":14242},[7807,7808],[21,14244,7813],{"className":14245},[42,7812],[11,14247,14248,14249,5614,14252,14255],{},"Equivariance renders position invisible: ",[203,14250,14251],{},"dog bites man",[203,14253,14254],{},"man bites dog"," produce the same set of outputs, merely reordered. Order\nmust therefore be injected explicitly, the job of positional encoding below.",[616,14257,14259],{"id":14258},"multi-head-attention","Multi-head attention",[11,14261,14262,14263,14265,14266,14281,14282],{},"A single attention function forces every position to mix all of its relationships\ninto one set of weights. ",[158,14264,14259],{}," runs ",[21,14267,14269],{"className":14268},[24],[21,14270,14272],{"className":14271,"ariaHidden":29},[28],[21,14273,14275,14278],{"className":14274},[33],[21,14276],{"className":14277,"style":867},[37],[21,14279,47],{"className":14280},[42,46]," attention functions in\nparallel, each with its own projections, so different heads can specialize: one\ntracking syntactic agreement, another long-range coreference, another local\nadjacency.",[168,14283,14284],{},[15,14285,8876],{"href":14286,"ariaDescribedBy":14287,"dataFootnoteRef":6,"id":14288},"#user-content-fn-chollet-mha",[174],"user-content-fnref-chollet-mha",[177,14290,14291,14313,14610,14819],{"type":179},[11,14292,14293,14296,14297,14312],{},[158,14294,14295],{},"Definition (Multi-head attention)."," Project the inputs ",[21,14298,14300],{"className":14299},[24],[21,14301,14303],{"className":14302,"ariaHidden":29},[28],[21,14304,14306,14309],{"className":14305},[33],[21,14307],{"className":14308,"style":867},[37],[21,14310,47],{"className":14311},[42,46]," times with\nindependent matrices, attend in each subspace, then concatenate and project:",[21,14314,14316],{"className":14315},[1582],[21,14317,14319],{"className":14318},[24],[21,14320,14322,14382],{"className":14321,"ariaHidden":29},[28],[21,14323,14325,14328,14373,14376,14379],{"className":14324},[33],[21,14326],{"className":14327,"style":38},[37],[21,14329,14331,14338],{"className":14330},[353],[21,14332,14334],{"className":14333},[353],[21,14335,14337],{"className":14336},[42,3265],"head",[21,14339,14341],{"className":14340},[51],[21,14342,14344,14365],{"className":14343},[55,56],[21,14345,14347,14362],{"className":14346},[60],[21,14348,14350],{"className":14349,"style":243},[64],[21,14351,14353,14356],{"style":14352},"top:-2.55em;margin-right:0.05em;",[21,14354],{"className":14355,"style":73},[72],[21,14357,14359],{"className":14358},[77,78,79,80],[21,14360,3798],{"className":14361},[42,46,80],[21,14363,89],{"className":14364},[88],[21,14366,14368],{"className":14367},[60],[21,14369,14371],{"className":14370,"style":96},[64],[21,14372],{},[21,14374],{"className":14375,"style":686},[401],[21,14377,691],{"className":14378},[690],[21,14380],{"className":14381,"style":686},[401],[21,14383,14385,14389,14395,14398,14401,14465,14468,14471,14474,14535,14538,14541,14544,14604,14607],{"className":14384},[33],[21,14386],{"className":14387,"style":14388},[37],"height:1.4744em;vertical-align:-0.4296em;",[21,14390,14392],{"className":14391},[353],[21,14393,4714],{"className":14394},[42,3265],[21,14396,1369],{"className":14397},[221],[21,14399,12294],{"className":14400,"style":12293},[42,46],[21,14402,14404,14407],{"className":14403},[42],[21,14405,713],{"className":14406,"style":712},[42,46],[21,14408,14410],{"className":14409},[51],[21,14411,14413,14456],{"className":14412},[55,56],[21,14414,14416,14453],{"className":14415},[60],[21,14417,14420,14432],{"className":14418,"style":14419},[64],"height:1.0448em;",[21,14421,14423,14426],{"style":14422},"top:-2.4065em;margin-left:-0.1389em;margin-right:0.05em;",[21,14424],{"className":14425,"style":73},[72],[21,14427,14429],{"className":14428},[77,78,79,80],[21,14430,739],{"className":14431},[42,46,80],[21,14433,14435,14438],{"style":14434},"top:-3.2198em;margin-right:0.05em;",[21,14436],{"className":14437,"style":73},[72],[21,14439,14441],{"className":14440},[77,78,79,80],[21,14442,14444,14447,14450],{"className":14443},[42,80],[21,14445,1369],{"className":14446},[221,80],[21,14448,3798],{"className":14449},[42,46,80],[21,14451,1388],{"className":14452},[273,80],[21,14454,89],{"className":14455},[88],[21,14457,14459],{"className":14458},[60],[21,14460,14463],{"className":14461,"style":14462},[64],"height:0.4296em;",[21,14464],{},[21,14466,1378],{"className":14467},[1377],[21,14469],{"className":14470,"style":402},[401],[21,14472,12294],{"className":14473,"style":12293},[42,46],[21,14475,14477,14480],{"className":14476},[42],[21,14478,713],{"className":14479,"style":712},[42,46],[21,14481,14483],{"className":14482},[51],[21,14484,14486,14526],{"className":14485},[55,56],[21,14487,14489,14523],{"className":14488},[60],[21,14490,14492,14503],{"className":14491,"style":14419},[64],[21,14493,14494,14497],{"style":14422},[21,14495],{"className":14496,"style":73},[72],[21,14498,14500],{"className":14499},[77,78,79,80],[21,14501,920],{"className":14502,"style":919},[42,46,80],[21,14504,14505,14508],{"style":14434},[21,14506],{"className":14507,"style":73},[72],[21,14509,14511],{"className":14510},[77,78,79,80],[21,14512,14514,14517,14520],{"className":14513},[42,80],[21,14515,1369],{"className":14516},[221,80],[21,14518,3798],{"className":14519},[42,46,80],[21,14521,1388],{"className":14522},[273,80],[21,14524,89],{"className":14525},[88],[21,14527,14529],{"className":14528},[60],[21,14530,14533],{"className":14531,"style":14532},[64],"height:0.2935em;",[21,14534],{},[21,14536,1378],{"className":14537},[1377],[21,14539],{"className":14540,"style":402},[401],[21,14542,12294],{"className":14543,"style":12293},[42,46],[21,14545,14547,14550],{"className":14546},[42],[21,14548,713],{"className":14549,"style":712},[42,46],[21,14551,14553],{"className":14552},[51],[21,14554,14556,14596],{"className":14555},[55,56],[21,14557,14559,14593],{"className":14558},[60],[21,14560,14562,14573],{"className":14561,"style":14419},[64],[21,14563,14564,14567],{"style":14422},[21,14565],{"className":14566,"style":73},[72],[21,14568,14570],{"className":14569},[77,78,79,80],[21,14571,1088],{"className":14572,"style":1087},[42,46,80],[21,14574,14575,14578],{"style":14434},[21,14576],{"className":14577,"style":73},[72],[21,14579,14581],{"className":14580},[77,78,79,80],[21,14582,14584,14587,14590],{"className":14583},[42,80],[21,14585,1369],{"className":14586},[221,80],[21,14588,3798],{"className":14589},[42,46,80],[21,14591,1388],{"className":14592},[273,80],[21,14594,89],{"className":14595},[88],[21,14597,14599],{"className":14598},[60],[21,14600,14602],{"className":14601,"style":14532},[64],[21,14603],{},[21,14605,1388],{"className":14606},[273],[21,14608,1378],{"className":14609},[1377],[21,14611,14613],{"className":14612},[1582],[21,14614,14616],{"className":14615},[24],[21,14617,14619,14650],{"className":14618,"ariaHidden":29},[28],[21,14620,14622,14625,14632,14635,14638,14641,14644,14647],{"className":14621},[33],[21,14623],{"className":14624,"style":1365},[37],[21,14626,14628],{"className":14627},[353],[21,14629,14631],{"className":14630},[42,3265],"MHA",[21,14633,1369],{"className":14634},[221],[21,14636,12294],{"className":14637,"style":12293},[42,46],[21,14639,1388],{"className":14640},[273],[21,14642],{"className":14643,"style":686},[401],[21,14645,691],{"className":14646},[690],[21,14648],{"className":14649,"style":686},[401],[21,14651,14653,14656,14663,14666,14709,14712,14715,14718,14721,14724,14727,14770,14773,14776,14816],{"className":14652},[33],[21,14654],{"className":14655,"style":1365},[37],[21,14657,14659],{"className":14658},[353],[21,14660,14662],{"className":14661},[42,3265],"Concat",[21,14664,1369],{"className":14665},[221],[21,14667,14669,14675],{"className":14668},[353],[21,14670,14672],{"className":14671},[353],[21,14673,14337],{"className":14674},[42,3265],[21,14676,14678],{"className":14677},[51],[21,14679,14681,14701],{"className":14680},[55,56],[21,14682,14684,14698],{"className":14683},[60],[21,14685,14687],{"className":14686,"style":5491},[64],[21,14688,14689,14692],{"style":14352},[21,14690],{"className":14691,"style":73},[72],[21,14693,14695],{"className":14694},[77,78,79,80],[21,14696,116],{"className":14697},[42,80],[21,14699,89],{"className":14700},[88],[21,14702,14704],{"className":14703},[60],[21,14705,14707],{"className":14706,"style":96},[64],[21,14708],{},[21,14710,1378],{"className":14711},[1377],[21,14713],{"className":14714,"style":402},[401],[21,14716,5523],{"className":14717},[2521],[21,14719],{"className":14720,"style":402},[401],[21,14722,1378],{"className":14723},[1377],[21,14725],{"className":14726,"style":402},[401],[21,14728,14730,14736],{"className":14729},[353],[21,14731,14733],{"className":14732},[353],[21,14734,14337],{"className":14735},[42,3265],[21,14737,14739],{"className":14738},[51],[21,14740,14742,14762],{"className":14741},[55,56],[21,14743,14745,14759],{"className":14744},[60],[21,14746,14748],{"className":14747,"style":2424},[64],[21,14749,14750,14753],{"style":14352},[21,14751],{"className":14752,"style":73},[72],[21,14754,14756],{"className":14755},[77,78,79,80],[21,14757,47],{"className":14758},[42,46,80],[21,14760,89],{"className":14761},[88],[21,14763,14765],{"className":14764},[60],[21,14766,14768],{"className":14767,"style":96},[64],[21,14769],{},[21,14771,1388],{"className":14772},[273],[21,14774],{"className":14775,"style":402},[401],[21,14777,14779,14782],{"className":14778},[42],[21,14780,713],{"className":14781,"style":712},[42,46],[21,14783,14785],{"className":14784},[51],[21,14786,14788,14808],{"className":14787},[55,56],[21,14789,14791,14805],{"className":14790},[60],[21,14792,14794],{"className":14793,"style":726},[64],[21,14795,14796,14799],{"style":729},[21,14797],{"className":14798,"style":73},[72],[21,14800,14802],{"className":14801},[77,78,79,80],[21,14803,11577],{"className":14804,"style":11576},[42,46,80],[21,14806,89],{"className":14807},[88],[21,14809,14811],{"className":14810},[60],[21,14812,14814],{"className":14813,"style":96},[64],[21,14815],{},[21,14817,2126],{"className":14818},[42],[11,14820,14821,14822,14898,14899,15004],{},"Each head works in a ",[21,14823,14825],{"className":14824},[24],[21,14826,14828,14883],{"className":14827,"ariaHidden":29},[28],[21,14829,14831,14834,14874,14877,14880],{"className":14830},[33],[21,14832],{"className":14833,"style":38},[37],[21,14835,14837,14840],{"className":14836},[42],[21,14838,804],{"className":14839},[42,46],[21,14841,14843],{"className":14842},[51],[21,14844,14846,14866],{"className":14845},[55,56],[21,14847,14849,14863],{"className":14848},[60],[21,14850,14852],{"className":14851,"style":2424},[64],[21,14853,14854,14857],{"style":68},[21,14855],{"className":14856,"style":73},[72],[21,14858,14860],{"className":14859},[77,78,79,80],[21,14861,230],{"className":14862,"style":229},[42,46,80],[21,14864,89],{"className":14865},[88],[21,14867,14869],{"className":14868},[60],[21,14870,14872],{"className":14871,"style":96},[64],[21,14873],{},[21,14875],{"className":14876,"style":686},[401],[21,14878,691],{"className":14879},[690],[21,14881],{"className":14882,"style":686},[401],[21,14884,14886,14889,14892,14895],{"className":14885},[33],[21,14887],{"className":14888,"style":1365},[37],[21,14890,804],{"className":14891},[42,46],[21,14893,8240],{"className":14894},[42],[21,14896,47],{"className":14897},[42,46]," dimensional subspace, so the total compute\nmatches a single full-width head; ",[21,14900,14902],{"className":14901},[24],[21,14903,14905,14960],{"className":14904,"ariaHidden":29},[28],[21,14906,14908,14911,14951,14954,14957],{"className":14907},[33],[21,14909],{"className":14910,"style":886},[37],[21,14912,14914,14917],{"className":14913},[42],[21,14915,713],{"className":14916,"style":712},[42,46],[21,14918,14920],{"className":14919},[51],[21,14921,14923,14943],{"className":14922},[55,56],[21,14924,14926,14940],{"className":14925},[60],[21,14927,14929],{"className":14928,"style":726},[64],[21,14930,14931,14934],{"style":729},[21,14932],{"className":14933,"style":73},[72],[21,14935,14937],{"className":14936},[77,78,79,80],[21,14938,11577],{"className":14939,"style":11576},[42,46,80],[21,14941,89],{"className":14942},[88],[21,14944,14946],{"className":14945},[60],[21,14947,14949],{"className":14948,"style":96},[64],[21,14950],{},[21,14952],{"className":14953,"style":686},[401],[21,14955,3730],{"className":14956},[690],[21,14958],{"className":14959,"style":686},[401],[21,14961,14963,14966],{"className":14962},[33],[21,14964],{"className":14965,"style":765},[37],[21,14967,14969,14972],{"className":14968},[42],[21,14970,773],{"className":14971},[42,772],[21,14973,14975],{"className":14974},[51],[21,14976,14978],{"className":14977},[55],[21,14979,14981],{"className":14980},[60],[21,14982,14984],{"className":14983,"style":765},[64],[21,14985,14986,14989],{"style":788},[21,14987],{"className":14988,"style":73},[72],[21,14990,14992],{"className":14991},[77,78,79,80],[21,14993,14995,14998,15001],{"className":14994},[42,80],[21,14996,804],{"className":14997},[42,46,80],[21,14999,2550],{"className":15000},[1218,80],[21,15002,804],{"className":15003},[42,46,80]," recombines\nthe heads.",[2128,15006],{"hash":15007},"f1304c22c8b585b2a338c2145110fcb5c41434dba20c62581f1d8aa3db154a18",[11,15009,15010,15011,5614,15088,15103,15104,15235,15236,15270,15271,15305,15306,15322,15323,15338,15339,15603,15604,1378],{},"The shapes are worth tracing once.\nFix ",[21,15012,15014],{"className":15013},[24],[21,15015,15017,15035],{"className":15016,"ariaHidden":29},[28],[21,15018,15020,15023,15026,15029,15032],{"className":15019},[33],[21,15021],{"className":15022,"style":867},[37],[21,15024,804],{"className":15025},[42,46],[21,15027],{"className":15028,"style":686},[401],[21,15030,691],{"className":15031},[690],[21,15033],{"className":15034,"style":686},[401],[21,15036,15038,15041],{"className":15037},[33],[21,15039],{"className":15040,"style":38},[37],[21,15042,15044,15047],{"className":15043},[42],[21,15045,804],{"className":15046},[42,46],[21,15048,15050],{"className":15049},[51],[21,15051,15053,15080],{"className":15052},[55,56],[21,15054,15056,15077],{"className":15055},[60],[21,15057,15059],{"className":15058,"style":2424},[64],[21,15060,15061,15064],{"style":68},[21,15062],{"className":15063,"style":73},[72],[21,15065,15067],{"className":15066},[77,78,79,80],[21,15068,15070],{"className":15069},[42,80],[21,15071,15073],{"className":15072},[42,1965,80],[21,15074,15076],{"className":15075},[42,80],"model",[21,15078,89],{"className":15079},[88],[21,15081,15083],{"className":15082},[60],[21,15084,15086],{"className":15085,"style":96},[64],[21,15087],{},[21,15089,15091],{"className":15090},[24],[21,15092,15094],{"className":15093,"ariaHidden":29},[28],[21,15095,15097,15100],{"className":15096},[33],[21,15098],{"className":15099,"style":867},[37],[21,15101,47],{"className":15102},[42,46]," heads, and set the per-head width ",[21,15105,15107],{"className":15106},[24],[21,15108,15110,15165,15220],{"className":15109,"ariaHidden":29},[28],[21,15111,15113,15116,15156,15159,15162],{"className":15112},[33],[21,15114],{"className":15115,"style":38},[37],[21,15117,15119,15122],{"className":15118},[42],[21,15120,804],{"className":15121},[42,46],[21,15123,15125],{"className":15124},[51],[21,15126,15128,15148],{"className":15127},[55,56],[21,15129,15131,15145],{"className":15130},[60],[21,15132,15134],{"className":15133,"style":2424},[64],[21,15135,15136,15139],{"style":68},[21,15137],{"className":15138,"style":73},[72],[21,15140,15142],{"className":15141},[77,78,79,80],[21,15143,230],{"className":15144,"style":229},[42,46,80],[21,15146,89],{"className":15147},[88],[21,15149,15151],{"className":15150},[60],[21,15152,15154],{"className":15153,"style":96},[64],[21,15155],{},[21,15157],{"className":15158,"style":686},[401],[21,15160,691],{"className":15161},[690],[21,15163],{"className":15164,"style":686},[401],[21,15166,15168,15171,15211,15214,15217],{"className":15167},[33],[21,15169],{"className":15170,"style":38},[37],[21,15172,15174,15177],{"className":15173},[42],[21,15175,804],{"className":15176},[42,46],[21,15178,15180],{"className":15179},[51],[21,15181,15183,15203],{"className":15182},[55,56],[21,15184,15186,15200],{"className":15185},[60],[21,15187,15189],{"className":15188,"style":2478},[64],[21,15190,15191,15194],{"style":68},[21,15192],{"className":15193,"style":73},[72],[21,15195,15197],{"className":15196},[77,78,79,80],[21,15198,297],{"className":15199,"style":202},[42,46,80],[21,15201,89],{"className":15202},[88],[21,15204,15206],{"className":15205},[60],[21,15207,15209],{"className":15208,"style":96},[64],[21,15210],{},[21,15212],{"className":15213,"style":686},[401],[21,15215,691],{"className":15216},[690],[21,15218],{"className":15219,"style":686},[401],[21,15221,15223,15226,15229,15232],{"className":15222},[33],[21,15224],{"className":15225,"style":1365},[37],[21,15227,804],{"className":15228},[42,46],[21,15230,8240],{"className":15231},[42],[21,15233,47],{"className":15234},[42,46]," (so ",[21,15237,15239],{"className":15238},[24],[21,15240,15242,15260],{"className":15241,"ariaHidden":29},[28],[21,15243,15245,15248,15251,15254,15257],{"className":15244},[33],[21,15246],{"className":15247,"style":867},[37],[21,15249,47],{"className":15250},[42,46],[21,15252],{"className":15253,"style":686},[401],[21,15255,691],{"className":15256},[690],[21,15258],{"className":15259,"style":686},[401],[21,15261,15263,15266],{"className":15262},[33],[21,15264],{"className":15265,"style":112},[37],[21,15267,15269],{"className":15268},[42],"8"," heads of a ",[21,15272,15274],{"className":15273},[24],[21,15275,15277,15295],{"className":15276,"ariaHidden":29},[28],[21,15278,15280,15283,15286,15289,15292],{"className":15279},[33],[21,15281],{"className":15282,"style":867},[37],[21,15284,804],{"className":15285},[42,46],[21,15287],{"className":15288,"style":686},[401],[21,15290,691],{"className":15291},[690],[21,15293],{"className":15294,"style":686},[401],[21,15296,15298,15301],{"className":15297},[33],[21,15299],{"className":15300,"style":112},[37],[21,15302,15304],{"className":15303},[42],"512"," model each work in ",[21,15307,15309],{"className":15308},[24],[21,15310,15312],{"className":15311,"ariaHidden":29},[28],[21,15313,15315,15318],{"className":15314},[33],[21,15316],{"className":15317,"style":112},[37],[21,15319,15321],{"className":15320},[42],"64"," dimensions).\nFor head ",[21,15324,15326],{"className":15325},[24],[21,15327,15329],{"className":15328,"ariaHidden":29},[28],[21,15330,15332,15335],{"className":15331},[33],[21,15333],{"className":15334,"style":3932},[37],[21,15336,3798],{"className":15337},[42,46]," the projections are ",[21,15340,15342],{"className":15341},[24],[21,15343,15345,15552],{"className":15344,"ariaHidden":29},[28],[21,15346,15348,15351,15411,15414,15417,15477,15480,15483,15543,15546,15549],{"className":15347},[33],[21,15349],{"className":15350,"style":14388},[37],[21,15352,15354,15357],{"className":15353},[42],[21,15355,713],{"className":15356,"style":712},[42,46],[21,15358,15360],{"className":15359},[51],[21,15361,15363,15403],{"className":15362},[55,56],[21,15364,15366,15400],{"className":15365},[60],[21,15367,15369,15380],{"className":15368,"style":14419},[64],[21,15370,15371,15374],{"style":14422},[21,15372],{"className":15373,"style":73},[72],[21,15375,15377],{"className":15376},[77,78,79,80],[21,15378,739],{"className":15379},[42,46,80],[21,15381,15382,15385],{"style":14434},[21,15383],{"className":15384,"style":73},[72],[21,15386,15388],{"className":15387},[77,78,79,80],[21,15389,15391,15394,15397],{"className":15390},[42,80],[21,15392,1369],{"className":15393},[221,80],[21,15395,3798],{"className":15396},[42,46,80],[21,15398,1388],{"className":15399},[273,80],[21,15401,89],{"className":15402},[88],[21,15404,15406],{"className":15405},[60],[21,15407,15409],{"className":15408,"style":14462},[64],[21,15410],{},[21,15412,1378],{"className":15413},[1377],[21,15415],{"className":15416,"style":402},[401],[21,15418,15420,15423],{"className":15419},[42],[21,15421,713],{"className":15422,"style":712},[42,46],[21,15424,15426],{"className":15425},[51],[21,15427,15429,15469],{"className":15428},[55,56],[21,15430,15432,15466],{"className":15431},[60],[21,15433,15435,15446],{"className":15434,"style":14419},[64],[21,15436,15437,15440],{"style":14422},[21,15438],{"className":15439,"style":73},[72],[21,15441,15443],{"className":15442},[77,78,79,80],[21,15444,920],{"className":15445,"style":919},[42,46,80],[21,15447,15448,15451],{"style":14434},[21,15449],{"className":15450,"style":73},[72],[21,15452,15454],{"className":15453},[77,78,79,80],[21,15455,15457,15460,15463],{"className":15456},[42,80],[21,15458,1369],{"className":15459},[221,80],[21,15461,3798],{"className":15462},[42,46,80],[21,15464,1388],{"className":15465},[273,80],[21,15467,89],{"className":15468},[88],[21,15470,15472],{"className":15471},[60],[21,15473,15475],{"className":15474,"style":14532},[64],[21,15476],{},[21,15478,1378],{"className":15479},[1377],[21,15481],{"className":15482,"style":402},[401],[21,15484,15486,15489],{"className":15485},[42],[21,15487,713],{"className":15488,"style":712},[42,46],[21,15490,15492],{"className":15491},[51],[21,15493,15495,15535],{"className":15494},[55,56],[21,15496,15498,15532],{"className":15497},[60],[21,15499,15501,15512],{"className":15500,"style":14419},[64],[21,15502,15503,15506],{"style":14422},[21,15504],{"className":15505,"style":73},[72],[21,15507,15509],{"className":15508},[77,78,79,80],[21,15510,1088],{"className":15511,"style":1087},[42,46,80],[21,15513,15514,15517],{"style":14434},[21,15515],{"className":15516,"style":73},[72],[21,15518,15520],{"className":15519},[77,78,79,80],[21,15521,15523,15526,15529],{"className":15522},[42,80],[21,15524,1369],{"className":15525},[221,80],[21,15527,3798],{"className":15528},[42,46,80],[21,15530,1388],{"className":15531},[273,80],[21,15533,89],{"className":15534},[88],[21,15536,15538],{"className":15537},[60],[21,15539,15541],{"className":15540,"style":14532},[64],[21,15542],{},[21,15544],{"className":15545,"style":686},[401],[21,15547,3730],{"className":15548},[690],[21,15550],{"className":15551,"style":686},[401],[21,15553,15555,15559],{"className":15554},[33],[21,15556],{"className":15557,"style":15558},[37],"height:0.888em;",[21,15560,15562,15565],{"className":15561},[42],[21,15563,773],{"className":15564},[42,772],[21,15566,15568],{"className":15567},[51],[21,15569,15571],{"className":15570},[55],[21,15572,15574],{"className":15573},[60],[21,15575,15577],{"className":15576,"style":15558},[64],[21,15578,15579,15582],{"style":788},[21,15580],{"className":15581,"style":73},[72],[21,15583,15585],{"className":15584},[77,78,79,80],[21,15586,15588,15591,15594,15597,15600],{"className":15587},[42,80],[21,15589,804],{"className":15590},[42,46,80],[21,15592,2550],{"className":15593},[1218,80],[21,15595,804],{"className":15596},[42,46,80],[21,15598,8240],{"className":15599},[42,80],[21,15601,47],{"className":15602},[42,46,80],", so from an input ",[21,15605,15607],{"className":15606},[24],[21,15608,15610,15628],{"className":15609,"ariaHidden":29},[28],[21,15611,15613,15616,15619,15622,15625],{"className":15612},[33],[21,15614],{"className":15615,"style":4493},[37],[21,15617,12294],{"className":15618,"style":12293},[42,46],[21,15620],{"className":15621,"style":686},[401],[21,15623,3730],{"className":15624},[690],[21,15626],{"className":15627,"style":686},[401],[21,15629,15631,15634],{"className":15630},[33],[21,15632],{"className":15633,"style":765},[37],[21,15635,15637,15640],{"className":15636},[42],[21,15638,773],{"className":15639},[42,772],[21,15641,15643],{"className":15642},[51],[21,15644,15646],{"className":15645},[55],[21,15647,15649],{"className":15648},[60],[21,15650,15652],{"className":15651,"style":765},[64],[21,15653,15654,15657],{"style":788},[21,15655],{"className":15656,"style":73},[72],[21,15658,15660],{"className":15659},[77,78,79,80],[21,15661,15663,15666,15669],{"className":15662},[42,80],[21,15664,134],{"className":15665},[42,46,80],[21,15667,2550],{"className":15668},[1218,80],[21,15670,804],{"className":15671},[42,46,80],[21,15673,15675],{"className":15674},[1582],[21,15676,15678],{"className":15677},[24],[21,15679,15681,16261],{"className":15680,"ariaHidden":29},[28],[21,15682,15684,15688,15872,15876,15879,16061,16064,16067,16245,16248,16251,16255,16258],{"className":15683},[33],[21,15685],{"className":15686,"style":15687},[37],"height:3.0224em;vertical-align:-1.9776em;",[21,15689,15691],{"className":15690},[2521,2522],[21,15692,15694,15863],{"className":15693},[55,56],[21,15695,15697,15860],{"className":15696},[60],[21,15698,15700,15728],{"className":15699,"style":14419},[64],[21,15701,15703,15707],{"style":15702},"top:-1.2422em;",[21,15704],{"className":15705,"style":15706},[72],"height:3.0448em;",[21,15708,15710],{"className":15709},[77,78,79,80],[21,15711,15713,15716,15719,15722,15725],{"className":15712},[42,80],[21,15714,134],{"className":15715},[42,46,80],[21,15717,2550],{"className":15718},[1218,80],[21,15720,804],{"className":15721},[42,46,80],[21,15723,8240],{"className":15724},[42,80],[21,15726,47],{"className":15727},[42,46,80],[21,15729,15731,15734],{"style":15730},"top:-3.0448em;",[21,15732],{"className":15733,"style":15706},[72],[21,15735,15737],{"className":15736},[2521,2522],[21,15738,15740,15851],{"className":15739},[55,56],[21,15741,15743,15848],{"className":15742},[60],[21,15744,15746,15777],{"className":15745,"style":14419},[64],[21,15747,15750,15753],{"className":15748,"style":15749},[2612],"top:-1.9672em;",[21,15751],{"className":15752,"style":15706},[72],[21,15754,15756,15763,15770],{"className":15755,"style":2621},[2620],[21,15757,15759],{"className":15758,"style":2626},[2625],[2628,15760,15761],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},[2636,15762],{"d":2638},[21,15764,15766],{"className":15765,"style":2626},[2642],[2628,15767,15768],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},[2636,15769],{"d":2648},[21,15771,15773],{"className":15772,"style":2626},[2652],[2628,15774,15775],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},[2636,15776],{"d":2658},[21,15778,15779,15782],{"style":15730},[21,15780],{"className":15781,"style":15706},[72],[21,15783,15785,15788],{"className":15784},[42],[21,15786,12294],{"className":15787,"style":12293},[42,46],[21,15789,15791,15794],{"className":15790},[42],[21,15792,713],{"className":15793,"style":712},[42,46],[21,15795,15797],{"className":15796},[51],[21,15798,15800,15840],{"className":15799},[55,56],[21,15801,15803,15837],{"className":15802},[60],[21,15804,15806,15817],{"className":15805,"style":14419},[64],[21,15807,15808,15811],{"style":14422},[21,15809],{"className":15810,"style":73},[72],[21,15812,15814],{"className":15813},[77,78,79,80],[21,15815,739],{"className":15816},[42,46,80],[21,15818,15819,15822],{"style":14434},[21,15820],{"className":15821,"style":73},[72],[21,15823,15825],{"className":15824},[77,78,79,80],[21,15826,15828,15831,15834],{"className":15827},[42,80],[21,15829,1369],{"className":15830},[221,80],[21,15832,3798],{"className":15833},[42,46,80],[21,15835,1388],{"className":15836},[273,80],[21,15838,89],{"className":15839},[88],[21,15841,15843],{"className":15842},[60],[21,15844,15846],{"className":15845,"style":14462},[64],[21,15847],{},[21,15849,89],{"className":15850},[88],[21,15852,15854],{"className":15853},[60],[21,15855,15858],{"className":15856,"style":15857},[64],"height:1.0776em;",[21,15859],{},[21,15861,89],{"className":15862},[88],[21,15864,15866],{"className":15865},[60],[21,15867,15870],{"className":15868,"style":15869},[64],"height:1.9776em;",[21,15871],{},[21,15873],{"className":15874,"style":15875},[401],"margin-right:1em;",[21,15877],{"className":15878,"style":402},[401],[21,15880,15882],{"className":15881},[2521,2522],[21,15883,15885,16052],{"className":15884},[55,56],[21,15886,15888,16049],{"className":15887},[60],[21,15889,15891,15918],{"className":15890,"style":14419},[64],[21,15892,15894,15897],{"style":15893},"top:-1.3783em;",[21,15895],{"className":15896,"style":15706},[72],[21,15898,15900],{"className":15899},[77,78,79,80],[21,15901,15903,15906,15909,15912,15915],{"className":15902},[42,80],[21,15904,134],{"className":15905},[42,46,80],[21,15907,2550],{"className":15908},[1218,80],[21,15910,804],{"className":15911},[42,46,80],[21,15913,8240],{"className":15914},[42,80],[21,15916,47],{"className":15917},[42,46,80],[21,15919,15920,15923],{"style":15730},[21,15921],{"className":15922,"style":15706},[72],[21,15924,15926],{"className":15925},[2521,2522],[21,15927,15929,16040],{"className":15928},[55,56],[21,15930,15932,16037],{"className":15931},[60],[21,15933,15935,15966],{"className":15934,"style":14419},[64],[21,15936,15939,15942],{"className":15937,"style":15938},[2612],"top:-2.1033em;",[21,15940],{"className":15941,"style":15706},[72],[21,15943,15945,15952,15959],{"className":15944,"style":2621},[2620],[21,15946,15948],{"className":15947,"style":2626},[2625],[2628,15949,15950],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},[2636,15951],{"d":2638},[21,15953,15955],{"className":15954,"style":2626},[2642],[2628,15956,15957],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},[2636,15958],{"d":2648},[21,15960,15962],{"className":15961,"style":2626},[2652],[2628,15963,15964],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},[2636,15965],{"d":2658},[21,15967,15968,15971],{"style":15730},[21,15969],{"className":15970,"style":15706},[72],[21,15972,15974,15977],{"className":15973},[42],[21,15975,12294],{"className":15976,"style":12293},[42,46],[21,15978,15980,15983],{"className":15979},[42],[21,15981,713],{"className":15982,"style":712},[42,46],[21,15984,15986],{"className":15985},[51],[21,15987,15989,16029],{"className":15988},[55,56],[21,15990,15992,16026],{"className":15991},[60],[21,15993,15995,16006],{"className":15994,"style":14419},[64],[21,15996,15997,16000],{"style":14422},[21,15998],{"className":15999,"style":73},[72],[21,16001,16003],{"className":16002},[77,78,79,80],[21,16004,920],{"className":16005,"style":919},[42,46,80],[21,16007,16008,16011],{"style":14434},[21,16009],{"className":16010,"style":73},[72],[21,16012,16014],{"className":16013},[77,78,79,80],[21,16015,16017,16020,16023],{"className":16016},[42,80],[21,16018,1369],{"className":16019},[221,80],[21,16021,3798],{"className":16022},[42,46,80],[21,16024,1388],{"className":16025},[273,80],[21,16027,89],{"className":16028},[88],[21,16030,16032],{"className":16031},[60],[21,16033,16035],{"className":16034,"style":14532},[64],[21,16036],{},[21,16038,89],{"className":16039},[88],[21,16041,16043],{"className":16042},[60],[21,16044,16047],{"className":16045,"style":16046},[64],"height:0.9415em;",[21,16048],{},[21,16050,89],{"className":16051},[88],[21,16053,16055],{"className":16054},[60],[21,16056,16059],{"className":16057,"style":16058},[64],"height:1.8415em;",[21,16060],{},[21,16062],{"className":16063,"style":15875},[401],[21,16065],{"className":16066,"style":402},[401],[21,16068,16070],{"className":16069},[2521,2522],[21,16071,16073,16237],{"className":16072},[55,56],[21,16074,16076,16234],{"className":16075},[60],[21,16077,16079,16105],{"className":16078,"style":14419},[64],[21,16080,16081,16084],{"style":15893},[21,16082],{"className":16083,"style":15706},[72],[21,16085,16087],{"className":16086},[77,78,79,80],[21,16088,16090,16093,16096,16099,16102],{"className":16089},[42,80],[21,16091,134],{"className":16092},[42,46,80],[21,16094,2550],{"className":16095},[1218,80],[21,16097,804],{"className":16098},[42,46,80],[21,16100,8240],{"className":16101},[42,80],[21,16103,47],{"className":16104},[42,46,80],[21,16106,16107,16110],{"style":15730},[21,16108],{"className":16109,"style":15706},[72],[21,16111,16113],{"className":16112},[2521,2522],[21,16114,16116,16226],{"className":16115},[55,56],[21,16117,16119,16223],{"className":16118},[60],[21,16120,16122,16152],{"className":16121,"style":14419},[64],[21,16123,16125,16128],{"className":16124,"style":15938},[2612],[21,16126],{"className":16127,"style":15706},[72],[21,16129,16131,16138,16145],{"className":16130,"style":2621},[2620],[21,16132,16134],{"className":16133,"style":2626},[2625],[2628,16135,16136],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},[2636,16137],{"d":2638},[21,16139,16141],{"className":16140,"style":2626},[2642],[2628,16142,16143],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},[2636,16144],{"d":2648},[21,16146,16148],{"className":16147,"style":2626},[2652],[2628,16149,16150],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},[2636,16151],{"d":2658},[21,16153,16154,16157],{"style":15730},[21,16155],{"className":16156,"style":15706},[72],[21,16158,16160,16163],{"className":16159},[42],[21,16161,12294],{"className":16162,"style":12293},[42,46],[21,16164,16166,16169],{"className":16165},[42],[21,16167,713],{"className":16168,"style":712},[42,46],[21,16170,16172],{"className":16171},[51],[21,16173,16175,16215],{"className":16174},[55,56],[21,16176,16178,16212],{"className":16177},[60],[21,16179,16181,16192],{"className":16180,"style":14419},[64],[21,16182,16183,16186],{"style":14422},[21,16184],{"className":16185,"style":73},[72],[21,16187,16189],{"className":16188},[77,78,79,80],[21,16190,1088],{"className":16191,"style":1087},[42,46,80],[21,16193,16194,16197],{"style":14434},[21,16195],{"className":16196,"style":73},[72],[21,16198,16200],{"className":16199},[77,78,79,80],[21,16201,16203,16206,16209],{"className":16202},[42,80],[21,16204,1369],{"className":16205},[221,80],[21,16207,3798],{"className":16208},[42,46,80],[21,16210,1388],{"className":16211},[273,80],[21,16213,89],{"className":16214},[88],[21,16216,16218],{"className":16217},[60],[21,16219,16221],{"className":16220,"style":14532},[64],[21,16222],{},[21,16224,89],{"className":16225},[88],[21,16227,16229],{"className":16228},[60],[21,16230,16232],{"className":16231,"style":16046},[64],[21,16233],{},[21,16235,89],{"className":16236},[88],[21,16238,16240],{"className":16239},[60],[21,16241,16243],{"className":16242,"style":16058},[64],[21,16244],{},[21,16246],{"className":16247,"style":686},[401],[21,16249],{"className":16250,"style":686},[401],[21,16252,16254],{"className":16253},[690],"⟹",[21,16256],{"className":16257,"style":686},[401],[21,16259],{"className":16260,"style":686},[401],[21,16262,16264,16268,16430,16433],{"className":16263},[33],[21,16265],{"className":16266,"style":16267},[37],"height:2.3924em;vertical-align:-1.698em;",[21,16269,16271],{"className":16270},[2521,2522],[21,16272,16274,16421],{"className":16273},[55,56],[21,16275,16277,16418],{"className":16276},[60],[21,16278,16280,16307],{"className":16279,"style":867},[64],[21,16281,16283,16286],{"style":16282},"top:-1.477em;",[21,16284],{"className":16285,"style":1678},[72],[21,16287,16289],{"className":16288},[77,78,79,80],[21,16290,16292,16295,16298,16301,16304],{"className":16291},[42,80],[21,16293,134],{"className":16294},[42,46,80],[21,16296,2550],{"className":16297},[1218,80],[21,16299,804],{"className":16300},[42,46,80],[21,16302,8240],{"className":16303},[42,80],[21,16305,47],{"className":16306},[42,46,80],[21,16308,16309,16312],{"style":2593},[21,16310],{"className":16311,"style":1678},[72],[21,16313,16315],{"className":16314},[2521,2522],[21,16316,16318,16409],{"className":16317},[55,56],[21,16319,16321,16406],{"className":16320},[60],[21,16322,16324,16355],{"className":16323,"style":867},[64],[21,16325,16328,16331],{"className":16326,"style":16327},[2612],"top:-2.202em;",[21,16329],{"className":16330,"style":1678},[72],[21,16332,16334,16341,16348],{"className":16333,"style":2621},[2620],[21,16335,16337],{"className":16336,"style":2626},[2625],[2628,16338,16339],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},[2636,16340],{"d":2638},[21,16342,16344],{"className":16343,"style":2626},[2642],[2628,16345,16346],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},[2636,16347],{"d":2648},[21,16349,16351],{"className":16350,"style":2626},[2652],[2628,16352,16353],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},[2636,16354],{"d":2658},[21,16356,16357,16360],{"style":2593},[21,16358],{"className":16359,"style":1678},[72],[21,16361,16363],{"className":16362},[42],[21,16364,16366,16372],{"className":16365},[353],[21,16367,16369],{"className":16368},[353],[21,16370,14337],{"className":16371},[42,3265],[21,16373,16375],{"className":16374},[51],[21,16376,16378,16398],{"className":16377},[55,56],[21,16379,16381,16395],{"className":16380},[60],[21,16382,16384],{"className":16383,"style":243},[64],[21,16385,16386,16389],{"style":14352},[21,16387],{"className":16388,"style":73},[72],[21,16390,16392],{"className":16391},[77,78,79,80],[21,16393,3798],{"className":16394},[42,46,80],[21,16396,89],{"className":16397},[88],[21,16399,16401],{"className":16400},[60],[21,16402,16404],{"className":16403,"style":96},[64],[21,16405],{},[21,16407,89],{"className":16408},[88],[21,16410,16412],{"className":16411},[60],[21,16413,16416],{"className":16414,"style":16415},[64],"height:0.798em;",[21,16417],{},[21,16419,89],{"className":16420},[88],[21,16422,16424],{"className":16423},[60],[21,16425,16428],{"className":16426,"style":16427},[64],"height:1.698em;",[21,16429],{},[21,16431],{"className":16432,"style":402},[401],[21,16434,1378],{"className":16435},[1377],[11,16437,16438,16439,16484,16485,16500,16501,16534,16535,16640],{},"each head returns an ",[21,16440,16442],{"className":16441},[24],[21,16443,16445,16463],{"className":16444,"ariaHidden":29},[28],[21,16446,16448,16451,16454,16457,16460],{"className":16447},[33],[21,16449],{"className":16450,"style":4149},[37],[21,16452,134],{"className":16453},[42,46],[21,16455],{"className":16456,"style":1087},[401],[21,16458,2550],{"className":16459},[1218],[21,16461],{"className":16462,"style":1087},[401],[21,16464,16466,16469,16472,16475,16478,16481],{"className":16465},[33],[21,16467],{"className":16468,"style":1365},[37],[21,16470,1369],{"className":16471},[221],[21,16473,804],{"className":16474},[42,46],[21,16476,8240],{"className":16477},[42],[21,16479,47],{"className":16480},[42,46],[21,16482,1388],{"className":16483},[273]," block. Concatenating the ",[21,16486,16488],{"className":16487},[24],[21,16489,16491],{"className":16490,"ariaHidden":29},[28],[21,16492,16494,16497],{"className":16493},[33],[21,16495],{"className":16496,"style":867},[37],[21,16498,47],{"className":16499},[42,46]," blocks along\nthe width rebuilds a full ",[21,16502,16504],{"className":16503},[24],[21,16505,16507,16525],{"className":16506,"ariaHidden":29},[28],[21,16508,16510,16513,16516,16519,16522],{"className":16509},[33],[21,16511],{"className":16512,"style":4149},[37],[21,16514,134],{"className":16515},[42,46],[21,16517],{"className":16518,"style":1087},[401],[21,16520,2550],{"className":16521},[1218],[21,16523],{"className":16524,"style":1087},[401],[21,16526,16528,16531],{"className":16527},[33],[21,16529],{"className":16530,"style":867},[37],[21,16532,804],{"className":16533},[42,46]," matrix, which ",[21,16536,16538],{"className":16537},[24],[21,16539,16541,16596],{"className":16540,"ariaHidden":29},[28],[21,16542,16544,16547,16587,16590,16593],{"className":16543},[33],[21,16545],{"className":16546,"style":886},[37],[21,16548,16550,16553],{"className":16549},[42],[21,16551,713],{"className":16552,"style":712},[42,46],[21,16554,16556],{"className":16555},[51],[21,16557,16559,16579],{"className":16558},[55,56],[21,16560,16562,16576],{"className":16561},[60],[21,16563,16565],{"className":16564,"style":726},[64],[21,16566,16567,16570],{"style":729},[21,16568],{"className":16569,"style":73},[72],[21,16571,16573],{"className":16572},[77,78,79,80],[21,16574,11577],{"className":16575,"style":11576},[42,46,80],[21,16577,89],{"className":16578},[88],[21,16580,16582],{"className":16581},[60],[21,16583,16585],{"className":16584,"style":96},[64],[21,16586],{},[21,16588],{"className":16589,"style":686},[401],[21,16591,3730],{"className":16592},[690],[21,16594],{"className":16595,"style":686},[401],[21,16597,16599,16602],{"className":16598},[33],[21,16600],{"className":16601,"style":765},[37],[21,16603,16605,16608],{"className":16604},[42],[21,16606,773],{"className":16607},[42,772],[21,16609,16611],{"className":16610},[51],[21,16612,16614],{"className":16613},[55],[21,16615,16617],{"className":16616},[60],[21,16618,16620],{"className":16619,"style":765},[64],[21,16621,16622,16625],{"style":788},[21,16623],{"className":16624,"style":73},[72],[21,16626,16628],{"className":16627},[77,78,79,80],[21,16629,16631,16634,16637],{"className":16630},[42,80],[21,16632,804],{"className":16633},[42,46,80],[21,16635,2550],{"className":16636},[1218,80],[21,16638,804],{"className":16639},[42,46,80]," mixes back into the model width:",[21,16642,16644],{"className":16643},[1582],[21,16645,16647],{"className":16646},[24],[21,16648,16650,16902,17075],{"className":16649,"ariaHidden":29},[28],[21,16651,16653,16657,16887,16890,16893,16896,16899],{"className":16652},[33],[21,16654],{"className":16655,"style":16656},[37],"height:2.3924em;vertical-align:-1.6424em;",[21,16658,16660],{"className":16659},[2521,2522],[21,16661,16663,16878],{"className":16662},[55,56],[21,16664,16666,16875],{"className":16665},[60],[21,16667,16670,16691],{"className":16668,"style":16669},[64],"height:0.75em;",[21,16671,16673,16676],{"style":16672},"top:-1.4159em;",[21,16674],{"className":16675,"style":1678},[72],[21,16677,16679],{"className":16678},[77,78,79,80],[21,16680,16682,16685,16688],{"className":16681},[42,80],[21,16683,134],{"className":16684},[42,46,80],[21,16686,2550],{"className":16687},[1218,80],[21,16689,804],{"className":16690},[42,46,80],[21,16692,16693,16696],{"style":2593},[21,16694],{"className":16695,"style":1678},[72],[21,16697,16699],{"className":16698},[2521,2522],[21,16700,16702,16866],{"className":16701},[55,56],[21,16703,16705,16863],{"className":16704},[60],[21,16706,16708,16739],{"className":16707,"style":16669},[64],[21,16709,16712,16715],{"className":16710,"style":16711},[2612],"top:-2.102em;",[21,16713],{"className":16714,"style":1678},[72],[21,16716,16718,16725,16732],{"className":16717,"style":2621},[2620],[21,16719,16721],{"className":16720,"style":2626},[2625],[2628,16722,16723],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},[2636,16724],{"d":2638},[21,16726,16728],{"className":16727,"style":2626},[2642],[2628,16729,16730],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},[2636,16731],{"d":2648},[21,16733,16735],{"className":16734,"style":2626},[2652],[2628,16736,16737],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},[2636,16738],{"d":2658},[21,16740,16741,16744],{"style":2593},[21,16742],{"className":16743,"style":1678},[72],[21,16745,16747,16753,16756,16799,16802,16805,16808,16811,16814,16817,16860],{"className":16746},[42],[21,16748,16750],{"className":16749},[353],[21,16751,14662],{"className":16752},[42,3265],[21,16754,1369],{"className":16755},[221],[21,16757,16759,16765],{"className":16758},[353],[21,16760,16762],{"className":16761},[353],[21,16763,14337],{"className":16764},[42,3265],[21,16766,16768],{"className":16767},[51],[21,16769,16771,16791],{"className":16770},[55,56],[21,16772,16774,16788],{"className":16773},[60],[21,16775,16777],{"className":16776,"style":5491},[64],[21,16778,16779,16782],{"style":14352},[21,16780],{"className":16781,"style":73},[72],[21,16783,16785],{"className":16784},[77,78,79,80],[21,16786,116],{"className":16787},[42,80],[21,16789,89],{"className":16790},[88],[21,16792,16794],{"className":16793},[60],[21,16795,16797],{"className":16796,"style":96},[64],[21,16798],{},[21,16800,1378],{"className":16801},[1377],[21,16803],{"className":16804,"style":402},[401],[21,16806,5523],{"className":16807},[2521],[21,16809],{"className":16810,"style":402},[401],[21,16812,1378],{"className":16813},[1377],[21,16815],{"className":16816,"style":402},[401],[21,16818,16820,16826],{"className":16819},[353],[21,16821,16823],{"className":16822},[353],[21,16824,14337],{"className":16825},[42,3265],[21,16827,16829],{"className":16828},[51],[21,16830,16832,16852],{"className":16831},[55,56],[21,16833,16835,16849],{"className":16834},[60],[21,16836,16838],{"className":16837,"style":2424},[64],[21,16839,16840,16843],{"style":14352},[21,16841],{"className":16842,"style":73},[72],[21,16844,16846],{"className":16845},[77,78,79,80],[21,16847,47],{"className":16848},[42,46,80],[21,16850,89],{"className":16851},[88],[21,16853,16855],{"className":16854},[60],[21,16856,16858],{"className":16857,"style":96},[64],[21,16859],{},[21,16861,1388],{"className":16862},[273],[21,16864,89],{"className":16865},[88],[21,16867,16869],{"className":16868},[60],[21,16870,16873],{"className":16871,"style":16872},[64],"height:0.898em;",[21,16874],{},[21,16876,89],{"className":16877},[88],[21,16879,16881],{"className":16880},[60],[21,16882,16885],{"className":16883,"style":16884},[64],"height:1.6424em;",[21,16886],{},[21,16888],{"className":16889,"style":686},[401],[21,16891],{"className":16892,"style":1087},[401],[21,16894,1219],{"className":16895},[1218],[21,16897],{"className":16898,"style":686},[401],[21,16900],{"className":16901,"style":1087},[401],[21,16903,16905,16909,17060,17063,17066,17069,17072],{"className":16904},[33],[21,16906],{"className":16907,"style":16908},[37],"height:2.2258em;vertical-align:-1.5424em;",[21,16910,16912],{"className":16911},[2521,2522],[21,16913,16915,17051],{"className":16914},[55,56],[21,16916,16918,17048],{"className":16917},[60],[21,16919,16921,16942],{"className":16920,"style":2200},[64],[21,16922,16924,16927],{"style":16923},"top:-1.5159em;",[21,16925],{"className":16926,"style":1678},[72],[21,16928,16930],{"className":16929},[77,78,79,80],[21,16931,16933,16936,16939],{"className":16932},[42,80],[21,16934,804],{"className":16935},[42,46,80],[21,16937,2550],{"className":16938},[1218,80],[21,16940,804],{"className":16941},[42,46,80],[21,16943,16944,16947],{"style":2593},[21,16945],{"className":16946,"style":1678},[72],[21,16948,16950],{"className":16949},[2521,2522],[21,16951,16953,17040],{"className":16952},[55,56],[21,16954,16956,17037],{"className":16955},[60],[21,16957,16959,16989],{"className":16958,"style":2200},[64],[21,16960,16962,16965],{"className":16961,"style":16327},[2612],[21,16963],{"className":16964,"style":1678},[72],[21,16966,16968,16975,16982],{"className":16967,"style":2621},[2620],[21,16969,16971],{"className":16970,"style":2626},[2625],[2628,16972,16973],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},[2636,16974],{"d":2638},[21,16976,16978],{"className":16977,"style":2626},[2642],[2628,16979,16980],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},[2636,16981],{"d":2648},[21,16983,16985],{"className":16984,"style":2626},[2652],[2628,16986,16987],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},[2636,16988],{"d":2658},[21,16990,16991,16994],{"style":2593},[21,16992],{"className":16993,"style":1678},[72],[21,16995,16997],{"className":16996},[42],[21,16998,17000,17003],{"className":16999},[42],[21,17001,713],{"className":17002,"style":712},[42,46],[21,17004,17006],{"className":17005},[51],[21,17007,17009,17029],{"className":17008},[55,56],[21,17010,17012,17026],{"className":17011},[60],[21,17013,17015],{"className":17014,"style":726},[64],[21,17016,17017,17020],{"style":729},[21,17018],{"className":17019,"style":73},[72],[21,17021,17023],{"className":17022},[77,78,79,80],[21,17024,11577],{"className":17025,"style":11576},[42,46,80],[21,17027,89],{"className":17028},[88],[21,17030,17032],{"className":17031},[60],[21,17033,17035],{"className":17034,"style":96},[64],[21,17036],{},[21,17038,89],{"className":17039},[88],[21,17041,17043],{"className":17042},[60],[21,17044,17046],{"className":17045,"style":16415},[64],[21,17047],{},[21,17049,89],{"className":17050},[88],[21,17052,17054],{"className":17053},[60],[21,17055,17058],{"className":17056,"style":17057},[64],"height:1.5424em;",[21,17059],{},[21,17061],{"className":17062,"style":686},[401],[21,17064],{"className":17065,"style":686},[401],[21,17067,691],{"className":17068},[690],[21,17070],{"className":17071,"style":686},[401],[21,17073],{"className":17074,"style":686},[401],[21,17076,17078,17081,17205,17208],{"className":17077},[33],[21,17079],{"className":17080,"style":16656},[37],[21,17082,17084],{"className":17083},[2521,2522],[21,17085,17087,17197],{"className":17086},[55,56],[21,17088,17090,17194],{"className":17089},[60],[21,17091,17093,17113],{"className":17092,"style":16669},[64],[21,17094,17095,17098],{"style":16672},[21,17096],{"className":17097,"style":1678},[72],[21,17099,17101],{"className":17100},[77,78,79,80],[21,17102,17104,17107,17110],{"className":17103},[42,80],[21,17105,134],{"className":17106},[42,46,80],[21,17108,2550],{"className":17109},[1218,80],[21,17111,804],{"className":17112},[42,46,80],[21,17114,17115,17118],{"style":2593},[21,17116],{"className":17117,"style":1678},[72],[21,17119,17121],{"className":17120},[2521,2522],[21,17122,17124,17186],{"className":17123},[55,56],[21,17125,17127,17183],{"className":17126},[60],[21,17128,17130,17160],{"className":17129,"style":16669},[64],[21,17131,17133,17136],{"className":17132,"style":16711},[2612],[21,17134],{"className":17135,"style":1678},[72],[21,17137,17139,17146,17153],{"className":17138,"style":2621},[2620],[21,17140,17142],{"className":17141,"style":2626},[2625],[2628,17143,17144],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2634},[2636,17145],{"d":2638},[21,17147,17149],{"className":17148,"style":2626},[2642],[2628,17150,17151],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2645},[2636,17152],{"d":2648},[21,17154,17156],{"className":17155,"style":2626},[2652],[2628,17157,17158],{"xmlns":2630,"width":2631,"height":2632,"viewBox":2633,"preserveAspectRatio":2655},[2636,17159],{"d":2658},[21,17161,17162,17165],{"style":2593},[21,17163],{"className":17164,"style":1678},[72],[21,17166,17168,17174,17177,17180],{"className":17167},[42],[21,17169,17171],{"className":17170},[353],[21,17172,14631],{"className":17173},[42,3265],[21,17175,1369],{"className":17176},[221],[21,17178,12294],{"className":17179,"style":12293},[42,46],[21,17181,1388],{"className":17182},[273],[21,17184,89],{"className":17185},[88],[21,17187,17189],{"className":17188},[60],[21,17190,17192],{"className":17191,"style":16872},[64],[21,17193],{},[21,17195,89],{"className":17196},[88],[21,17198,17200],{"className":17199},[60],[21,17201,17203],{"className":17202,"style":16884},[64],[21,17204],{},[21,17206],{"className":17207,"style":402},[401],[21,17209,2126],{"className":17210},[42],[11,17212,17213,17214,17277,17278,17293,17294,17346],{},"Because ",[21,17215,17217],{"className":17216},[24],[21,17218,17220,17238,17268],{"className":17219,"ariaHidden":29},[28],[21,17221,17223,17226,17229,17232,17235],{"className":17222},[33],[21,17224],{"className":17225,"style":867},[37],[21,17227,47],{"className":17228},[42,46],[21,17230],{"className":17231,"style":1087},[401],[21,17233,1219],{"className":17234},[1218],[21,17236],{"className":17237,"style":1087},[401],[21,17239,17241,17244,17247,17250,17253,17256,17259,17262,17265],{"className":17240},[33],[21,17242],{"className":17243,"style":1365},[37],[21,17245,1369],{"className":17246},[221],[21,17248,804],{"className":17249},[42,46],[21,17251,8240],{"className":17252},[42],[21,17254,47],{"className":17255},[42,46],[21,17257,1388],{"className":17258},[273],[21,17260],{"className":17261,"style":686},[401],[21,17263,691],{"className":17264},[690],[21,17266],{"className":17267,"style":686},[401],[21,17269,17271,17274],{"className":17270},[33],[21,17272],{"className":17273,"style":867},[37],[21,17275,804],{"className":17276},[42,46],", the total projection and attention cost is the same\nas one full-width head; heads split the width rather than add to it. The\nexpressive gain is that the model holds ",[21,17279,17281],{"className":17280},[24],[21,17282,17284],{"className":17283,"ariaHidden":29},[28],[21,17285,17287,17290],{"className":17286},[33],[21,17288],{"className":17289,"style":867},[37],[21,17291,47],{"className":17292},[42,46]," distinct routing tables at once\ninstead of averaging them into one — one head can lock onto the previous token\nwhile another tracks long-range coreference, and ",[21,17295,17297],{"className":17296},[24],[21,17298,17300],{"className":17299,"ariaHidden":29},[28],[21,17301,17303,17306],{"className":17302},[33],[21,17304],{"className":17305,"style":886},[37],[21,17307,17309,17312],{"className":17308},[42],[21,17310,713],{"className":17311,"style":712},[42,46],[21,17313,17315],{"className":17314},[51],[21,17316,17318,17338],{"className":17317},[55,56],[21,17319,17321,17335],{"className":17320},[60],[21,17322,17324],{"className":17323,"style":726},[64],[21,17325,17326,17329],{"style":729},[21,17327],{"className":17328,"style":73},[72],[21,17330,17332],{"className":17331},[77,78,79,80],[21,17333,11577],{"className":17334,"style":11576},[42,46,80],[21,17336,89],{"className":17337},[88],[21,17339,17341],{"className":17340},[60],[21,17342,17344],{"className":17343,"style":96},[64],[21,17345],{}," learns how to combine\ntheir outputs.",[2128,17348],{"hash":17349},"084444461fe4f278f3000a37434b8480999198ffaf9bd2347d72be5e67ede4b1",[616,17351,17353],{"id":17352},"positional-encoding","Positional encoding",[11,17355,17356,17357,17372,17373,17389,17390,17393,17394,17422,17423],{},"Because self-attention is permutation-equivariant, the model cannot tell position\n",[21,17358,17360],{"className":17359},[24],[21,17361,17363],{"className":17362,"ariaHidden":29},[28],[21,17364,17366,17369],{"className":17365},[33],[21,17367],{"className":17368,"style":112},[37],[21,17370,116],{"className":17371},[42]," from position ",[21,17374,17376],{"className":17375},[24],[21,17377,17379],{"className":17378,"ariaHidden":29},[28],[21,17380,17382,17385],{"className":17381},[33],[21,17383],{"className":17384,"style":112},[37],[21,17386,17388],{"className":17387},[42],"5"," unless we tell it. The fix is to ",[163,17391,17392],{},"add"," a position-dependent\nvector ",[21,17395,17397],{"className":17396},[24],[21,17398,17400],{"className":17399,"ariaHidden":29},[28],[21,17401,17403,17406,17413,17416,17419],{"className":17402},[33],[21,17404],{"className":17405,"style":1365},[37],[21,17407,17409],{"className":17408},[353],[21,17410,17412],{"className":17411},[42,3265],"PE",[21,17414,1369],{"className":17415},[221],[21,17417,11],{"className":17418},[42,46],[21,17420,1388],{"className":17421},[273]," to each input embedding before the first layer, so\nthat order rides along in the representation itself.",[168,17424,17425],{},[15,17426,17430],{"href":17427,"ariaDescribedBy":17428,"dataFootnoteRef":6,"id":17429},"#user-content-fn-chollet-pos",[174],"user-content-fnref-chollet-pos","4",[177,17432,17433,17486,17897],{"type":179},[11,17434,17435,17438,17439,17454,17455,17470,17471,1378],{},[158,17436,17437],{},"Definition (Sinusoidal positional encoding)."," For position ",[21,17440,17442],{"className":17441},[24],[21,17443,17445],{"className":17444,"ariaHidden":29},[28],[21,17446,17448,17451],{"className":17447},[33],[21,17449],{"className":17450,"style":198},[37],[21,17452,11],{"className":17453},[42,46]," and embedding\ndimension index ",[21,17456,17458],{"className":17457},[24],[21,17459,17461],{"className":17460,"ariaHidden":29},[28],[21,17462,17464,17467],{"className":17463},[33],[21,17465],{"className":17466,"style":3932},[37],[21,17468,3798],{"className":17469},[42,46],", with model width ",[21,17472,17474],{"className":17473},[24],[21,17475,17477],{"className":17476,"ariaHidden":29},[28],[21,17478,17480,17483],{"className":17479},[33],[21,17481],{"className":17482,"style":867},[37],[21,17484,804],{"className":17485},[42,46],[21,17487,17489],{"className":17488},[1582],[21,17490,17492],{"className":17491},[24],[21,17493,17495,17537,17730,17751],{"className":17494,"ariaHidden":29},[28],[21,17496,17498,17501,17507,17510,17513,17516,17519,17522,17525,17528,17531,17534],{"className":17497},[33],[21,17499],{"className":17500,"style":1365},[37],[21,17502,17504],{"className":17503},[353],[21,17505,17412],{"className":17506},[42,3265],[21,17508,1369],{"className":17509},[221],[21,17511,11],{"className":17512},[42,46],[21,17514,1378],{"className":17515},[1377],[21,17517],{"className":17518,"style":402},[401],[21,17520,5228],{"className":17521},[42],[21,17523,3798],{"className":17524},[42,46],[21,17526,1388],{"className":17527},[273],[21,17529],{"className":17530,"style":686},[401],[21,17532,691],{"className":17533},[690],[21,17535],{"className":17536,"style":686},[401],[21,17538,17540,17544,17551,17554,17557,17685,17688,17691,17694,17697,17703,17706,17709,17712,17715,17718,17721,17724,17727],{"className":17539},[33],[21,17541],{"className":17542,"style":17543},[37],"height:1.854em;vertical-align:-0.704em;",[21,17545,17547],{"className":17546},[353],[21,17548,17550],{"className":17549},[42,3265],"sin",[21,17552],{"className":17553,"style":3270},[401],[21,17555],{"className":17556,"style":402},[401],[21,17558,17560,17566,17679],{"className":17559},[2521],[21,17561,17563],{"className":17562,"style":3281},[221,3280],[21,17564,1369],{"className":17565},[3285,3286],[21,17567,17569,17572,17676],{"className":17568},[42],[21,17570],{"className":17571},[221,1657],[21,17573,17575],{"className":17574},[1661],[21,17576,17578,17667],{"className":17577},[55,56],[21,17579,17581,17664],{"className":17580},[60],[21,17582,17585,17645,17653],{"className":17583,"style":17584},[64],"height:1.1076em;",[21,17586,17588,17591],{"style":17587},"top:-2.296em;",[21,17589],{"className":17590,"style":1678},[72],[21,17592,17594,17598],{"className":17593},[42],[21,17595,17597],{"className":17596},[42],"1000",[21,17599,17601,17604],{"className":17600},[42],[21,17602,1373],{"className":17603},[42],[21,17605,17607],{"className":17606},[51],[21,17608,17610],{"className":17609},[55],[21,17611,17613],{"className":17612},[60],[21,17614,17617],{"className":17615,"style":17616},[64],"height:0.814em;",[21,17618,17620,17623],{"style":17619},"top:-2.989em;margin-right:0.05em;",[21,17621],{"className":17622,"style":73},[72],[21,17624,17626],{"className":17625},[77,78,79,80],[21,17627,17629,17633,17636,17639,17642],{"className":17628},[42,80],[21,17630],{"className":17631,"style":17632},[401,80],"margin-right:0.1952em;",[21,17634,5228],{"className":17635},[42,80],[21,17637,3798],{"className":17638},[42,46,80],[21,17640,8240],{"className":17641},[42,80],[21,17643,804],{"className":17644},[42,46,80],[21,17646,17647,17650],{"style":1858},[21,17648],{"className":17649,"style":1678},[72],[21,17651],{"className":17652,"style":1866},[1865],[21,17654,17655,17658],{"style":1869},[21,17656],{"className":17657,"style":1678},[72],[21,17659,17661],{"className":17660},[42],[21,17662,11],{"className":17663},[42,46],[21,17665,89],{"className":17666},[88],[21,17668,17670],{"className":17669},[60],[21,17671,17674],{"className":17672,"style":17673},[64],"height:0.704em;",[21,17675],{},[21,17677],{"className":17678},[273,1657],[21,17680,17682],{"className":17681,"style":3281},[273,3280],[21,17683,1388],{"className":17684},[3285,3286],[21,17686],{"className":17687,"style":402},[401],[21,17689,1378],{"className":17690},[1377],[21,17692],{"className":17693,"style":1958},[401],[21,17695],{"className":17696,"style":402},[401],[21,17698,17700],{"className":17699},[353],[21,17701,17412],{"className":17702},[42,3265],[21,17704,1369],{"className":17705},[221],[21,17707,11],{"className":17708},[42,46],[21,17710,1378],{"className":17711},[1377],[21,17713],{"className":17714,"style":402},[401],[21,17716,5228],{"className":17717},[42],[21,17719,3798],{"className":17720},[42,46],[21,17722],{"className":17723,"style":1087},[401],[21,17725,11681],{"className":17726},[1218],[21,17728],{"className":17729,"style":1087},[401],[21,17731,17733,17736,17739,17742,17745,17748],{"className":17732},[33],[21,17734],{"className":17735,"style":1365},[37],[21,17737,116],{"className":17738},[42],[21,17740,1388],{"className":17741},[273],[21,17743],{"className":17744,"style":686},[401],[21,17746,691],{"className":17747},[690],[21,17749],{"className":17750,"style":686},[401],[21,17752,17754,17757,17764,17767,17770,17891,17894],{"className":17753},[33],[21,17755],{"className":17756,"style":17543},[37],[21,17758,17760],{"className":17759},[353],[21,17761,17763],{"className":17762},[42,3265],"cos",[21,17765],{"className":17766,"style":3270},[401],[21,17768],{"className":17769,"style":402},[401],[21,17771,17773,17779,17885],{"className":17772},[2521],[21,17774,17776],{"className":17775,"style":3281},[221,3280],[21,17777,1369],{"className":17778},[3285,3286],[21,17780,17782,17785,17882],{"className":17781},[42],[21,17783],{"className":17784},[221,1657],[21,17786,17788],{"className":17787},[1661],[21,17789,17791,17874],{"className":17790},[55,56],[21,17792,17794,17871],{"className":17793},[60],[21,17795,17797,17852,17860],{"className":17796,"style":17584},[64],[21,17798,17799,17802],{"style":17587},[21,17800],{"className":17801,"style":1678},[72],[21,17803,17805,17808],{"className":17804},[42],[21,17806,17597],{"className":17807},[42],[21,17809,17811,17814],{"className":17810},[42],[21,17812,1373],{"className":17813},[42],[21,17815,17817],{"className":17816},[51],[21,17818,17820],{"className":17819},[55],[21,17821,17823],{"className":17822},[60],[21,17824,17826],{"className":17825,"style":17616},[64],[21,17827,17828,17831],{"style":17619},[21,17829],{"className":17830,"style":73},[72],[21,17832,17834],{"className":17833},[77,78,79,80],[21,17835,17837,17840,17843,17846,17849],{"className":17836},[42,80],[21,17838],{"className":17839,"style":17632},[401,80],[21,17841,5228],{"className":17842},[42,80],[21,17844,3798],{"className":17845},[42,46,80],[21,17847,8240],{"className":17848},[42,80],[21,17850,804],{"className":17851},[42,46,80],[21,17853,17854,17857],{"style":1858},[21,17855],{"className":17856,"style":1678},[72],[21,17858],{"className":17859,"style":1866},[1865],[21,17861,17862,17865],{"style":1869},[21,17863],{"className":17864,"style":1678},[72],[21,17866,17868],{"className":17867},[42],[21,17869,11],{"className":17870},[42,46],[21,17872,89],{"className":17873},[88],[21,17875,17877],{"className":17876},[60],[21,17878,17880],{"className":17879,"style":17673},[64],[21,17881],{},[21,17883],{"className":17884},[273,1657],[21,17886,17888],{"className":17887,"style":3281},[273,3280],[21,17889,1388],{"className":17890},[3285,3286],[21,17892],{"className":17893,"style":402},[401],[21,17895,2126],{"className":17896},[42],[11,17898,17899,17900,17919,17920,17957],{},"Each dimension is a sinusoid; the wavelength grows geometrically from ",[21,17901,17903],{"className":17902},[24],[21,17904,17906],{"className":17905,"ariaHidden":29},[28],[21,17907,17909,17912,17915],{"className":17908},[33],[21,17910],{"className":17911,"style":112},[37],[21,17913,5228],{"className":17914},[42],[21,17916,17918],{"className":17917,"style":202},[42,46],"π"," to\n",[21,17921,17923],{"className":17922},[24],[21,17924,17926,17945],{"className":17925,"ariaHidden":29},[28],[21,17927,17929,17932,17936,17939,17942],{"className":17928},[33],[21,17930],{"className":17931,"style":112},[37],[21,17933,17935],{"className":17934},[42],"10000",[21,17937],{"className":17938,"style":1087},[401],[21,17940,1219],{"className":17941},[1218],[21,17943],{"className":17944,"style":1087},[401],[21,17946,17948,17951,17954],{"className":17947},[33],[21,17949],{"className":17950,"style":112},[37],[21,17952,5228],{"className":17953},[42],[21,17955,17918],{"className":17956,"style":202},[42,46]," across dimensions, so the vector encodes position at many\nresolutions at once.",[11,17959,17960,17961,17964,17965,18011,18012,18039],{},"The geometric spread of frequencies is the key property: low-index dimensions oscillate\nquickly to distinguish neighbors, high-index dimensions oscillate slowly to\ndistinguish distant regions. The encoding also admits ",[163,17962,17963],{},"relative"," shifts, since\n",[21,17966,17968],{"className":17967},[24],[21,17969,17971,17998],{"className":17970,"ariaHidden":29},[28],[21,17972,17974,17977,17983,17986,17989,17992,17995],{"className":17973},[33],[21,17975],{"className":17976,"style":1365},[37],[21,17978,17980],{"className":17979},[353],[21,17981,17412],{"className":17982},[42,3265],[21,17984,1369],{"className":17985},[221],[21,17987,11],{"className":17988},[42,46],[21,17990],{"className":17991,"style":1087},[401],[21,17993,11681],{"className":17994},[1218],[21,17996],{"className":17997,"style":1087},[401],[21,17999,18001,18004,18008],{"className":18000},[33],[21,18002],{"className":18003,"style":1365},[37],[21,18005,18007],{"className":18006},[42],"Δ",[21,18009,1388],{"className":18010},[273]," is a fixed linear function of ",[21,18013,18015],{"className":18014},[24],[21,18016,18018],{"className":18017,"ariaHidden":29},[28],[21,18019,18021,18024,18030,18033,18036],{"className":18020},[33],[21,18022],{"className":18023,"style":1365},[37],[21,18025,18027],{"className":18026},[353],[21,18028,17412],{"className":18029},[42,3265],[21,18031,1369],{"className":18032},[221],[21,18034,11],{"className":18035},[42,46],[21,18037,1388],{"className":18038},[273],",\nso the model can learn to attend by offset, not just absolute index.",[2128,18041],{"hash":18042},"685cf0eaa8fde0a43cac0b1af564addde3061d23fa6822815d8442148858c098",[11,18044,18045],{},"Learned positional embeddings — a trainable lookup table indexed by position —\nare the common alternative; they fit the data more tightly but, unlike the\nsinusoids, do not extrapolate to sequences longer than those seen in training.",[636,18047,18048,18064],{},[639,18049,18050],{},[642,18051,18052,18055,18058,18061],{},[645,18053,18054],{},"Scheme",[645,18056,18057],{},"Form",[645,18059,18060],{},"Generalizes past training length?",[645,18062,18063],{},"Cost",[658,18065,18066,18129,18456],{},[642,18067,18068,18071,18108,18126],{},[663,18069,18070],{},"Sinusoidal",[663,18072,18073,18074,18107],{},"fixed ",[21,18075,18077],{"className":18076},[24],[21,18078,18080],{"className":18079,"ariaHidden":29},[28],[21,18081,18083,18086,18092,18095,18098,18101],{"className":18082},[33],[21,18084],{"className":18085,"style":1365},[37],[21,18087,18089],{"className":18088},[353],[21,18090,17550],{"className":18091},[42,3265],[21,18093],{"className":18094,"style":402},[401],[21,18096,8240],{"className":18097},[42],[21,18099],{"className":18100,"style":402},[401],[21,18102,18104],{"className":18103},[353],[21,18105,17763],{"className":18106},[42,3265]," of position",[663,18109,18110,18111,1388],{},"yes (deterministic for any ",[21,18112,18114],{"className":18113},[24],[21,18115,18117],{"className":18116,"ariaHidden":29},[28],[21,18118,18120,18123],{"className":18119},[33],[21,18121],{"className":18122,"style":198},[37],[21,18124,11],{"className":18125},[42,46],[663,18127,18128],{},"none",[642,18130,18131,18134,18294,18376],{},[663,18132,18133],{},"Learned",[663,18135,18136,18137],{},"trainable table ",[21,18138,18140],{"className":18139},[24],[21,18141,18143,18206],{"className":18142,"ariaHidden":29},[28],[21,18144,18146,18149,18197,18200,18203],{"className":18145},[33],[21,18147],{"className":18148,"style":701},[37],[21,18150,18152,18155],{"className":18151},[42],[21,18153,5782],{"className":18154,"style":9194},[42,46],[21,18156,18158],{"className":18157},[51],[21,18159,18161,18189],{"className":18160},[55,56],[21,18162,18164,18186],{"className":18163},[60],[21,18165,18167],{"className":18166,"style":2478},[64],[21,18168,18170,18173],{"style":18169},"top:-2.55em;margin-left:-0.0576em;margin-right:0.05em;",[21,18171],{"className":18172,"style":73},[72],[21,18174,18176],{"className":18175},[77,78,79,80],[21,18177,18179],{"className":18178},[42,80],[21,18180,18182],{"className":18181},[42,1965,80],[21,18183,18185],{"className":18184},[42,80],"pos",[21,18187,89],{"className":18188},[88],[21,18190,18192],{"className":18191},[60],[21,18193,18195],{"className":18194,"style":267},[64],[21,18196],{},[21,18198],{"className":18199,"style":686},[401],[21,18201,3730],{"className":18202},[690],[21,18204],{"className":18205,"style":686},[401],[21,18207,18209,18212],{"className":18208},[33],[21,18210],{"className":18211,"style":765},[37],[21,18213,18215,18218],{"className":18214},[42],[21,18216,773],{"className":18217},[42,772],[21,18219,18221],{"className":18220},[51],[21,18222,18224],{"className":18223},[55],[21,18225,18227],{"className":18226},[60],[21,18228,18230],{"className":18229,"style":765},[64],[21,18231,18232,18235],{"style":788},[21,18233],{"className":18234,"style":73},[72],[21,18236,18238],{"className":18237},[77,78,79,80],[21,18239,18241,18288,18291],{"className":18240},[42,80],[21,18242,18244,18247],{"className":18243},[42,80],[21,18245,134],{"className":18246},[42,46,80],[21,18248,18250],{"className":18249},[51],[21,18251,18253,18280],{"className":18252},[55,56],[21,18254,18256,18277],{"className":18255},[60],[21,18257,18259],{"className":18258,"style":1161},[64],[21,18260,18261,18264],{"style":1164},[21,18262],{"className":18263,"style":824},[72],[21,18265,18267],{"className":18266},[77,828,829,80],[21,18268,18270],{"className":18269},[42,80],[21,18271,18273],{"className":18272},[353,80],[21,18274,18276],{"className":18275},[42,3265,80],"max",[21,18278,89],{"className":18279},[88],[21,18281,18283],{"className":18282},[60],[21,18284,18286],{"className":18285,"style":1183},[64],[21,18287],{},[21,18289,2550],{"className":18290},[1218,80],[21,18292,804],{"className":18293},[42,46,80],[663,18295,18296,18297,1388],{},"no (undefined for ",[21,18298,18300],{"className":18299},[24],[21,18301,18303,18323],{"className":18302,"ariaHidden":29},[28],[21,18304,18306,18310,18313,18316,18320],{"className":18305},[33],[21,18307],{"className":18308,"style":18309},[37],"height:0.7335em;vertical-align:-0.1944em;",[21,18311,11],{"className":18312},[42,46],[21,18314],{"className":18315,"style":686},[401],[21,18317,18319],{"className":18318},[690],">",[21,18321],{"className":18322,"style":686},[401],[21,18324,18326,18330],{"className":18325},[33],[21,18327],{"className":18328,"style":18329},[37],"height:0.5806em;vertical-align:-0.15em;",[21,18331,18333,18336],{"className":18332},[42],[21,18334,134],{"className":18335},[42,46],[21,18337,18339],{"className":18338},[51],[21,18340,18342,18368],{"className":18341},[55,56],[21,18343,18345,18365],{"className":18344},[60],[21,18346,18348],{"className":18347,"style":2478},[64],[21,18349,18350,18353],{"style":68},[21,18351],{"className":18352,"style":73},[72],[21,18354,18356],{"className":18355},[77,78,79,80],[21,18357,18359],{"className":18358},[42,80],[21,18360,18362],{"className":18361},[353,80],[21,18363,18276],{"className":18364},[42,3265,80],[21,18366,89],{"className":18367},[88],[21,18369,18371],{"className":18370},[60],[21,18372,18374],{"className":18373,"style":96},[64],[21,18375],{},[663,18377,18378,18455],{},[21,18379,18381],{"className":18380},[24],[21,18382,18384,18446],{"className":18383,"ariaHidden":29},[28],[21,18385,18387,18391,18437,18440,18443],{"className":18386},[33],[21,18388],{"className":18389,"style":18390},[37],"height:0.5945em;vertical-align:-0.15em;",[21,18392,18394,18397],{"className":18393},[42],[21,18395,134],{"className":18396},[42,46],[21,18398,18400],{"className":18399},[51],[21,18401,18403,18429],{"className":18402},[55,56],[21,18404,18406,18426],{"className":18405},[60],[21,18407,18409],{"className":18408,"style":2478},[64],[21,18410,18411,18414],{"style":68},[21,18412],{"className":18413,"style":73},[72],[21,18415,18417],{"className":18416},[77,78,79,80],[21,18418,18420],{"className":18419},[42,80],[21,18421,18423],{"className":18422},[353,80],[21,18424,18276],{"className":18425},[42,3265,80],[21,18427,89],{"className":18428},[88],[21,18430,18432],{"className":18431},[60],[21,18433,18435],{"className":18434,"style":96},[64],[21,18436],{},[21,18438],{"className":18439,"style":1087},[401],[21,18441,1219],{"className":18442},[1218],[21,18444],{"className":18445,"style":1087},[401],[21,18447,18449,18452],{"className":18448},[33],[21,18450],{"className":18451,"style":867},[37],[21,18453,804],{"className":18454},[42,46]," params",[642,18457,18458,18461,18498,18501],{},[663,18459,18460],{},"Relative",[663,18462,18463,18464],{},"bias on scores by offset ",[21,18465,18467],{"className":18466},[24],[21,18468,18470,18489],{"className":18469,"ariaHidden":29},[28],[21,18471,18473,18477,18480,18483,18486],{"className":18472},[33],[21,18474],{"className":18475,"style":18476},[37],"height:0.7429em;vertical-align:-0.0833em;",[21,18478,3798],{"className":18479},[42,46],[21,18481],{"className":18482,"style":1087},[401],[21,18484,7294],{"className":18485},[1218],[21,18487],{"className":18488,"style":1087},[401],[21,18490,18492,18495],{"className":18491},[33],[21,18493],{"className":18494,"style":1287},[37],[21,18496,257],{"className":18497,"style":256},[42,46],[663,18499,18500],{},"yes",[663,18502,18503],{},"per-head bias",[616,18505,18507],{"id":18506},"the-transformer-block","The Transformer block",[11,18509,18510,18511,18514,18515,18518,18519,18522,18523,2126,18527],{},"Self-attention mixes information ",[163,18512,18513],{},"across"," positions; it does no nonlinear\nprocessing ",[163,18516,18517],{},"within"," a position. The ",[158,18520,18521],{},"Transformer block"," pairs the two: a\nmulti-head self-attention sublayer followed by a position-wise feed-forward\nnetwork, each sublayer wrapped in a residual connection and ",[15,18524,18526],{"href":18525},"\u002Fdeep-learning\u002Fregularization\u002Fnormalization","layer\nnormalization",[168,18528,18529],{},[15,18530,17388],{"href":18531,"ariaDescribedBy":18532,"dataFootnoteRef":6,"id":18533},"#user-content-fn-chollet-block",[174],"user-content-fnref-chollet-block",[177,18535,18536,18542,18801],{"type":179},[11,18537,18538,18541],{},[158,18539,18540],{},"Definition (Position-wise feed-forward network)."," A two-layer MLP applied\nindependently and identically to every position:",[21,18543,18545],{"className":18544},[1582],[21,18546,18548],{"className":18547},[24],[21,18549,18551,18582,18650,18752],{"className":18550,"ariaHidden":29},[28],[21,18552,18554,18557,18564,18567,18570,18573,18576,18579],{"className":18553},[33],[21,18555],{"className":18556,"style":1365},[37],[21,18558,18560],{"className":18559},[353],[21,18561,18563],{"className":18562},[42,3265],"FFN",[21,18565,1369],{"className":18566},[221],[21,18568,705],{"className":18569},[42,46],[21,18571,1388],{"className":18572},[273],[21,18574],{"className":18575,"style":686},[401],[21,18577,691],{"className":18578},[690],[21,18580],{"className":18581,"style":686},[401],[21,18583,18585,18588,18595,18598,18601,18641,18644,18647],{"className":18584},[33],[21,18586],{"className":18587,"style":1365},[37],[21,18589,18591],{"className":18590},[353],[21,18592,18594],{"className":18593},[42,3265],"ReLU",[21,18596,1369],{"className":18597},[221],[21,18599,705],{"className":18600},[42,46],[21,18602,18604,18607],{"className":18603},[42],[21,18605,713],{"className":18606,"style":712},[42,46],[21,18608,18610],{"className":18609},[51],[21,18611,18613,18633],{"className":18612},[55,56],[21,18614,18616,18630],{"className":18615},[60],[21,18617,18619],{"className":18618,"style":5491},[64],[21,18620,18621,18624],{"style":729},[21,18622],{"className":18623,"style":73},[72],[21,18625,18627],{"className":18626},[77,78,79,80],[21,18628,116],{"className":18629},[42,80],[21,18631,89],{"className":18632},[88],[21,18634,18636],{"className":18635},[60],[21,18637,18639],{"className":18638,"style":96},[64],[21,18640],{},[21,18642],{"className":18643,"style":1087},[401],[21,18645,11681],{"className":18646},[1218],[21,18648],{"className":18649,"style":1087},[401],[21,18651,18653,18656,18697,18700,18703,18743,18746,18749],{"className":18652},[33],[21,18654],{"className":18655,"style":1365},[37],[21,18657,18659,18663],{"className":18658},[42],[21,18660,18662],{"className":18661},[42,46],"b",[21,18664,18666],{"className":18665},[51],[21,18667,18669,18689],{"className":18668},[55,56],[21,18670,18672,18686],{"className":18671},[60],[21,18673,18675],{"className":18674,"style":5491},[64],[21,18676,18677,18680],{"style":68},[21,18678],{"className":18679,"style":73},[72],[21,18681,18683],{"className":18682},[77,78,79,80],[21,18684,116],{"className":18685},[42,80],[21,18687,89],{"className":18688},[88],[21,18690,18692],{"className":18691},[60],[21,18693,18695],{"className":18694,"style":96},[64],[21,18696],{},[21,18698,1388],{"className":18699},[273],[21,18701],{"className":18702,"style":402},[401],[21,18704,18706,18709],{"className":18705},[42],[21,18707,713],{"className":18708,"style":712},[42,46],[21,18710,18712],{"className":18711},[51],[21,18713,18715,18735],{"className":18714},[55,56],[21,18716,18718,18732],{"className":18717},[60],[21,18719,18721],{"className":18720,"style":5491},[64],[21,18722,18723,18726],{"style":729},[21,18724],{"className":18725,"style":73},[72],[21,18727,18729],{"className":18728},[77,78,79,80],[21,18730,5228],{"className":18731},[42,80],[21,18733,89],{"className":18734},[88],[21,18736,18738],{"className":18737},[60],[21,18739,18741],{"className":18740,"style":96},[64],[21,18742],{},[21,18744],{"className":18745,"style":1087},[401],[21,18747,11681],{"className":18748},[1218],[21,18750],{"className":18751,"style":1087},[401],[21,18753,18755,18758,18798],{"className":18754},[33],[21,18756],{"className":18757,"style":2331},[37],[21,18759,18761,18764],{"className":18760},[42],[21,18762,18662],{"className":18763},[42,46],[21,18765,18767],{"className":18766},[51],[21,18768,18770,18790],{"className":18769},[55,56],[21,18771,18773,18787],{"className":18772},[60],[21,18774,18776],{"className":18775,"style":5491},[64],[21,18777,18778,18781],{"style":68},[21,18779],{"className":18780,"style":73},[72],[21,18782,18784],{"className":18783},[77,78,79,80],[21,18785,5228],{"className":18786},[42,80],[21,18788,89],{"className":18789},[88],[21,18791,18793],{"className":18792},[60],[21,18794,18796],{"className":18795,"style":96},[64],[21,18797],{},[21,18799,1378],{"className":18800},[1377],[11,18802,18803,18804,18822,18823,18857],{},"with an inner width typically ",[21,18805,18807],{"className":18806},[24],[21,18808,18810],{"className":18809,"ariaHidden":29},[28],[21,18811,18813,18816,18819],{"className":18812},[33],[21,18814],{"className":18815,"style":867},[37],[21,18817,17430],{"className":18818},[42],[21,18820,804],{"className":18821},[42,46],". It is the per-token nonlinearity that\nattention lacks; sharing weights across positions makes it a ",[21,18824,18826],{"className":18825},[24],[21,18827,18829,18848],{"className":18828,"ariaHidden":29},[28],[21,18830,18832,18836,18839,18842,18845],{"className":18831},[33],[21,18833],{"className":18834,"style":18835},[37],"height:0.7278em;vertical-align:-0.0833em;",[21,18837,116],{"className":18838},[42],[21,18840],{"className":18841,"style":1087},[401],[21,18843,2550],{"className":18844},[1218],[21,18846],{"className":18847,"style":1087},[401],[21,18849,18851,18854],{"className":18850},[33],[21,18852],{"className":18853,"style":112},[37],[21,18855,116],{"className":18856},[42],"\nconvolution in disguise.",[11,18859,18860,18861,18940,18941,18945],{},"Each sublayer is wrapped identically, ",[21,18862,18864],{"className":18863},[24],[21,18865,18867,18889,18917],{"className":18866,"ariaHidden":29},[28],[21,18868,18870,18874,18880,18883,18886],{"className":18869},[33],[21,18871],{"className":18872,"style":18873},[37],"height:0.6151em;",[21,18875,18877],{"className":18876},[353],[21,18878,1969],{"className":18879},[42,3265],[21,18881],{"className":18882,"style":686},[401],[21,18884,691],{"className":18885},[690],[21,18887],{"className":18888,"style":686},[401],[21,18890,18892,18895,18902,18905,18908,18911,18914],{"className":18891},[33],[21,18893],{"className":18894,"style":1365},[37],[21,18896,18898],{"className":18897},[353],[21,18899,18901],{"className":18900},[42,3265],"LayerNorm",[21,18903,1369],{"className":18904},[221],[21,18906,705],{"className":18907},[42,46],[21,18909],{"className":18910,"style":1087},[401],[21,18912,11681],{"className":18913},[1218],[21,18915],{"className":18916,"style":1087},[401],[21,18918,18920,18923,18930,18933,18936],{"className":18919},[33],[21,18921],{"className":18922,"style":1365},[37],[21,18924,18926],{"className":18925},[353],[21,18927,18929],{"className":18928},[42,3265],"Sublayer",[21,18931,1369],{"className":18932},[221],[21,18934,705],{"className":18935},[42,46],[21,18937,18939],{"className":18938},[273],"))",". The residual skip lets gradients reach early\nlayers undiminished — the same mechanism that powers ",[15,18942,18944],{"href":18943},"\u002Fdeep-learning\u002Farchitectures\u002Fcnn-architectures","residual\nnetworks"," — and LayerNorm holds\neach position's activations at a stable scale across depth.",[2128,18947],{"hash":18948},"1081d93c9d7cfa7915db69065a0b79c97f6afe1c41346cf0827b9cef51f21481",[5230,18950,18952],{"id":18951},"encoder-and-decoder","Encoder and decoder",[11,18954,18955,18956,18959],{},"The original Transformer is an ",[158,18957,18958],{},"encoder-decoder",". The encoder stacks identical\nblocks with unrestricted self-attention — every position sees every other. The\ndecoder adds two twists that make it a valid autoregressive generator.",[636,18961,18962,18977],{},[639,18963,18964],{},[642,18965,18966,18969,18971,18974],{},[645,18967,18968],{},"Component",[645,18970,12270],{},[645,18972,18973],{},"Cross-attention",[645,18975,18976],{},"Sees the future?",[658,18978,18979,18993],{},[642,18980,18981,18984,18987,18990],{},[663,18982,18983],{},"Encoder block",[663,18985,18986],{},"full (all-pairs)",[663,18988,18989],{},"—",[663,18991,18992],{},"yes (whole input visible)",[642,18994,18995,18998,19003,19006],{},[663,18996,18997],{},"Decoder block",[663,18999,19000],{},[163,19001,19002],{},"masked",[663,19004,19005],{},"attends to encoder output",[663,19007,19008],{},"no (causal mask)",[19010,19011,19012,19251],"ul",{},[19013,19014,19015,19018,19019,19034,19035,19064,19065,19142,19143,19176,19177,19250],"li",{},[158,19016,19017],{},"Masked self-attention."," To generate token ",[21,19020,19022],{"className":19021},[24],[21,19023,19025],{"className":19024,"ariaHidden":29},[28],[21,19026,19028,19031],{"className":19027},[33],[21,19029],{"className":19030,"style":18873},[37],[21,19032,84],{"className":19033},[42,46]," using only tokens ",[21,19036,19038],{"className":19037},[24],[21,19039,19041,19055],{"className":19040,"ariaHidden":29},[28],[21,19042,19044,19048,19052],{"className":19043},[33],[21,19045],{"className":19046,"style":19047},[37],"height:0.5782em;vertical-align:-0.0391em;",[21,19049,19051],{"className":19050},[690],"\u003C",[21,19053],{"className":19054,"style":686},[401],[21,19056,19058,19061],{"className":19057},[33],[21,19059],{"className":19060,"style":18873},[37],[21,19062,84],{"className":19063},[42,46],", the\ndecoder sets ",[21,19066,19068],{"className":19067},[24],[21,19069,19071,19129],{"className":19070,"ariaHidden":29},[28],[21,19072,19074,19077,19120,19123,19126],{"className":19073},[33],[21,19075],{"className":19076,"style":701},[37],[21,19078,19080,19083],{"className":19079},[42],[21,19081,9195],{"className":19082,"style":9194},[42,46],[21,19084,19086],{"className":19085},[51],[21,19087,19089,19112],{"className":19088},[55,56],[21,19090,19092,19109],{"className":19091},[60],[21,19093,19095],{"className":19094,"style":243},[64],[21,19096,19097,19100],{"style":18169},[21,19098],{"className":19099,"style":73},[72],[21,19101,19103],{"className":19102},[77,78,79,80],[21,19104,19106],{"className":19105},[42,80],[21,19107,12930],{"className":19108,"style":256},[42,46,80],[21,19110,89],{"className":19111},[88],[21,19113,19115],{"className":19114},[60],[21,19116,19118],{"className":19117,"style":267},[64],[21,19119],{},[21,19121],{"className":19122,"style":686},[401],[21,19124,691],{"className":19125},[690],[21,19127],{"className":19128,"style":686},[401],[21,19130,19132,19135,19138],{"className":19131},[33],[21,19133],{"className":19134,"style":4149},[37],[21,19136,7294],{"className":19137},[42],[21,19139,19141],{"className":19140},[42],"∞"," for ",[21,19144,19146],{"className":19145},[24],[21,19147,19149,19167],{"className":19148,"ariaHidden":29},[28],[21,19150,19152,19155,19158,19161,19164],{"className":19151},[33],[21,19153],{"className":19154,"style":1287},[37],[21,19156,257],{"className":19157,"style":256},[42,46],[21,19159],{"className":19160,"style":686},[401],[21,19162,18319],{"className":19163},[690],[21,19165],{"className":19166,"style":686},[401],[21,19168,19170,19173],{"className":19169},[33],[21,19171],{"className":19172,"style":3932},[37],[21,19174,3798],{"className":19175},[42,46]," before the softmax (the masking\nbranch of the algorithm above), forcing ",[21,19178,19180],{"className":19179},[24],[21,19181,19183,19241],{"className":19182,"ariaHidden":29},[28],[21,19184,19186,19189,19232,19235,19238],{"className":19185},[33],[21,19187],{"className":19188,"style":701},[37],[21,19190,19192,19195],{"className":19191},[42],[21,19193,3644],{"className":19194},[42,46],[21,19196,19198],{"className":19197},[51],[21,19199,19201,19224],{"className":19200},[55,56],[21,19202,19204,19221],{"className":19203},[60],[21,19205,19207],{"className":19206,"style":243},[64],[21,19208,19209,19212],{"style":68},[21,19210],{"className":19211,"style":73},[72],[21,19213,19215],{"className":19214},[77,78,79,80],[21,19216,19218],{"className":19217},[42,80],[21,19219,12930],{"className":19220,"style":256},[42,46,80],[21,19222,89],{"className":19223},[88],[21,19225,19227],{"className":19226},[60],[21,19228,19230],{"className":19229,"style":267},[64],[21,19231],{},[21,19233],{"className":19234,"style":686},[401],[21,19236,691],{"className":19237},[690],[21,19239],{"className":19240,"style":686},[401],[21,19242,19244,19247],{"className":19243},[33],[21,19245],{"className":19246,"style":112},[37],[21,19248,1373],{"className":19249},[42]," on the future. Generation\nstays causal.",[19013,19252,19253,19256],{},[158,19254,19255],{},"Cross-attention."," A second attention sublayer takes its queries from the\ndecoder but its keys and values from the encoder's output, letting each\ngenerated token read the entire source sequence — the channel through which a\ntranslation model reads the sentence it is translating.",[11,19258,19259,19260,19326,19327,19342,19343,19358,19359,19464,19465,19579,19580,19613,19614,19665,19666,19699,19700,19703],{},"Cross-attention is the setting where the ",[21,19261,19263],{"className":19262},[24],[21,19264,19266,19317],{"className":19265,"ariaHidden":29},[28],[21,19267,19269,19272,19275,19278,19314],{"className":19268},[33],[21,19270],{"className":19271,"style":2331},[37],[21,19273,134],{"className":19274},[42,46],[21,19276],{"className":19277,"style":686},[401],[21,19279,19281,19308,19311],{"className":19280},[690],[21,19282,19284],{"className":19283},[690],[21,19285,19287],{"className":19286},[42,2347],[21,19288,19290],{"className":19289},[2351],[21,19291,19293,19296,19305],{"className":19292},[2355],[21,19294],{"className":19295,"style":2331},[37],[21,19297,19299],{"className":19298},[2362],[21,19300,19302],{"className":19301},[42],[21,19303,2369],{"className":19304},[690],[21,19306],{"className":19307},[2373],[21,19309],{"className":19310},[401,2377],[21,19312,691],{"className":19313},[690],[21,19315],{"className":19316,"style":686},[401],[21,19318,19320,19323],{"className":19319},[33],[21,19321],{"className":19322,"style":130},[37],[21,19324,2186],{"className":19325},[42,46]," shapes matter. In an\nencoder of length ",[21,19328,19330],{"className":19329},[24],[21,19331,19333],{"className":19332,"ariaHidden":29},[28],[21,19334,19336,19339],{"className":19335},[33],[21,19337],{"className":19338,"style":130},[37],[21,19340,2186],{"className":19341},[42,46]," and a decoder generating length ",[21,19344,19346],{"className":19345},[24],[21,19347,19349],{"className":19348,"ariaHidden":29},[28],[21,19350,19352,19355],{"className":19351},[33],[21,19353],{"className":19354,"style":130},[37],[21,19356,134],{"className":19357},[42,46],", the decoder supplies\n",[21,19360,19362],{"className":19361},[24],[21,19363,19365,19383],{"className":19364,"ariaHidden":29},[28],[21,19366,19368,19371,19374,19377,19380],{"className":19367},[33],[21,19369],{"className":19370,"style":2166},[37],[21,19372,739],{"className":19373},[42,46],[21,19375],{"className":19376,"style":686},[401],[21,19378,3730],{"className":19379},[690],[21,19381],{"className":19382,"style":686},[401],[21,19384,19386,19389],{"className":19385},[33],[21,19387],{"className":19388,"style":765},[37],[21,19390,19392,19395],{"className":19391},[42],[21,19393,773],{"className":19394},[42,772],[21,19396,19398],{"className":19397},[51],[21,19399,19401],{"className":19400},[55],[21,19402,19404],{"className":19403},[60],[21,19405,19407],{"className":19406,"style":765},[64],[21,19408,19409,19412],{"style":788},[21,19410],{"className":19411,"style":73},[72],[21,19413,19415],{"className":19414},[77,78,79,80],[21,19416,19418,19421,19424],{"className":19417},[42,80],[21,19419,134],{"className":19420},[42,46,80],[21,19422,2550],{"className":19423},[1218,80],[21,19425,19427,19430],{"className":19426},[42,80],[21,19428,804],{"className":19429},[42,46,80],[21,19431,19433],{"className":19432},[51],[21,19434,19436,19456],{"className":19435},[55,56],[21,19437,19439,19453],{"className":19438},[60],[21,19440,19442],{"className":19441,"style":817},[64],[21,19443,19444,19447],{"style":820},[21,19445],{"className":19446,"style":824},[72],[21,19448,19450],{"className":19449},[77,828,829,80],[21,19451,230],{"className":19452,"style":229},[42,46,80],[21,19454,89],{"className":19455},[88],[21,19457,19459],{"className":19458},[60],[21,19460,19462],{"className":19461,"style":842},[64],[21,19463],{}," while the encoder supplies ",[21,19466,19468],{"className":19467},[24],[21,19469,19471,19498],{"className":19470,"ariaHidden":29},[28],[21,19472,19474,19477,19480,19483,19486,19489,19492,19495],{"className":19473},[33],[21,19475],{"className":19476,"style":2166},[37],[21,19478,920],{"className":19479,"style":919},[42,46],[21,19481,1378],{"className":19482},[1377],[21,19484],{"className":19485,"style":402},[401],[21,19487,1088],{"className":19488,"style":1087},[42,46],[21,19490],{"className":19491,"style":686},[401],[21,19493,3730],{"className":19494},[690],[21,19496],{"className":19497,"style":686},[401],[21,19499,19501,19504],{"className":19500},[33],[21,19502],{"className":19503,"style":765},[37],[21,19505,19507,19510],{"className":19506},[42],[21,19508,773],{"className":19509},[42,772],[21,19511,19513],{"className":19512},[51],[21,19514,19516],{"className":19515},[55],[21,19517,19519],{"className":19518},[60],[21,19520,19522],{"className":19521,"style":765},[64],[21,19523,19524,19527],{"style":788},[21,19525],{"className":19526,"style":73},[72],[21,19528,19530],{"className":19529},[77,78,79,80],[21,19531,19533,19536,19539],{"className":19532},[42,80],[21,19534,2186],{"className":19535},[42,46,80],[21,19537,2550],{"className":19538},[1218,80],[21,19540,19542,19545],{"className":19541},[42,80],[21,19543,804],{"className":19544},[42,46,80],[21,19546,19548],{"className":19547},[51],[21,19549,19551,19571],{"className":19550},[55,56],[21,19552,19554,19568],{"className":19553},[60],[21,19555,19557],{"className":19556,"style":817},[64],[21,19558,19559,19562],{"style":820},[21,19560],{"className":19561,"style":824},[72],[21,19563,19565],{"className":19564},[77,828,829,80],[21,19566,230],{"className":19567,"style":229},[42,46,80],[21,19569,89],{"className":19570},[88],[21,19572,19574],{"className":19573},[60],[21,19575,19577],{"className":19576,"style":842},[64],[21,19578],{},", so the score matrix is a rectangular ",[21,19581,19583],{"className":19582},[24],[21,19584,19586,19604],{"className":19585,"ariaHidden":29},[28],[21,19587,19589,19592,19595,19598,19601],{"className":19588},[33],[21,19590],{"className":19591,"style":4149},[37],[21,19593,134],{"className":19594},[42,46],[21,19596],{"className":19597,"style":1087},[401],[21,19599,2550],{"className":19600},[1218],[21,19602],{"className":19603,"style":1087},[401],[21,19605,19607,19610],{"className":19606},[33],[21,19608],{"className":19609,"style":130},[37],[21,19611,2186],{"className":19612},[42,46],"\nlookup — every output position against every source position. Self-attention is\nthe special case ",[21,19615,19617],{"className":19616},[24],[21,19618,19620,19638,19656],{"className":19619,"ariaHidden":29},[28],[21,19621,19623,19626,19629,19632,19635],{"className":19622},[33],[21,19624],{"className":19625,"style":2166},[37],[21,19627,739],{"className":19628},[42,46],[21,19630],{"className":19631,"style":686},[401],[21,19633,691],{"className":19634},[690],[21,19636],{"className":19637,"style":686},[401],[21,19639,19641,19644,19647,19650,19653],{"className":19640},[33],[21,19642],{"className":19643,"style":2200},[37],[21,19645,920],{"className":19646,"style":919},[42,46],[21,19648],{"className":19649,"style":686},[401],[21,19651,691],{"className":19652},[690],[21,19654],{"className":19655,"style":686},[401],[21,19657,19659,19662],{"className":19658},[33],[21,19660],{"className":19661,"style":2200},[37],[21,19663,1088],{"className":19664,"style":1087},[42,46]," from one sequence, giving the square ",[21,19667,19669],{"className":19668},[24],[21,19670,19672,19690],{"className":19671,"ariaHidden":29},[28],[21,19673,19675,19678,19681,19684,19687],{"className":19674},[33],[21,19676],{"className":19677,"style":4149},[37],[21,19679,134],{"className":19680},[42,46],[21,19682],{"className":19683,"style":1087},[401],[21,19685,2550],{"className":19686},[1218],[21,19688],{"className":19689,"style":1087},[401],[21,19691,19693,19696],{"className":19692},[33],[21,19694],{"className":19695,"style":130},[37],[21,19697,134],{"className":19698},[42,46],"\nmatrix; cross-attention keeps queries and keys from ",[163,19701,19702],{},"different"," sequences.",[5230,19705,19707],{"id":19706},"the-causal-mask","The causal mask",[11,19709,19710,19711,19781,19782,19142,19856,5614,19890,19142,19966,19999,20000,20018,20019,20084,20085,20100,20101,20119],{},"The mask is a single additive matrix ",[21,19712,19714],{"className":19713},[24],[21,19715,19717,19737],{"className":19716,"ariaHidden":29},[28],[21,19718,19720,19723,19728,19731,19734],{"className":19719},[33],[21,19721],{"className":19722,"style":4493},[37],[21,19724,19727],{"className":19725,"style":19726},[42,46],"margin-right:0.109em;","M",[21,19729],{"className":19730,"style":686},[401],[21,19732,3730],{"className":19733},[690],[21,19735],{"className":19736,"style":686},[401],[21,19738,19740,19743],{"className":19739},[33],[21,19741],{"className":19742,"style":3740},[37],[21,19744,19746,19749],{"className":19745},[42],[21,19747,773],{"className":19748},[42,772],[21,19750,19752],{"className":19751},[51],[21,19753,19755],{"className":19754},[55],[21,19756,19758],{"className":19757},[60],[21,19759,19761],{"className":19760,"style":3740},[64],[21,19762,19763,19766],{"style":788},[21,19764],{"className":19765,"style":73},[72],[21,19767,19769],{"className":19768},[77,78,79,80],[21,19770,19772,19775,19778],{"className":19771},[42,80],[21,19773,134],{"className":19774},[42,46,80],[21,19776,2550],{"className":19777},[1218,80],[21,19779,134],{"className":19780},[42,46,80]," applied to\nthe scaled scores before the softmax: ",[21,19783,19785],{"className":19784},[24],[21,19786,19788,19847],{"className":19787,"ariaHidden":29},[28],[21,19789,19791,19794,19838,19841,19844],{"className":19790},[33],[21,19792],{"className":19793,"style":701},[37],[21,19795,19797,19800],{"className":19796},[42],[21,19798,19727],{"className":19799,"style":19726},[42,46],[21,19801,19803],{"className":19802},[51],[21,19804,19806,19830],{"className":19805},[55,56],[21,19807,19809,19827],{"className":19808},[60],[21,19810,19812],{"className":19811,"style":243},[64],[21,19813,19815,19818],{"style":19814},"top:-2.55em;margin-left:-0.109em;margin-right:0.05em;",[21,19816],{"className":19817,"style":73},[72],[21,19819,19821],{"className":19820},[77,78,79,80],[21,19822,19824],{"className":19823},[42,80],[21,19825,12930],{"className":19826,"style":256},[42,46,80],[21,19828,89],{"className":19829},[88],[21,19831,19833],{"className":19832},[60],[21,19834,19836],{"className":19835,"style":267},[64],[21,19837],{},[21,19839],{"className":19840,"style":686},[401],[21,19842,691],{"className":19843},[690],[21,19845],{"className":19846,"style":686},[401],[21,19848,19850,19853],{"className":19849},[33],[21,19851],{"className":19852,"style":112},[37],[21,19854,1373],{"className":19855},[42],[21,19857,19859],{"className":19858},[24],[21,19860,19862,19881],{"className":19861,"ariaHidden":29},[28],[21,19863,19865,19868,19871,19874,19878],{"className":19864},[33],[21,19866],{"className":19867,"style":1287},[37],[21,19869,257],{"className":19870,"style":256},[42,46],[21,19872],{"className":19873,"style":686},[401],[21,19875,19877],{"className":19876},[690],"≤",[21,19879],{"className":19880,"style":686},[401],[21,19882,19884,19887],{"className":19883},[33],[21,19885],{"className":19886,"style":3932},[37],[21,19888,3798],{"className":19889},[42,46],[21,19891,19893],{"className":19892},[24],[21,19894,19896,19954],{"className":19895,"ariaHidden":29},[28],[21,19897,19899,19902,19945,19948,19951],{"className":19898},[33],[21,19900],{"className":19901,"style":701},[37],[21,19903,19905,19908],{"className":19904},[42],[21,19906,19727],{"className":19907,"style":19726},[42,46],[21,19909,19911],{"className":19910},[51],[21,19912,19914,19937],{"className":19913},[55,56],[21,19915,19917,19934],{"className":19916},[60],[21,19918,19920],{"className":19919,"style":243},[64],[21,19921,19922,19925],{"style":19814},[21,19923],{"className":19924,"style":73},[72],[21,19926,19928],{"className":19927},[77,78,79,80],[21,19929,19931],{"className":19930},[42,80],[21,19932,12930],{"className":19933,"style":256},[42,46,80],[21,19935,89],{"className":19936},[88],[21,19938,19940],{"className":19939},[60],[21,19941,19943],{"className":19942,"style":267},[64],[21,19944],{},[21,19946],{"className":19947,"style":686},[401],[21,19949,691],{"className":19950},[690],[21,19952],{"className":19953,"style":686},[401],[21,19955,19957,19960,19963],{"className":19956},[33],[21,19958],{"className":19959,"style":4149},[37],[21,19961,7294],{"className":19962},[42],[21,19964,19141],{"className":19965},[42],[21,19967,19969],{"className":19968},[24],[21,19970,19972,19990],{"className":19971,"ariaHidden":29},[28],[21,19973,19975,19978,19981,19984,19987],{"className":19974},[33],[21,19976],{"className":19977,"style":1287},[37],[21,19979,257],{"className":19980,"style":256},[42,46],[21,19982],{"className":19983,"style":686},[401],[21,19985,18319],{"className":19986},[690],[21,19988],{"className":19989,"style":686},[401],[21,19991,19993,19996],{"className":19992},[33],[21,19994],{"className":19995,"style":3932},[37],[21,19997,3798],{"className":19998},[42,46],". Adding ",[21,20001,20003],{"className":20002},[24],[21,20004,20006],{"className":20005,"ariaHidden":29},[28],[21,20007,20009,20012,20015],{"className":20008},[33],[21,20010],{"className":20011,"style":4149},[37],[21,20013,7294],{"className":20014},[42],[21,20016,19141],{"className":20017},[42]," to a logit sends ",[21,20020,20022],{"className":20021},[24],[21,20023,20025,20075],{"className":20024,"ariaHidden":29},[28],[21,20026,20028,20031,20066,20069,20072],{"className":20027},[33],[21,20029],{"className":20030,"style":3740},[37],[21,20032,20034,20037],{"className":20033},[42],[21,20035,10682],{"className":20036},[42,46],[21,20038,20040],{"className":20039},[51],[21,20041,20043],{"className":20042},[55],[21,20044,20046],{"className":20045},[60],[21,20047,20049],{"className":20048,"style":3740},[64],[21,20050,20051,20054],{"style":788},[21,20052],{"className":20053,"style":73},[72],[21,20055,20057],{"className":20056},[77,78,79,80],[21,20058,20060,20063],{"className":20059},[42,80],[21,20061,7294],{"className":20062},[42,80],[21,20064,19141],{"className":20065},[42,80],[21,20067],{"className":20068,"style":686},[401],[21,20070,691],{"className":20071},[690],[21,20073],{"className":20074,"style":686},[401],[21,20076,20078,20081],{"className":20077},[33],[21,20079],{"className":20080,"style":112},[37],[21,20082,1373],{"className":20083},[42],", so the\nsoftmax assigns those future positions weight exactly ",[21,20086,20088],{"className":20087},[24],[21,20089,20091],{"className":20090,"ariaHidden":29},[28],[21,20092,20094,20097],{"className":20093},[33],[21,20095],{"className":20096,"style":112},[37],[21,20098,1373],{"className":20099},[42]," — the reason for the\n",[21,20102,20104],{"className":20103},[24],[21,20105,20107],{"className":20106,"ariaHidden":29},[28],[21,20108,20110,20113,20116],{"className":20109},[33],[21,20111],{"className":20112,"style":4149},[37],[21,20114,7294],{"className":20115},[42],[21,20117,19141],{"className":20118},[42]," rather than merely a large negative number. For four tokens,",[21,20121,20123],{"className":20122},[1582],[21,20124,20126],{"className":20125},[24],[21,20127,20129,20147,20578,20792,20817],{"className":20128,"ariaHidden":29},[28],[21,20130,20132,20135,20138,20141,20144],{"className":20131},[33],[21,20133],{"className":20134,"style":2200},[37],[21,20136,19727],{"className":20137,"style":19726},[42,46],[21,20139],{"className":20140,"style":686},[401],[21,20142,691],{"className":20143},[690],[21,20145],{"className":20146,"style":686},[401],[21,20148,20150,20154,20554,20557,20560,20563,20566,20569,20572,20575],{"className":20149},[33],[21,20151],{"className":20152,"style":20153},[37],"height:4.8em;vertical-align:-2.15em;",[21,20155,20157,20202,20516],{"className":20156},[2521],[21,20158,20160],{"className":20159},[221],[21,20161,20163],{"className":20162},[3285,8611],[21,20164,20166,20193],{"className":20165},[55,56],[21,20167,20169,20190],{"className":20168},[60],[21,20170,20173],{"className":20171,"style":20172},[64],"height:2.65em;",[21,20174,20176,20180],{"style":20175},"top:-4.65em;",[21,20177],{"className":20178,"style":20179},[72],"height:6.8em;",[21,20181,20183],{"style":20182},"width:0.667em;height:4.8em;",[2628,20184,20187],{"xmlns":2630,"width":8634,"height":20185,"viewBox":20186},"4.8em","0 0 667 4800",[2636,20188],{"d":20189},"M403 1759 V84 H666 V0 H319 V1759 v1200 v1759 v84 h347 v-84\nH403z M403 1759 V0 H319 V1759 v1200 v1759 v84 h84z",[21,20191,89],{"className":20192},[88],[21,20194,20196],{"className":20195},[60],[21,20197,20200],{"className":20198,"style":20199},[64],"height:2.15em;",[21,20201],{},[21,20203,20205],{"className":20204},[42],[21,20206,20208,20279,20282,20285,20355,20358,20361,20434,20437,20440],{"className":20207},[8658],[21,20209,20211],{"className":20210},[8662],[21,20212,20214,20271],{"className":20213},[55,56],[21,20215,20217,20268],{"className":20216},[60],[21,20218,20220,20232,20244,20256],{"className":20219,"style":20172},[64],[21,20221,20223,20226],{"style":20222},"top:-4.81em;",[21,20224],{"className":20225,"style":1678},[72],[21,20227,20229],{"className":20228},[42],[21,20230,1373],{"className":20231},[42],[21,20233,20235,20238],{"style":20234},"top:-3.61em;",[21,20236],{"className":20237,"style":1678},[72],[21,20239,20241],{"className":20240},[42],[21,20242,1373],{"className":20243},[42],[21,20245,20247,20250],{"style":20246},"top:-2.41em;",[21,20248],{"className":20249,"style":1678},[72],[21,20251,20253],{"className":20252},[42],[21,20254,1373],{"className":20255},[42],[21,20257,20259,20262],{"style":20258},"top:-1.21em;",[21,20260],{"className":20261,"style":1678},[72],[21,20263,20265],{"className":20264},[42],[21,20266,1373],{"className":20267},[42],[21,20269,89],{"className":20270},[88],[21,20272,20274],{"className":20273},[60],[21,20275,20277],{"className":20276,"style":20199},[64],[21,20278],{},[21,20280],{"className":20281,"style":8723},[8722],[21,20283],{"className":20284,"style":8723},[8722],[21,20286,20288],{"className":20287},[8662],[21,20289,20291,20347],{"className":20290},[55,56],[21,20292,20294,20344],{"className":20293},[60],[21,20295,20297,20311,20322,20333],{"className":20296,"style":20172},[64],[21,20298,20299,20302],{"style":20222},[21,20300],{"className":20301,"style":1678},[72],[21,20303,20305,20308],{"className":20304},[42],[21,20306,7294],{"className":20307},[42],[21,20309,19141],{"className":20310},[42],[21,20312,20313,20316],{"style":20234},[21,20314],{"className":20315,"style":1678},[72],[21,20317,20319],{"className":20318},[42],[21,20320,1373],{"className":20321},[42],[21,20323,20324,20327],{"style":20246},[21,20325],{"className":20326,"style":1678},[72],[21,20328,20330],{"className":20329},[42],[21,20331,1373],{"className":20332},[42],[21,20334,20335,20338],{"style":20258},[21,20336],{"className":20337,"style":1678},[72],[21,20339,20341],{"className":20340},[42],[21,20342,1373],{"className":20343},[42],[21,20345,89],{"className":20346},[88],[21,20348,20350],{"className":20349},[60],[21,20351,20353],{"className":20352,"style":20199},[64],[21,20354],{},[21,20356],{"className":20357,"style":8723},[8722],[21,20359],{"className":20360,"style":8723},[8722],[21,20362,20364],{"className":20363},[8662],[21,20365,20367,20426],{"className":20366},[55,56],[21,20368,20370,20423],{"className":20369},[60],[21,20371,20373,20387,20401,20412],{"className":20372,"style":20172},[64],[21,20374,20375,20378],{"style":20222},[21,20376],{"className":20377,"style":1678},[72],[21,20379,20381,20384],{"className":20380},[42],[21,20382,7294],{"className":20383},[42],[21,20385,19141],{"className":20386},[42],[21,20388,20389,20392],{"style":20234},[21,20390],{"className":20391,"style":1678},[72],[21,20393,20395,20398],{"className":20394},[42],[21,20396,7294],{"className":20397},[42],[21,20399,19141],{"className":20400},[42],[21,20402,20403,20406],{"style":20246},[21,20404],{"className":20405,"style":1678},[72],[21,20407,20409],{"className":20408},[42],[21,20410,1373],{"className":20411},[42],[21,20413,20414,20417],{"style":20258},[21,20415],{"className":20416,"style":1678},[72],[21,20418,20420],{"className":20419},[42],[21,20421,1373],{"className":20422},[42],[21,20424,89],{"className":20425},[88],[21,20427,20429],{"className":20428},[60],[21,20430,20432],{"className":20431,"style":20199},[64],[21,20433],{},[21,20435],{"className":20436,"style":8723},[8722],[21,20438],{"className":20439,"style":8723},[8722],[21,20441,20443],{"className":20442},[8662],[21,20444,20446,20508],{"className":20445},[55,56],[21,20447,20449,20505],{"className":20448},[60],[21,20450,20452,20466,20480,20494],{"className":20451,"style":20172},[64],[21,20453,20454,20457],{"style":20222},[21,20455],{"className":20456,"style":1678},[72],[21,20458,20460,20463],{"className":20459},[42],[21,20461,7294],{"className":20462},[42],[21,20464,19141],{"className":20465},[42],[21,20467,20468,20471],{"style":20234},[21,20469],{"className":20470,"style":1678},[72],[21,20472,20474,20477],{"className":20473},[42],[21,20475,7294],{"className":20476},[42],[21,20478,19141],{"className":20479},[42],[21,20481,20482,20485],{"style":20246},[21,20483],{"className":20484,"style":1678},[72],[21,20486,20488,20491],{"className":20487},[42],[21,20489,7294],{"className":20490},[42],[21,20492,19141],{"className":20493},[42],[21,20495,20496,20499],{"style":20258},[21,20497],{"className":20498,"style":1678},[72],[21,20500,20502],{"className":20501},[42],[21,20503,1373],{"className":20504},[42],[21,20506,89],{"className":20507},[88],[21,20509,20511],{"className":20510},[60],[21,20512,20514],{"className":20513,"style":20199},[64],[21,20515],{},[21,20517,20519],{"className":20518},[273],[21,20520,20522],{"className":20521},[3285,8611],[21,20523,20525,20546],{"className":20524},[55,56],[21,20526,20528,20543],{"className":20527},[60],[21,20529,20531],{"className":20530,"style":20172},[64],[21,20532,20533,20536],{"style":20175},[21,20534],{"className":20535,"style":20179},[72],[21,20537,20538],{"style":20182},[2628,20539,20540],{"xmlns":2630,"width":8634,"height":20185,"viewBox":20186},[2636,20541],{"d":20542},"M347 1759 V0 H0 V84 H263 V1759 v1200 v1759 H0 v84 H347z\nM347 1759 V0 H263 V1759 v1200 v1759 h84z",[21,20544,89],{"className":20545},[88],[21,20547,20549],{"className":20548},[60],[21,20550,20552],{"className":20551,"style":20199},[64],[21,20553],{},[21,20555],{"className":20556,"style":402},[401],[21,20558,1378],{"className":20559},[1377],[21,20561],{"className":20562,"style":1958},[401],[21,20564],{"className":20565,"style":402},[401],[21,20567,3644],{"className":20568},[42,46],[21,20570],{"className":20571,"style":686},[401],[21,20573,691],{"className":20574},[690],[21,20576],{"className":20577,"style":686},[401],[21,20579,20581,20584,20590,20593,20596,20783,20786,20789],{"className":20580},[33],[21,20582],{"className":20583,"style":4757},[37],[21,20585,20587],{"className":20586},[353],[21,20588,3266],{"className":20589},[42,3265],[21,20591],{"className":20592,"style":3270},[401],[21,20594],{"className":20595,"style":402},[401],[21,20597,20599,20605,20777],{"className":20598},[2521],[21,20600,20602],{"className":20601,"style":3281},[221,3280],[21,20603,1369],{"className":20604},[3285,79],[21,20606,20608,20611,20774],{"className":20607},[42],[21,20609],{"className":20610},[221,1657],[21,20612,20614],{"className":20613},[1661],[21,20615,20617,20766],{"className":20616},[55,56],[21,20618,20620,20763],{"className":20619},[60],[21,20621,20623,20715,20723],{"className":20622,"style":4797},[64],[21,20624,20625,20628],{"style":4800},[21,20626],{"className":20627,"style":1678},[72],[21,20629,20631],{"className":20630},[42],[21,20632,20634],{"className":20633},[42,3321],[21,20635,20637,20707],{"className":20636},[55,56],[21,20638,20640,20704],{"className":20639},[60],[21,20641,20643,20692],{"className":20642,"style":3978},[64],[21,20644,20646,20649],{"className":20645,"style":2593},[2612],[21,20647],{"className":20648,"style":1678},[72],[21,20650,20652],{"className":20651,"style":3341},[42],[21,20653,20655,20658],{"className":20654},[42],[21,20656,804],{"className":20657},[42,46],[21,20659,20661],{"className":20660},[51],[21,20662,20664,20684],{"className":20663},[55,56],[21,20665,20667,20681],{"className":20666},[60],[21,20668,20670],{"className":20669,"style":2424},[64],[21,20671,20672,20675],{"style":68},[21,20673],{"className":20674,"style":73},[72],[21,20676,20678],{"className":20677},[77,78,79,80],[21,20679,230],{"className":20680,"style":229},[42,46,80],[21,20682,89],{"className":20683},[88],[21,20685,20687],{"className":20686},[60],[21,20688,20690],{"className":20689,"style":96},[64],[21,20691],{},[21,20693,20694,20697],{"style":4030},[21,20695],{"className":20696,"style":1678},[72],[21,20698,20700],{"className":20699,"style":3392},[3391],[2628,20701,20702],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,20703],{"d":3399},[21,20705,89],{"className":20706},[88],[21,20708,20710],{"className":20709},[60],[21,20711,20713],{"className":20712,"style":4050},[64],[21,20714],{},[21,20716,20717,20720],{"style":1858},[21,20718],{"className":20719,"style":1678},[72],[21,20721],{"className":20722,"style":1866},[1865],[21,20724,20725,20728],{"style":1869},[21,20726],{"className":20727,"style":1678},[72],[21,20729,20731,20734],{"className":20730},[42],[21,20732,739],{"className":20733},[42,46],[21,20735,20737,20740],{"className":20736},[42],[21,20738,920],{"className":20739,"style":919},[42,46],[21,20741,20743],{"className":20742},[51],[21,20744,20746],{"className":20745},[55],[21,20747,20749],{"className":20748},[60],[21,20750,20752],{"className":20751,"style":765},[64],[21,20753,20754,20757],{"style":788},[21,20755],{"className":20756,"style":73},[72],[21,20758,20760],{"className":20759},[77,78,79,80],[21,20761,3137],{"className":20762},[42,80],[21,20764,89],{"className":20765},[88],[21,20767,20769],{"className":20768},[60],[21,20770,20772],{"className":20771,"style":4948},[64],[21,20773],{},[21,20775],{"className":20776},[273,1657],[21,20778,20780],{"className":20779,"style":3281},[273,3280],[21,20781,1388],{"className":20782},[3285,79],[21,20784],{"className":20785,"style":1087},[401],[21,20787,11681],{"className":20788},[1218],[21,20790],{"className":20791,"style":1087},[401],[21,20793,20795,20798,20801,20808,20811,20814],{"className":20794},[33],[21,20796],{"className":20797,"style":2166},[37],[21,20799,19727],{"className":20800,"style":19726},[42,46],[21,20802,20804],{"className":20803},[42,1965],[21,20805,20807],{"className":20806},[42]," pattern",[21,20809],{"className":20810,"style":686},[401],[21,20812,691],{"className":20813},[690],[21,20815],{"className":20816,"style":686},[401],[21,20818,20820,20823,21193,21196],{"className":20819},[33],[21,20821],{"className":20822,"style":20153},[37],[21,20824,20826,20863,21156],{"className":20825},[2521],[21,20827,20829],{"className":20828},[221],[21,20830,20832],{"className":20831},[3285,8611],[21,20833,20835,20855],{"className":20834},[55,56],[21,20836,20838,20852],{"className":20837},[60],[21,20839,20841],{"className":20840,"style":20172},[64],[21,20842,20843,20846],{"style":20175},[21,20844],{"className":20845,"style":20179},[72],[21,20847,20848],{"style":20182},[2628,20849,20850],{"xmlns":2630,"width":8634,"height":20185,"viewBox":20186},[2636,20851],{"d":20189},[21,20853,89],{"className":20854},[88],[21,20856,20858],{"className":20857},[60],[21,20859,20861],{"className":20860,"style":20199},[64],[21,20862],{},[21,20864,20866],{"className":20865},[42],[21,20867,20869,20937,20940,20943,21010,21013,21016,21083,21086,21089],{"className":20868},[8658],[21,20870,20872],{"className":20871},[8662],[21,20873,20875,20929],{"className":20874},[55,56],[21,20876,20878,20926],{"className":20877},[60],[21,20879,20881,20893,20904,20915],{"className":20880,"style":20172},[64],[21,20882,20883,20886],{"style":20222},[21,20884],{"className":20885,"style":1678},[72],[21,20887,20889],{"className":20888},[42],[21,20890,20892],{"className":20891},[42],"∙",[21,20894,20895,20898],{"style":20234},[21,20896],{"className":20897,"style":1678},[72],[21,20899,20901],{"className":20900},[42],[21,20902,20892],{"className":20903},[42],[21,20905,20906,20909],{"style":20246},[21,20907],{"className":20908,"style":1678},[72],[21,20910,20912],{"className":20911},[42],[21,20913,20892],{"className":20914},[42],[21,20916,20917,20920],{"style":20258},[21,20918],{"className":20919,"style":1678},[72],[21,20921,20923],{"className":20922},[42],[21,20924,20892],{"className":20925},[42],[21,20927,89],{"className":20928},[88],[21,20930,20932],{"className":20931},[60],[21,20933,20935],{"className":20934,"style":20199},[64],[21,20936],{},[21,20938],{"className":20939,"style":8723},[8722],[21,20941],{"className":20942,"style":8723},[8722],[21,20944,20946],{"className":20945},[8662],[21,20947,20949,21002],{"className":20948},[55,56],[21,20950,20952,20999],{"className":20951},[60],[21,20953,20955,20966,20977,20988],{"className":20954,"style":20172},[64],[21,20956,20957,20960],{"style":20222},[21,20958],{"className":20959,"style":1678},[72],[21,20961,20963],{"className":20962},[42],[21,20964,1373],{"className":20965},[42],[21,20967,20968,20971],{"style":20234},[21,20969],{"className":20970,"style":1678},[72],[21,20972,20974],{"className":20973},[42],[21,20975,20892],{"className":20976},[42],[21,20978,20979,20982],{"style":20246},[21,20980],{"className":20981,"style":1678},[72],[21,20983,20985],{"className":20984},[42],[21,20986,20892],{"className":20987},[42],[21,20989,20990,20993],{"style":20258},[21,20991],{"className":20992,"style":1678},[72],[21,20994,20996],{"className":20995},[42],[21,20997,20892],{"className":20998},[42],[21,21000,89],{"className":21001},[88],[21,21003,21005],{"className":21004},[60],[21,21006,21008],{"className":21007,"style":20199},[64],[21,21009],{},[21,21011],{"className":21012,"style":8723},[8722],[21,21014],{"className":21015,"style":8723},[8722],[21,21017,21019],{"className":21018},[8662],[21,21020,21022,21075],{"className":21021},[55,56],[21,21023,21025,21072],{"className":21024},[60],[21,21026,21028,21039,21050,21061],{"className":21027,"style":20172},[64],[21,21029,21030,21033],{"style":20222},[21,21031],{"className":21032,"style":1678},[72],[21,21034,21036],{"className":21035},[42],[21,21037,1373],{"className":21038},[42],[21,21040,21041,21044],{"style":20234},[21,21042],{"className":21043,"style":1678},[72],[21,21045,21047],{"className":21046},[42],[21,21048,1373],{"className":21049},[42],[21,21051,21052,21055],{"style":20246},[21,21053],{"className":21054,"style":1678},[72],[21,21056,21058],{"className":21057},[42],[21,21059,20892],{"className":21060},[42],[21,21062,21063,21066],{"style":20258},[21,21064],{"className":21065,"style":1678},[72],[21,21067,21069],{"className":21068},[42],[21,21070,20892],{"className":21071},[42],[21,21073,89],{"className":21074},[88],[21,21076,21078],{"className":21077},[60],[21,21079,21081],{"className":21080,"style":20199},[64],[21,21082],{},[21,21084],{"className":21085,"style":8723},[8722],[21,21087],{"className":21088,"style":8723},[8722],[21,21090,21092],{"className":21091},[8662],[21,21093,21095,21148],{"className":21094},[55,56],[21,21096,21098,21145],{"className":21097},[60],[21,21099,21101,21112,21123,21134],{"className":21100,"style":20172},[64],[21,21102,21103,21106],{"style":20222},[21,21104],{"className":21105,"style":1678},[72],[21,21107,21109],{"className":21108},[42],[21,21110,1373],{"className":21111},[42],[21,21113,21114,21117],{"style":20234},[21,21115],{"className":21116,"style":1678},[72],[21,21118,21120],{"className":21119},[42],[21,21121,1373],{"className":21122},[42],[21,21124,21125,21128],{"style":20246},[21,21126],{"className":21127,"style":1678},[72],[21,21129,21131],{"className":21130},[42],[21,21132,1373],{"className":21133},[42],[21,21135,21136,21139],{"style":20258},[21,21137],{"className":21138,"style":1678},[72],[21,21140,21142],{"className":21141},[42],[21,21143,20892],{"className":21144},[42],[21,21146,89],{"className":21147},[88],[21,21149,21151],{"className":21150},[60],[21,21152,21154],{"className":21153,"style":20199},[64],[21,21155],{},[21,21157,21159],{"className":21158},[273],[21,21160,21162],{"className":21161},[3285,8611],[21,21163,21165,21185],{"className":21164},[55,56],[21,21166,21168,21182],{"className":21167},[60],[21,21169,21171],{"className":21170,"style":20172},[64],[21,21172,21173,21176],{"style":20175},[21,21174],{"className":21175,"style":20179},[72],[21,21177,21178],{"style":20182},[2628,21179,21180],{"xmlns":2630,"width":8634,"height":20185,"viewBox":20186},[2636,21181],{"d":20542},[21,21183,89],{"className":21184},[88],[21,21186,21188],{"className":21187},[60],[21,21189,21191],{"className":21190,"style":20199},[64],[21,21192],{},[21,21194],{"className":21195,"style":402},[401],[21,21197,1378],{"className":21198},[1377],[11,21200,21201,21202,21217,21218,21245],{},"leaving a lower-triangular weight matrix: token ",[21,21203,21205],{"className":21204},[24],[21,21206,21208],{"className":21207,"ariaHidden":29},[28],[21,21209,21211,21214],{"className":21210},[33],[21,21212],{"className":21213,"style":3932},[37],[21,21215,3798],{"className":21216},[42,46]," attends only to tokens ",[21,21219,21221],{"className":21220},[24],[21,21222,21224],{"className":21223,"ariaHidden":29},[28],[21,21225,21227,21230,21233,21236,21239,21242],{"className":21226},[33],[21,21228],{"className":21229,"style":3932},[37],[21,21231,116],{"className":21232},[42],[21,21234],{"className":21235,"style":402},[401],[21,21237,5523],{"className":21238},[2521],[21,21240],{"className":21241,"style":402},[401],[21,21243,3798],{"className":21244},[42,46],". The mask costs nothing at training time (one addition), yet it is what\nlets the decoder train on a whole target sequence in parallel while still\npredicting each token from its prefix alone.",[2128,21247],{"hash":21248},"f28890382bf7cbd4ae1d5275638982b367015f083a1d98df7fc7281fbe5f1b0e",[616,21250,21252],{"id":21251},"recurrence-vs-convolution-vs-self-attention","Recurrence vs. convolution vs. self-attention",[11,21254,21255,21256,21259,21260,21263,21264,2126],{},"Why did attention displace both recurrence and convolution for sequences? Compare\nthe three on the costs that matter: how far information must travel to connect two\npositions (",[158,21257,21258],{},"maximum path length",", which bounds how easily long-range gradients\nflow), how much of the work is inherently ",[158,21261,21262],{},"sequential",", and the ",[158,21265,21266],{},"per-layer\ncompute",[636,21268,21269,21285],{},[639,21270,21271],{},[642,21272,21273,21276,21279,21282],{},[645,21274,21275],{},"Layer type",[645,21277,21278],{},"Per-layer complexity",[645,21280,21281],{},"Sequential ops",[645,21283,21284],{},"Max path length",[658,21286,21287,21402,21591],{},[642,21288,21289,21292,21350,21376],{},[663,21290,21291],{},"Recurrent",[663,21293,21294],{},[21,21295,21297],{"className":21296},[24],[21,21298,21300],{"className":21299,"ariaHidden":29},[28],[21,21301,21303,21306,21309,21312,21315,21318,21347],{"className":21302},[33],[21,21304],{"className":21305,"style":7304},[37],[21,21307,11577],{"className":21308,"style":11576},[42,46],[21,21310,1369],{"className":21311},[221],[21,21313,134],{"className":21314},[42,46],[21,21316],{"className":21317,"style":402},[401],[21,21319,21321,21324],{"className":21320},[42],[21,21322,804],{"className":21323},[42,46],[21,21325,21327],{"className":21326},[51],[21,21328,21330],{"className":21329},[55],[21,21331,21333],{"className":21332},[60],[21,21334,21336],{"className":21335,"style":7197},[64],[21,21337,21338,21341],{"style":788},[21,21339],{"className":21340,"style":73},[72],[21,21342,21344],{"className":21343},[77,78,79,80],[21,21345,5228],{"className":21346},[42,80],[21,21348,1388],{"className":21349},[273],[663,21351,21352],{},[21,21353,21355],{"className":21354},[24],[21,21356,21358],{"className":21357,"ariaHidden":29},[28],[21,21359,21361,21364,21367,21370,21373],{"className":21360},[33],[21,21362],{"className":21363,"style":1365},[37],[21,21365,11577],{"className":21366,"style":11576},[42,46],[21,21368,1369],{"className":21369},[221],[21,21371,134],{"className":21372},[42,46],[21,21374,1388],{"className":21375},[273],[663,21377,21378],{},[21,21379,21381],{"className":21380},[24],[21,21382,21384],{"className":21383,"ariaHidden":29},[28],[21,21385,21387,21390,21393,21396,21399],{"className":21386},[33],[21,21388],{"className":21389,"style":1365},[37],[21,21391,11577],{"className":21392,"style":11576},[42,46],[21,21394,1369],{"className":21395},[221],[21,21397,134],{"className":21398},[42,46],[21,21400,1388],{"className":21401},[273],[642,21403,21404,21424,21488,21514],{},[663,21405,21406,21407,1388],{},"Convolutional (kernel ",[21,21408,21410],{"className":21409},[24],[21,21411,21413],{"className":21412,"ariaHidden":29},[28],[21,21414,21416,21419],{"className":21415},[33],[21,21417],{"className":21418,"style":130},[37],[21,21420,21423],{"className":21421,"style":21422},[42,46],"margin-right:0.0269em;","w",[663,21425,21426],{},[21,21427,21429],{"className":21428},[24],[21,21430,21432],{"className":21431,"ariaHidden":29},[28],[21,21433,21435,21438,21441,21444,21447,21450,21453,21456,21485],{"className":21434},[33],[21,21436],{"className":21437,"style":7304},[37],[21,21439,11577],{"className":21440,"style":11576},[42,46],[21,21442,1369],{"className":21443},[221],[21,21445,21423],{"className":21446,"style":21422},[42,46],[21,21448],{"className":21449,"style":402},[401],[21,21451,134],{"className":21452},[42,46],[21,21454],{"className":21455,"style":402},[401],[21,21457,21459,21462],{"className":21458},[42],[21,21460,804],{"className":21461},[42,46],[21,21463,21465],{"className":21464},[51],[21,21466,21468],{"className":21467},[55],[21,21469,21471],{"className":21470},[60],[21,21472,21474],{"className":21473,"style":7197},[64],[21,21475,21476,21479],{"style":788},[21,21477],{"className":21478,"style":73},[72],[21,21480,21482],{"className":21481},[77,78,79,80],[21,21483,5228],{"className":21484},[42,80],[21,21486,1388],{"className":21487},[273],[663,21489,21490],{},[21,21491,21493],{"className":21492},[24],[21,21494,21496],{"className":21495,"ariaHidden":29},[28],[21,21497,21499,21502,21505,21508,21511],{"className":21498},[33],[21,21500],{"className":21501,"style":1365},[37],[21,21503,11577],{"className":21504,"style":11576},[42,46],[21,21506,1369],{"className":21507},[221],[21,21509,116],{"className":21510},[42],[21,21512,1388],{"className":21513},[273],[663,21515,21516],{},[21,21517,21519],{"className":21518},[24],[21,21520,21522],{"className":21521,"ariaHidden":29},[28],[21,21523,21525,21528,21531,21534,21582,21585,21588],{"className":21524},[33],[21,21526],{"className":21527,"style":1365},[37],[21,21529,11577],{"className":21530,"style":11576},[42,46],[21,21532,1369],{"className":21533},[221],[21,21535,21537,21545],{"className":21536},[353],[21,21538,21540],{"className":21539},[353],[21,21541,21544],{"className":21542,"style":21543},[42,3265],"margin-right:0.0139em;","log",[21,21546,21548],{"className":21547},[51],[21,21549,21551,21573],{"className":21550},[55,56],[21,21552,21554,21570],{"className":21553},[60],[21,21555,21558],{"className":21556,"style":21557},[64],"height:0.0573em;",[21,21559,21561,21564],{"style":21560},"top:-2.4559em;margin-right:0.05em;",[21,21562],{"className":21563,"style":73},[72],[21,21565,21567],{"className":21566},[77,78,79,80],[21,21568,21423],{"className":21569,"style":21422},[42,46,80],[21,21571,89],{"className":21572},[88],[21,21574,21576],{"className":21575},[60],[21,21577,21580],{"className":21578,"style":21579},[64],"height:0.2441em;",[21,21581],{},[21,21583],{"className":21584,"style":402},[401],[21,21586,134],{"className":21587},[42,46],[21,21589,1388],{"className":21590},[273],[642,21592,21593,21595,21650,21676],{},[663,21594,12270],{},[663,21596,21597],{},[21,21598,21600],{"className":21599},[24],[21,21601,21603],{"className":21602,"ariaHidden":29},[28],[21,21604,21606,21609,21612,21615,21644,21647],{"className":21605},[33],[21,21607],{"className":21608,"style":7304},[37],[21,21610,11577],{"className":21611,"style":11576},[42,46],[21,21613,1369],{"className":21614},[221],[21,21616,21618,21621],{"className":21617},[42],[21,21619,134],{"className":21620},[42,46],[21,21622,21624],{"className":21623},[51],[21,21625,21627],{"className":21626},[55],[21,21628,21630],{"className":21629},[60],[21,21631,21633],{"className":21632,"style":7197},[64],[21,21634,21635,21638],{"style":788},[21,21636],{"className":21637,"style":73},[72],[21,21639,21641],{"className":21640},[77,78,79,80],[21,21642,5228],{"className":21643},[42,80],[21,21645,804],{"className":21646},[42,46],[21,21648,1388],{"className":21649},[273],[663,21651,21652],{},[21,21653,21655],{"className":21654},[24],[21,21656,21658],{"className":21657,"ariaHidden":29},[28],[21,21659,21661,21664,21667,21670,21673],{"className":21660},[33],[21,21662],{"className":21663,"style":1365},[37],[21,21665,11577],{"className":21666,"style":11576},[42,46],[21,21668,1369],{"className":21669},[221],[21,21671,116],{"className":21672},[42],[21,21674,1388],{"className":21675},[273],[663,21677,21678],{},[21,21679,21681],{"className":21680},[24],[21,21682,21684],{"className":21683,"ariaHidden":29},[28],[21,21685,21687,21690,21693,21696,21699],{"className":21686},[33],[21,21688],{"className":21689,"style":1365},[37],[21,21691,11577],{"className":21692,"style":11576},[42,46],[21,21694,1369],{"className":21695},[221],[21,21697,116],{"className":21698},[42],[21,21700,1388],{"className":21701},[273],[11,21703,21704,21705,21729,21730,21733,21734,21758,21759,21812,21813,21846,21847,21880],{},"Two columns decide it. Self-attention's maximum path length is ",[21,21706,21708],{"className":21707},[24],[21,21709,21711],{"className":21710,"ariaHidden":29},[28],[21,21712,21714,21717,21720,21723,21726],{"className":21713},[33],[21,21715],{"className":21716,"style":1365},[37],[21,21718,11577],{"className":21719,"style":11576},[42,46],[21,21721,1369],{"className":21722},[221],[21,21724,116],{"className":21725},[42],[21,21727,1388],{"className":21728},[273],": ",[163,21731,21732],{},"any"," two\npositions interact directly, so the gradient between them never has to pass through a\nlong chain. Its sequential depth is ",[21,21735,21737],{"className":21736},[24],[21,21738,21740],{"className":21739,"ariaHidden":29},[28],[21,21741,21743,21746,21749,21752,21755],{"className":21742},[33],[21,21744],{"className":21745,"style":1365},[37],[21,21747,11577],{"className":21748,"style":11576},[42,46],[21,21750,1369],{"className":21751},[221],[21,21753,116],{"className":21754},[42],[21,21756,1388],{"className":21757},[273],", so the whole layer parallelizes\nacross positions on a GPU, unlike the strictly stepwise recurrence. The price is\nthe ",[21,21760,21762],{"className":21761},[24],[21,21763,21765],{"className":21764,"ariaHidden":29},[28],[21,21766,21768,21771,21774,21777,21806,21809],{"className":21767},[33],[21,21769],{"className":21770,"style":7304},[37],[21,21772,11577],{"className":21773,"style":11576},[42,46],[21,21775,1369],{"className":21776},[221],[21,21778,21780,21783],{"className":21779},[42],[21,21781,134],{"className":21782},[42,46],[21,21784,21786],{"className":21785},[51],[21,21787,21789],{"className":21788},[55],[21,21790,21792],{"className":21791},[60],[21,21793,21795],{"className":21794,"style":7197},[64],[21,21796,21797,21800],{"style":788},[21,21798],{"className":21799,"style":73},[72],[21,21801,21803],{"className":21802},[77,78,79,80],[21,21804,5228],{"className":21805},[42,80],[21,21807,804],{"className":21808},[42,46],[21,21810,1388],{"className":21811},[273]," term: the ",[21,21814,21816],{"className":21815},[24],[21,21817,21819,21837],{"className":21818,"ariaHidden":29},[28],[21,21820,21822,21825,21828,21831,21834],{"className":21821},[33],[21,21823],{"className":21824,"style":4149},[37],[21,21826,134],{"className":21827},[42,46],[21,21829],{"className":21830,"style":1087},[401],[21,21832,2550],{"className":21833},[1218],[21,21835],{"className":21836,"style":1087},[401],[21,21838,21840,21843],{"className":21839},[33],[21,21841],{"className":21842,"style":130},[37],[21,21844,134],{"className":21845},[42,46]," attention matrix is quadratic in sequence\nlength, cheap when ",[21,21848,21850],{"className":21849},[24],[21,21851,21853,21871],{"className":21852,"ariaHidden":29},[28],[21,21854,21856,21859,21862,21865,21868],{"className":21855},[33],[21,21857],{"className":21858,"style":19047},[37],[21,21860,134],{"className":21861},[42,46],[21,21863],{"className":21864,"style":686},[401],[21,21866,19051],{"className":21867},[690],[21,21869],{"className":21870,"style":686},[401],[21,21872,21874,21877],{"className":21873},[33],[21,21875],{"className":21876,"style":867},[37],[21,21878,804],{"className":21879},[42,46]," (the usual regime) but the dominant cost for long\nsequences, the bottleneck that later efficient-attention variants attack.",[2128,21882],{"hash":21883},"ccb2861b1d8e505198e6b51f70fe7b1c94937c2596a9ad750b31991176c26408",[616,21885,21887],{"id":21886},"where-attention-came-from","Where attention came from",[11,21889,21890],{},"Attention postdates Goodfellow's 2016 text, so the primary sources carry the\ndetail Goodfellow cannot.",[19010,21892,21893,22204,22444],{},[19013,21894,21895,21898,21899,21902,21903,22099,22100,22103,22104,22203],{},[158,21896,21897],{},"Where attention came from."," The dot-product form here is the streamlined\nversion. The idea started as ",[158,21900,21901],{},"additive attention"," (Bahdanau et al., ICLR 2015),\nwhich scored a query against each key with a small MLP, ",[21,21904,21906],{"className":21905},[24],[21,21907,21909,21949,22047],{"className":21908,"ariaHidden":29},[28],[21,21910,21912,21915,21922,21925,21928,21931,21934,21937,21940,21943,21946],{"className":21911},[33],[21,21913],{"className":21914,"style":1365},[37],[21,21916,21918],{"className":21917},[353],[21,21919,21921],{"className":21920},[42,3265],"score",[21,21923,1369],{"className":21924},[221],[21,21926,203],{"className":21927,"style":202},[42,46],[21,21929,1378],{"className":21930},[1377],[21,21932],{"className":21933,"style":402},[401],[21,21935,230],{"className":21936,"style":229},[42,46],[21,21938,1388],{"className":21939},[273],[21,21941],{"className":21942,"style":686},[401],[21,21944,691],{"className":21945},[690],[21,21947],{"className":21948,"style":686},[401],[21,21950,21952,21956,21985,21988,21992,21995,22035,22038,22041,22044],{"className":21951},[33],[21,21953],{"className":21954,"style":21955},[37],"height:1.1352em;vertical-align:-0.2861em;",[21,21957,21959,21962],{"className":21958},[42],[21,21960,297],{"className":21961,"style":202},[42,46],[21,21963,21965],{"className":21964},[51],[21,21966,21968],{"className":21967},[55],[21,21969,21971],{"className":21970},[60],[21,21972,21974],{"className":21973,"style":765},[64],[21,21975,21976,21979],{"style":788},[21,21977],{"className":21978,"style":73},[72],[21,21980,21982],{"className":21981},[77,78,79,80],[21,21983,3137],{"className":21984},[42,80],[21,21986],{"className":21987,"style":402},[401],[21,21989,21991],{"className":21990},[353],"tanh",[21,21993,1369],{"className":21994},[221],[21,21996,21998,22001],{"className":21997},[42],[21,21999,713],{"className":22000,"style":712},[42,46],[21,22002,22004],{"className":22003},[51],[21,22005,22007,22027],{"className":22006},[55,56],[21,22008,22010,22024],{"className":22009},[60],[21,22011,22013],{"className":22012,"style":2478},[64],[21,22014,22015,22018],{"style":729},[21,22016],{"className":22017,"style":73},[72],[21,22019,22021],{"className":22020},[77,78,79,80],[21,22022,203],{"className":22023,"style":202},[42,46,80],[21,22025,89],{"className":22026},[88],[21,22028,22030],{"className":22029},[60],[21,22031,22033],{"className":22032,"style":267},[64],[21,22034],{},[21,22036,203],{"className":22037,"style":202},[42,46],[21,22039],{"className":22040,"style":1087},[401],[21,22042,11681],{"className":22043},[1218],[21,22045],{"className":22046,"style":1087},[401],[21,22048,22050,22053,22093,22096],{"className":22049},[33],[21,22051],{"className":22052,"style":1365},[37],[21,22054,22056,22059],{"className":22055},[42],[21,22057,713],{"className":22058,"style":712},[42,46],[21,22060,22062],{"className":22061},[51],[21,22063,22065,22085],{"className":22064},[55,56],[21,22066,22068,22082],{"className":22067},[60],[21,22069,22071],{"className":22070,"style":2424},[64],[21,22072,22073,22076],{"style":729},[21,22074],{"className":22075,"style":73},[72],[21,22077,22079],{"className":22078},[77,78,79,80],[21,22080,230],{"className":22081,"style":229},[42,46,80],[21,22083,89],{"className":22084},[88],[21,22086,22088],{"className":22087},[60],[21,22089,22091],{"className":22090,"style":96},[64],[21,22092],{},[21,22094,230],{"className":22095,"style":229},[42,46],[21,22097,1388],{"className":22098},[273],", to let a translation decoder read every encoder\nstate instead of one fixed vector. Vaswani et al., ",[203,22101,22102],{},"Attention Is All You Need","\n(NeurIPS 2017), replaced the MLP with the cheap scaled dot product, added the\n",[21,22105,22107],{"className":22106},[24],[21,22108,22110],{"className":22109,"ariaHidden":29},[28],[21,22111,22113,22116,22119],{"className":22112},[33],[21,22114],{"className":22115,"style":5131},[37],[21,22117,5135],{"className":22118},[42],[21,22120,22122],{"className":22121},[42,3321],[21,22123,22125,22195],{"className":22124},[55,56],[21,22126,22128,22192],{"className":22127},[60],[21,22129,22131,22180],{"className":22130,"style":3978},[64],[21,22132,22134,22137],{"className":22133,"style":2593},[2612],[21,22135],{"className":22136,"style":1678},[72],[21,22138,22140],{"className":22139,"style":3341},[42],[21,22141,22143,22146],{"className":22142},[42],[21,22144,804],{"className":22145},[42,46],[21,22147,22149],{"className":22148},[51],[21,22150,22152,22172],{"className":22151},[55,56],[21,22153,22155,22169],{"className":22154},[60],[21,22156,22158],{"className":22157,"style":2424},[64],[21,22159,22160,22163],{"style":68},[21,22161],{"className":22162,"style":73},[72],[21,22164,22166],{"className":22165},[77,78,79,80],[21,22167,230],{"className":22168,"style":229},[42,46,80],[21,22170,89],{"className":22171},[88],[21,22173,22175],{"className":22174},[60],[21,22176,22178],{"className":22177,"style":96},[64],[21,22179],{},[21,22181,22182,22185],{"style":4030},[21,22183],{"className":22184,"style":1678},[72],[21,22186,22188],{"className":22187,"style":3392},[3391],[2628,22189,22190],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,22191],{"d":3399},[21,22193,89],{"className":22194},[88],[21,22196,22198],{"className":22197},[60],[21,22199,22201],{"className":22200,"style":4050},[64],[21,22202],{}," correction derived above, and removed recurrence entirely.",[19013,22205,22206,22260,22261,22264,22265,22304,22305,22329,22330,22333,22334,22367,22368,22418,22419,22443],{},[158,22207,22208,22209,22259],{},"Reducing the ",[21,22210,22212],{"className":22211},[24],[21,22213,22215],{"className":22214,"ariaHidden":29},[28],[21,22216,22218,22221,22224,22227,22256],{"className":22217},[33],[21,22219],{"className":22220,"style":7304},[37],[21,22222,11577],{"className":22223,"style":11576},[42,46],[21,22225,1369],{"className":22226},[221],[21,22228,22230,22233],{"className":22229},[42],[21,22231,134],{"className":22232},[42,46],[21,22234,22236],{"className":22235},[51],[21,22237,22239],{"className":22238},[55],[21,22240,22242],{"className":22241},[60],[21,22243,22245],{"className":22244,"style":7197},[64],[21,22246,22247,22250],{"style":788},[21,22248],{"className":22249,"style":73},[72],[21,22251,22253],{"className":22252},[77,78,79,80],[21,22254,5228],{"className":22255},[42,80],[21,22257,1388],{"className":22258},[273]," cost."," The quadratic score matrix is the target of a\nwhole line of work. Two directions matter. ",[163,22262,22263],{},"Sparse and linear approximations"," —\nLongformer (Beltagy et al., 2020) with sliding-window attention, and the\nPerformer\u002Flinear-attention family — reduce the asymptotic cost to ",[21,22266,22268],{"className":22267},[24],[21,22269,22271],{"className":22270,"ariaHidden":29},[28],[21,22272,22274,22277,22280,22283,22286,22289,22295,22298,22301],{"className":22273},[33],[21,22275],{"className":22276,"style":1365},[37],[21,22278,11577],{"className":22279,"style":11576},[42,46],[21,22281,1369],{"className":22282},[221],[21,22284,134],{"className":22285},[42,46],[21,22287],{"className":22288,"style":402},[401],[21,22290,22292],{"className":22291},[353],[21,22293,21544],{"className":22294,"style":21543},[42,3265],[21,22296],{"className":22297,"style":402},[401],[21,22299,134],{"className":22300},[42,46],[21,22302,1388],{"className":22303},[273]," or\n",[21,22306,22308],{"className":22307},[24],[21,22309,22311],{"className":22310,"ariaHidden":29},[28],[21,22312,22314,22317,22320,22323,22326],{"className":22313},[33],[21,22315],{"className":22316,"style":1365},[37],[21,22318,11577],{"className":22319,"style":11576},[42,46],[21,22321,1369],{"className":22322},[221],[21,22324,134],{"className":22325},[42,46],[21,22327,1388],{"className":22328},[273]," by never forming the full matrix. ",[163,22331,22332],{},"Exact but IO-aware kernels"," —\nFlashAttention (Dao et al., NeurIPS 2022) — keep the exact softmax but tile the\ncomputation so the ",[21,22335,22337],{"className":22336},[24],[21,22338,22340,22358],{"className":22339,"ariaHidden":29},[28],[21,22341,22343,22346,22349,22352,22355],{"className":22342},[33],[21,22344],{"className":22345,"style":4149},[37],[21,22347,134],{"className":22348},[42,46],[21,22350],{"className":22351,"style":1087},[401],[21,22353,2550],{"className":22354},[1218],[21,22356],{"className":22357,"style":1087},[401],[21,22359,22361,22364],{"className":22360},[33],[21,22362],{"className":22363,"style":130},[37],[21,22365,134],{"className":22366},[42,46]," matrix never leaves fast on-chip memory, cutting the\nmemory footprint from ",[21,22369,22371],{"className":22370},[24],[21,22372,22374],{"className":22373,"ariaHidden":29},[28],[21,22375,22377,22380,22383,22386,22415],{"className":22376},[33],[21,22378],{"className":22379,"style":7304},[37],[21,22381,11577],{"className":22382,"style":11576},[42,46],[21,22384,1369],{"className":22385},[221],[21,22387,22389,22392],{"className":22388},[42],[21,22390,134],{"className":22391},[42,46],[21,22393,22395],{"className":22394},[51],[21,22396,22398],{"className":22397},[55],[21,22399,22401],{"className":22400},[60],[21,22402,22404],{"className":22403,"style":7197},[64],[21,22405,22406,22409],{"style":788},[21,22407],{"className":22408,"style":73},[72],[21,22410,22412],{"className":22411},[77,78,79,80],[21,22413,5228],{"className":22414},[42,80],[21,22416,1388],{"className":22417},[273]," to ",[21,22420,22422],{"className":22421},[24],[21,22423,22425],{"className":22424,"ariaHidden":29},[28],[21,22426,22428,22431,22434,22437,22440],{"className":22427},[33],[21,22429],{"className":22430,"style":1365},[37],[21,22432,11577],{"className":22433,"style":11576},[42,46],[21,22435,1369],{"className":22436},[221],[21,22438,134],{"className":22439},[42,46],[21,22441,1388],{"className":22442},[273]," with no change to the result. The second\napproach is now the default in production Transformers.",[19013,22445,22446,22449,22450,22452],{},[158,22447,22448],{},"Relative and rotary position."," The additive sinusoidal encoding is only the\nfirst of several schemes. Shaw et al. (2018) added learned relative-position biases\ninside the scores; RoPE (Su et al., 2021) rotates the query and key vectors by an\nangle proportional to position, so a dot product depends only on the ",[163,22451,17963],{},"\noffset — the standard choice in modern decoder-only LLMs.",[11,22454,22455,22456,2126],{},"The full architecture built from this mechanism — the stacked encoder–decoder,\ncausal masking, and the parameter\u002Fcompute accounting — is the subject of\n",[15,22457,22459],{"href":22458},"\u002Fdeep-learning\u002Farchitectures\u002Fthe-transformer-architecture","the next lesson",[616,22461,22463],{"id":22462},"takeaways","Takeaways",[19010,22465,22466,22668,22871,22930,22974,22979],{},[19013,22467,22468,22470,22471,2126],{},[158,22469,160],{}," is a differentiable soft lookup: a query scores a set of keys, a\nsoftmax turns the scores into weights, and the output is the weighted blend of\nthe values, ",[21,22472,22474],{"className":22473},[24],[21,22475,22477,22525],{"className":22476,"ariaHidden":29},[28],[21,22478,22480,22483,22489,22492,22495,22498,22501,22504,22507,22510,22513,22516,22519,22522],{"className":22479},[33],[21,22481],{"className":22482,"style":1365},[37],[21,22484,22486],{"className":22485},[353],[21,22487,4714],{"className":22488},[42,3265],[21,22490,1369],{"className":22491},[221],[21,22493,739],{"className":22494},[42,46],[21,22496,1378],{"className":22497},[1377],[21,22499],{"className":22500,"style":402},[401],[21,22502,920],{"className":22503,"style":919},[42,46],[21,22505,1378],{"className":22506},[1377],[21,22508],{"className":22509,"style":402},[401],[21,22511,1088],{"className":22512,"style":1087},[42,46],[21,22514,1388],{"className":22515},[273],[21,22517],{"className":22518,"style":686},[401],[21,22520,691],{"className":22521},[690],[21,22523],{"className":22524,"style":686},[401],[21,22526,22528,22531,22537,22540,22543,22572,22575,22659,22662,22665],{"className":22527},[33],[21,22529],{"className":22530,"style":5131},[37],[21,22532,22534],{"className":22533},[353],[21,22535,3266],{"className":22536},[42,3265],[21,22538,1369],{"className":22539},[221],[21,22541,739],{"className":22542},[42,46],[21,22544,22546,22549],{"className":22545},[42],[21,22547,920],{"className":22548,"style":919},[42,46],[21,22550,22552],{"className":22551},[51],[21,22553,22555],{"className":22554},[55],[21,22556,22558],{"className":22557},[60],[21,22559,22561],{"className":22560,"style":765},[64],[21,22562,22563,22566],{"style":788},[21,22564],{"className":22565,"style":73},[72],[21,22567,22569],{"className":22568},[77,78,79,80],[21,22570,3137],{"className":22571},[42,80],[21,22573,8240],{"className":22574},[42],[21,22576,22578],{"className":22577},[42,3321],[21,22579,22581,22651],{"className":22580},[55,56],[21,22582,22584,22648],{"className":22583},[60],[21,22585,22587,22636],{"className":22586,"style":3978},[64],[21,22588,22590,22593],{"className":22589,"style":2593},[2612],[21,22591],{"className":22592,"style":1678},[72],[21,22594,22596],{"className":22595,"style":3341},[42],[21,22597,22599,22602],{"className":22598},[42],[21,22600,804],{"className":22601},[42,46],[21,22603,22605],{"className":22604},[51],[21,22606,22608,22628],{"className":22607},[55,56],[21,22609,22611,22625],{"className":22610},[60],[21,22612,22614],{"className":22613,"style":2424},[64],[21,22615,22616,22619],{"style":68},[21,22617],{"className":22618,"style":73},[72],[21,22620,22622],{"className":22621},[77,78,79,80],[21,22623,230],{"className":22624,"style":229},[42,46,80],[21,22626,89],{"className":22627},[88],[21,22629,22631],{"className":22630},[60],[21,22632,22634],{"className":22633,"style":96},[64],[21,22635],{},[21,22637,22638,22641],{"style":4030},[21,22639],{"className":22640,"style":1678},[72],[21,22642,22644],{"className":22643,"style":3392},[3391],[2628,22645,22646],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,22647],{"d":3399},[21,22649,89],{"className":22650},[88],[21,22652,22654],{"className":22653},[60],[21,22655,22657],{"className":22656,"style":4050},[64],[21,22658],{},[21,22660,1388],{"className":22661},[273],[21,22663],{"className":22664,"style":402},[401],[21,22666,1088],{"className":22667,"style":1087},[42,46],[19013,22669,22670,22671,22769,22770,22870],{},"The ",[158,22672,22673],{},[21,22674,22676],{"className":22675},[24],[21,22677,22679],{"className":22678,"ariaHidden":29},[28],[21,22680,22682,22685],{"className":22681},[33],[21,22683],{"className":22684,"style":3965},[37],[21,22686,22688],{"className":22687},[42,3321],[21,22689,22691,22761],{"className":22690},[55,56],[21,22692,22694,22758],{"className":22693},[60],[21,22695,22697,22746],{"className":22696,"style":3978},[64],[21,22698,22700,22703],{"className":22699,"style":2593},[2612],[21,22701],{"className":22702,"style":1678},[72],[21,22704,22706],{"className":22705,"style":3341},[42],[21,22707,22709,22712],{"className":22708},[42],[21,22710,804],{"className":22711},[42,46],[21,22713,22715],{"className":22714},[51],[21,22716,22718,22738],{"className":22717},[55,56],[21,22719,22721,22735],{"className":22720},[60],[21,22722,22724],{"className":22723,"style":2424},[64],[21,22725,22726,22729],{"style":68},[21,22727],{"className":22728,"style":73},[72],[21,22730,22732],{"className":22731},[77,78,79,80],[21,22733,230],{"className":22734,"style":229},[42,46,80],[21,22736,89],{"className":22737},[88],[21,22739,22741],{"className":22740},[60],[21,22742,22744],{"className":22743,"style":96},[64],[21,22745],{},[21,22747,22748,22751],{"style":4030},[21,22749],{"className":22750,"style":1678},[72],[21,22752,22754],{"className":22753,"style":3392},[3391],[2628,22755,22756],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,22757],{"d":3399},[21,22759,89],{"className":22760},[88],[21,22762,22764],{"className":22763},[60],[21,22765,22767],{"className":22766,"style":4050},[64],[21,22768],{}," scaling cancels the dot product's variance growth\n(",[21,22771,22773],{"className":22772},[24],[21,22774,22776,22803,22824],{"className":22775,"ariaHidden":29},[28],[21,22777,22779,22782,22788,22791,22794,22797,22800],{"className":22778},[33],[21,22780],{"className":22781,"style":1365},[37],[21,22783,22785],{"className":22784},[353],[21,22786,5931],{"className":22787},[42,3265],[21,22789,1369],{"className":22790},[221],[21,22792,203],{"className":22793,"style":202},[42,46],[21,22795],{"className":22796,"style":1087},[401],[21,22798,1219],{"className":22799},[1218],[21,22801],{"className":22802,"style":1087},[401],[21,22804,22806,22809,22812,22815,22818,22821],{"className":22805},[33],[21,22807],{"className":22808,"style":1365},[37],[21,22810,230],{"className":22811,"style":229},[42,46],[21,22813,1388],{"className":22814},[273],[21,22816],{"className":22817,"style":686},[401],[21,22819,691],{"className":22820},[690],[21,22822],{"className":22823,"style":686},[401],[21,22825,22827,22830],{"className":22826},[33],[21,22828],{"className":22829,"style":38},[37],[21,22831,22833,22836],{"className":22832},[42],[21,22834,804],{"className":22835},[42,46],[21,22837,22839],{"className":22838},[51],[21,22840,22842,22862],{"className":22841},[55,56],[21,22843,22845,22859],{"className":22844},[60],[21,22846,22848],{"className":22847,"style":2424},[64],[21,22849,22850,22853],{"style":68},[21,22851],{"className":22852,"style":73},[72],[21,22854,22856],{"className":22855},[77,78,79,80],[21,22857,230],{"className":22858,"style":229},[42,46,80],[21,22860,89],{"className":22861},[88],[21,22863,22865],{"className":22864},[60],[21,22866,22868],{"className":22867,"style":96},[64],[21,22869],{},"), keeping the softmax out of its\nsaturated, zero-gradient regime.",[19013,22872,22873,22875,22876,22909,22910,22925,22926,22929],{},[158,22874,12270],{}," (",[21,22877,22879],{"className":22878},[24],[21,22880,22882],{"className":22881,"ariaHidden":29},[28],[21,22883,22885,22888,22891,22894,22897,22900,22903,22906],{"className":22884},[33],[21,22886],{"className":22887,"style":2166},[37],[21,22889,739],{"className":22890},[42,46],[21,22892,1378],{"className":22893},[1377],[21,22895],{"className":22896,"style":402},[401],[21,22898,920],{"className":22899,"style":919},[42,46],[21,22901,1378],{"className":22902},[1377],[21,22904],{"className":22905,"style":402},[401],[21,22907,1088],{"className":22908,"style":1087},[42,46]," from one sequence) yields an all-pairs routing\nmatrix ",[21,22911,22913],{"className":22912},[24],[21,22914,22916],{"className":22915,"ariaHidden":29},[28],[21,22917,22919,22922],{"className":22918},[33],[21,22920],{"className":22921,"style":2200},[37],[21,22923,3644],{"className":22924},[42,46],"; being permutation-equivariant, it needs ",[158,22927,22928],{},"positional encoding"," —\nsinusoidal or learned — to know order.",[19013,22931,22932,22935,22936,22951,22952,22973],{},[158,22933,22934],{},"Multi-head"," attention runs ",[21,22937,22939],{"className":22938},[24],[21,22940,22942],{"className":22941,"ariaHidden":29},[28],[21,22943,22945,22948],{"className":22944},[33],[21,22946],{"className":22947,"style":867},[37],[21,22949,47],{"className":22950},[42,46]," parallel attentions in ",[21,22953,22955],{"className":22954},[24],[21,22956,22958],{"className":22957,"ariaHidden":29},[28],[21,22959,22961,22964,22967,22970],{"className":22960},[33],[21,22962],{"className":22963,"style":1365},[37],[21,22965,804],{"className":22966},[42,46],[21,22968,8240],{"className":22969},[42],[21,22971,47],{"className":22972},[42,46],"-dimensional\nsubspaces, capturing several relationship patterns at once for the cost of one.",[19013,22975,22670,22976,22978],{},[158,22977,18521],{}," = multi-head self-attention + position-wise FFN, each\nin a residual + LayerNorm wrapper; decoders add a causal mask and\ncross-attention.",[19013,22980,22981,22982,22985,22986,23010,23011,22875,23014,23038,23039,23092],{},"Against recurrence and convolution, self-attention wins on ",[158,22983,22984],{},"path length","\n(",[21,22987,22989],{"className":22988},[24],[21,22990,22992],{"className":22991,"ariaHidden":29},[28],[21,22993,22995,22998,23001,23004,23007],{"className":22994},[33],[21,22996],{"className":22997,"style":1365},[37],[21,22999,11577],{"className":23000,"style":11576},[42,46],[21,23002,1369],{"className":23003},[221],[21,23005,116],{"className":23006},[42],[21,23008,1388],{"className":23009},[273],") and ",[158,23012,23013],{},"parallelism",[21,23015,23017],{"className":23016},[24],[21,23018,23020],{"className":23019,"ariaHidden":29},[28],[21,23021,23023,23026,23029,23032,23035],{"className":23022},[33],[21,23024],{"className":23025,"style":1365},[37],[21,23027,11577],{"className":23028,"style":11576},[42,46],[21,23030,1369],{"className":23031},[221],[21,23033,116],{"className":23034},[42],[21,23036,1388],{"className":23037},[273]," sequential ops), paying an ",[21,23040,23042],{"className":23041},[24],[21,23043,23045],{"className":23044,"ariaHidden":29},[28],[21,23046,23048,23051,23054,23057,23086,23089],{"className":23047},[33],[21,23049],{"className":23050,"style":7304},[37],[21,23052,11577],{"className":23053,"style":11576},[42,46],[21,23055,1369],{"className":23056},[221],[21,23058,23060,23063],{"className":23059},[42],[21,23061,134],{"className":23062},[42,46],[21,23064,23066],{"className":23065},[51],[21,23067,23069],{"className":23068},[55],[21,23070,23072],{"className":23071},[60],[21,23073,23075],{"className":23074,"style":7197},[64],[21,23076,23077,23080],{"style":788},[21,23078],{"className":23079,"style":73},[72],[21,23081,23083],{"className":23082},[77,78,79,80],[21,23084,5228],{"className":23085},[42,80],[21,23087,804],{"className":23088},[42,46],[21,23090,1388],{"className":23091},[273],"\nper-layer cost — the trade that made it the dominant sequence architecture.",[23094,23095,23098,23103],"section",{"className":23096,"dataFootnotes":6},[23097],"footnotes",[616,23099,23102],{"className":23100,"id":174},[23101],"sr-only","Footnotes",[23104,23105,23106,23123,23242,23345,23358],"ol",{},[19013,23107,23109,9562,23112,23115,23116],{"id":23108},"user-content-fn-gf-attention",[158,23110,23111],{},"Goodfellow",[163,23113,23114],{},"Deep Learning",", §10.11 — Explicit Memory & attention over memory cells: content-based addressing as a differentiable soft lookup (the Transformer itself postdates the 2016 text). ",[15,23117,23122],{"href":23118,"ariaLabel":23119,"className":23120,"dataFootnoteBackref":6},"#user-content-fnref-gf-attention","Back to reference 1",[23121],"data-footnote-backref","↩",[19013,23124,23126,9562,23129,23132,23133,23232,23233,23236,23237],{"id":23125},"user-content-fn-chollet-attn",[158,23127,23128],{},"Chollet",[163,23130,23131],{},"Deep Learning with Python",", Ch. 11 — Deep Learning for Text: scaled dot-product attention, the ",[21,23134,23136],{"className":23135},[24],[21,23137,23139],{"className":23138,"ariaHidden":29},[28],[21,23140,23142,23145,23148],{"className":23141},[33],[21,23143],{"className":23144,"style":5131},[37],[21,23146,5135],{"className":23147},[42],[21,23149,23151],{"className":23150},[42,3321],[21,23152,23154,23224],{"className":23153},[55,56],[21,23155,23157,23221],{"className":23156},[60],[21,23158,23160,23209],{"className":23159,"style":3978},[64],[21,23161,23163,23166],{"className":23162,"style":2593},[2612],[21,23164],{"className":23165,"style":1678},[72],[21,23167,23169],{"className":23168,"style":3341},[42],[21,23170,23172,23175],{"className":23171},[42],[21,23173,804],{"className":23174},[42,46],[21,23176,23178],{"className":23177},[51],[21,23179,23181,23201],{"className":23180},[55,56],[21,23182,23184,23198],{"className":23183},[60],[21,23185,23187],{"className":23186,"style":2424},[64],[21,23188,23189,23192],{"style":68},[21,23190],{"className":23191,"style":73},[72],[21,23193,23195],{"className":23194},[77,78,79,80],[21,23196,230],{"className":23197,"style":229},[42,46,80],[21,23199,89],{"className":23200},[88],[21,23202,23204],{"className":23203},[60],[21,23205,23207],{"className":23206,"style":96},[64],[21,23208],{},[21,23210,23211,23214],{"style":4030},[21,23212],{"className":23213,"style":1678},[72],[21,23215,23217],{"className":23216,"style":3392},[3391],[2628,23218,23219],{"xmlns":2630,"width":2631,"height":3395,"viewBox":3396,"preserveAspectRatio":2634},[2636,23220],{"d":3399},[21,23222,89],{"className":23223},[88],[21,23225,23227],{"className":23226},[60],[21,23228,23230],{"className":23229,"style":4050},[64],[21,23231],{}," correction, and the ",[8360,23234,23235],{},"softmax(QK^T\u002F\\sqrt{d_k})V"," core of the Transformer. ",[15,23238,23122],{"href":23239,"ariaLabel":23240,"className":23241,"dataFootnoteBackref":6},"#user-content-fnref-chollet-attn","Back to reference 2",[23121],[19013,23243,23245,9562,23247,23249,23250,23265,23266,23287,23288,7790,23340],{"id":23244},"user-content-fn-chollet-mha",[158,23246,23128],{},[163,23248,23131],{},", Ch. 11 — Multi-Head Attention: ",[21,23251,23253],{"className":23252},[24],[21,23254,23256],{"className":23255,"ariaHidden":29},[28],[21,23257,23259,23262],{"className":23258},[33],[21,23260],{"className":23261,"style":867},[37],[21,23263,47],{"className":23264},[42,46]," parallel attention functions in ",[21,23267,23269],{"className":23268},[24],[21,23270,23272],{"className":23271,"ariaHidden":29},[28],[21,23273,23275,23278,23281,23284],{"className":23274},[33],[21,23276],{"className":23277,"style":1365},[37],[21,23279,804],{"className":23280},[42,46],[21,23282,8240],{"className":23283},[42],[21,23285,47],{"className":23286},[42,46],"-dimensional subspaces, concatenated and projected by ",[21,23289,23291],{"className":23290},[24],[21,23292,23294],{"className":23293,"ariaHidden":29},[28],[21,23295,23297,23300],{"className":23296},[33],[21,23298],{"className":23299,"style":886},[37],[21,23301,23303,23306],{"className":23302},[42],[21,23304,713],{"className":23305,"style":712},[42,46],[21,23307,23309],{"className":23308},[51],[21,23310,23312,23332],{"className":23311},[55,56],[21,23313,23315,23329],{"className":23314},[60],[21,23316,23318],{"className":23317,"style":726},[64],[21,23319,23320,23323],{"style":729},[21,23321],{"className":23322,"style":73},[72],[21,23324,23326],{"className":23325},[77,78,79,80],[21,23327,11577],{"className":23328,"style":11576},[42,46,80],[21,23330,89],{"className":23331},[88],[21,23333,23335],{"className":23334},[60],[21,23336,23338],{"className":23337,"style":96},[64],[21,23339],{},[15,23341,23122],{"href":23342,"ariaLabel":23343,"className":23344,"dataFootnoteBackref":6},"#user-content-fnref-chollet-mha","Back to reference 3",[23121],[19013,23346,23348,9562,23350,23352,23353],{"id":23347},"user-content-fn-chollet-pos",[158,23349,23128],{},[163,23351,23131],{},", Ch. 11 — Positional encoding: injecting order into a permutation-equivariant model, sinusoidal versus learned embeddings. ",[15,23354,23122],{"href":23355,"ariaLabel":23356,"className":23357,"dataFootnoteBackref":6},"#user-content-fnref-chollet-pos","Back to reference 4",[23121],[19013,23359,23361,9562,23363,23365,23366],{"id":23360},"user-content-fn-chollet-block",[158,23362,23128],{},[163,23364,23131],{},", Ch. 11 — The Transformer encoder block: multi-head self-attention plus a position-wise feed-forward network, each in a residual + normalization wrapper. ",[15,23367,23122],{"href":23368,"ariaLabel":23369,"className":23370,"dataFootnoteBackref":6},"#user-content-fnref-chollet-block","Back to reference 5",[23121],[23372,23373,23374],"style",{},"html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark-mode .shiki span {color: var(--shiki-dark-mode);background: var(--shiki-dark-mode-bg);font-style: var(--shiki-dark-mode-font-style);font-weight: var(--shiki-dark-mode-font-weight);text-decoration: var(--shiki-dark-mode-text-decoration);}html.dark-mode .shiki span {color: var(--shiki-dark-mode);background: var(--shiki-dark-mode-bg);font-style: var(--shiki-dark-mode-font-style);font-weight: var(--shiki-dark-mode-font-weight);text-decoration: var(--shiki-dark-mode-text-decoration);}",{"title":6,"searchDepth":8371,"depth":8371,"links":23376},[23377,23378,23383,23386,23387,23388,23392,23393,23394,23395],{"id":618,"depth":8371,"text":619},{"id":2133,"depth":8371,"text":2134,"children":23379},[23380,23382],{"id":5232,"depth":8377,"text":23381},"Why the dk​​ scaling",{"id":8419,"depth":8377,"text":8420},{"id":12269,"depth":8371,"text":12270,"children":23384},[23385],{"id":13149,"depth":8377,"text":13150},{"id":14258,"depth":8371,"text":14259},{"id":17352,"depth":8371,"text":17353},{"id":18506,"depth":8371,"text":18507,"children":23389},[23390,23391],{"id":18951,"depth":8377,"text":18952},{"id":19706,"depth":8377,"text":19707},{"id":21251,"depth":8371,"text":21252},{"id":21886,"depth":8371,"text":21887},{"id":22462,"depth":8371,"text":22463},{"id":174,"depth":8371,"text":23102},[],"computer-science","A recurrent network reads a\nsequence one step at a time, squeezing the entire past into a single hidden state\nht​. That bottleneck is the defect: information from position 1 reaches\nposition n only after n sequential hops, and the gradient connecting them\ndecays along the way: the very problem LSTMs and\nGRUs only partially patch.\nAttention removes the hops entirely. Instead of routing information through a\nchain, every position reads directly from every other position in one step,\nweighting what it reads by learned, content-dependent relevance.1",false,"md",{"moduleNumber":8389,"lessonNumber":8389,"order":23402},505,"Architectures",true,"\u002Fdeep-learning\u002Farchitectures\u002Fattention-and-transformers",[],"---\ntitle: Attention & Transformers\nmodule: Architectures\nmoduleNumber: 5\nlessonNumber: 5\norder: 505\nsummary: >\n  Attention replaces fixed wiring with content-based routing: every position\n  reads from every other through a soft, learned dot-product lookup. We derive\n  scaled dot-product attention and its $\\sqrt{d_k}$ correction, build it into\n  multi-head self-attention, inject order with positional encodings, and stack\n  the whole thing into the Transformer block that displaced recurrence and\n  convolution alike.\ntopics: [Architectures]\nsources:\n  - book: Goodfellow\n    ref: \"§10.11 — Explicit Memory & Attention (the Transformer postdates the 2016 text)\"\n  - book: Chollet\n    ref: \"Ch. 11 — Deep Learning for Text; The Transformer Architecture\"\n---\n\nA [recurrent network](\u002Fdeep-learning\u002Farchitectures\u002Frecurrent-networks) reads a\nsequence one step at a time, squeezing the entire past into a single hidden state\n$h_t$. That bottleneck is the defect: information from position $1$ reaches\nposition $n$ only after $n$ sequential hops, and the gradient connecting them\ndecays along the way: the very problem [LSTMs and\nGRUs](\u002Fdeep-learning\u002Farchitectures\u002Flstm-and-gru) only partially patch.\n**Attention** removes the hops entirely. Instead of routing information through a\nchain, every position reads directly from every other position in _one_ step,\nweighting what it reads by learned, content-dependent relevance.[^gf-attention]\n\n> **Definition (Attention).** A differentiable lookup. Given a query vector $q$, a\n> set of keys $\\{k_j\\}$, and an associated set of values $\\{v_j\\}$, attention\n> returns a convex combination $\\sum_j \\alpha_j v_j$ of the values, where the\n> weights $\\alpha_j$ measure the compatibility between $q$ and each key $k_j$. It\n> is a _soft_ dictionary: a hard lookup returns the one value whose key matches;\n> attention returns a weighted blend of all of them.\n\n## Queries, keys, and values\n\nThe vocabulary is borrowed from retrieval. A **query** is what a position is\nlooking for; a **key** advertises what each position offers; a **value** is the\ncontent actually returned. All three are linear projections of the inputs, so the\nmechanism is fully learned: three weight matrices determine what each token\nsearches for, what it exposes for matching, and what content it returns.\n\n| Object | Symbol | Shape | Role |\n| --- | --- | --- | --- |\n| Query | $q = x W_Q$ | $\\mathbb{R}^{d_k}$ | what this position is searching for |\n| Key | $k = x W_K$ | $\\mathbb{R}^{d_k}$ | what each position offers for matching |\n| Value | $v = x W_V$ | $\\mathbb{R}^{d_v}$ | the content returned when a key matches |\n| Score | $q \\cdot k_j$ | scalar | raw compatibility of the query with key $j$ |\n| Weight | $\\alpha_j$ | scalar in $(0,1)$ | softmax-normalized score, $\\sum_j \\alpha_j = 1$ |\n\nCompatibility is measured by a dot product: $q \\cdot k_j$ is large when the query\nand key point the same way. Normalizing the scores with a softmax turns them into\na probability distribution over positions, and the output is the expectation of\nthe values under that distribution,\n\n$$\n\\alpha_j = \\frac{\\exp(q \\cdot k_j)}{\\sum_{j'} \\exp(q \\cdot k_{j'})},\n\\qquad\n\\text{out} = \\sum_j \\alpha_j\\, v_j .\n$$\n\n$$\n% caption: Scaled dot-product attention as a dataflow: queries and keys give scaled scores, a softmax turns them into weights, and multiplying by $V$ blends the values.\n\\begin{tikzpicture}[>=stealth, font=\\footnotesize,\n  box\u002F.style={draw, minimum width=16mm, minimum height=9mm, align=center},\n  op\u002F.style={draw, minimum width=15mm, minimum height=9mm, align=center}]\n  \\definecolor{acc}{HTML}{2348F2}\n  \\node[box, draw=acc, text=acc] (Q) at (0,2.6)   {Q};\n  \\node[box, draw=acc, text=acc] (K) at (0,0)     {K};\n  \\node[box, draw=acc, text=acc] (V) at (0,-2.6)  {V};\n  \\node[op]  (mm1) at (2.8,1.3)  {\\texttt{matmul}};\n  \\node[box] (sc)  at (5.4,1.3)  {\\texttt{scale}};\n  \\node[box] (sm)  at (7.9,1.3)  {\\texttt{softmax}};\n  \\node[op]  (mm2) at (10.4,-0.6)  {\\texttt{matmul}};\n  \\node[box, draw=acc, text=acc] (out) at (13.0,-0.6) {\\texttt{output}};\n  \\draw[->, acc, thick] (Q) -- (mm1);\n  \\draw[->, acc, thick] (K) -- (mm1);\n  \\draw[->, thick] (mm1) -- (sc) node[midway, above, font=\\footnotesize] {\\texttt{scores}};\n  \\draw[->, thick] (sc) -- (sm);\n  \\draw[->, thick] (sm) -- (mm2);\n  \\node[font=\\footnotesize] at (9.7,0.9) {\\texttt{weights}};\n  \\draw[->, acc, thick] (V) -- (mm2);\n  \\draw[->, acc, thick] (mm2) -- (out);\n\\end{tikzpicture}\n$$\n\n## Scaled dot-product attention\n\nStack the $n$ queries into the rows of $Q$, the $m$ keys into the rows of $K$,\nand the $m$ values into the rows of $V$. Keeping the query count $n$ and the\nkey\u002Fvalue count $m$ separate matters: in self-attention $n = m$, but in\ncross-attention a decoder of length $n$ reads an encoder of length $m \\ne n$.\nQueries and keys must share a width $d_k$ so their dot product is defined; values\ncarry an independent width $d_v$. The shapes chain as follows.\n\n$$\n\\underbrace{Q}_{n \\times d_k} \\qquad\n\\underbrace{K}_{m \\times d_k} \\qquad\n\\underbrace{V}_{m \\times d_v} \\qquad\n\\underbrace{QK^\\top}_{n \\times m} \\qquad\n\\underbrace{\\softmax\\!\\parens{\\tfrac{QK^\\top}{\\sqrt{d_k}}}}_{n \\times m} \\qquad\n\\underbrace{AV}_{n \\times d_v} .\n$$\n\nEvery query-key score is a single matrix product $QK^\\top \\in \\mathbb{R}^{n\n\\times m}$: entry $(i,j)$ is the dot product $q_i \\cdot k_j$ of query $i$ with key\n$j$. Divide by $\\sqrt{d_k}$, run a softmax along each row so that row $i$ becomes\na probability distribution over the $m$ keys, and right-multiply by $V$. The\ninner dimension $m$ cancels in $A V$, so an $n \\times m$ weight matrix times an $m\n\\times d_v$ value matrix returns an $n \\times d_v$ output — one $d_v$-vector per\nquery, exactly the input query count. The whole operation collapses to one\nexpression.\n\n> **Definition (Scaled dot-product attention).** For queries $Q \\in \\mathbb{R}^{n\n> \\times d_k}$, keys $K \\in \\mathbb{R}^{m \\times d_k}$, and values $V \\in\n> \\mathbb{R}^{m \\times d_v}$,\n> $$\n> \\Attn(Q, K, V) = \\softmax\\!\\parens{\\frac{QK^\\top}{\\sqrt{d_k}}} V\n> \\;\\in\\; \\mathbb{R}^{n \\times d_v} .\n> $$\n> The softmax is applied independently to each row, so row $i$ of the output is a\n> convex combination of the $m$ value vectors, weighted by how well query $i$\n> matches every key.\n\nThe only unexplained piece is the $1\u002F\\sqrt{d_k}$ factor. It is essential:\nwithout it the softmax saturates and gradients vanish.[^chollet-attn]\n\n### Why the $\\sqrt{d_k}$ scaling\n\nTreat the entries of a query $q$ and a key $k$ as independent random variables\nwith mean $0$ and variance $1$. The dot product is a sum of $d_k$ such products,\nand we can compute the variance of that sum exactly.\n\n> **Lemma (Variance of the dot product).** If $q_1, \\dots, q_{d_k}$ and $k_1,\n> \\dots, k_{d_k}$ are mutually independent with $\\mathbb{E}[q_i] = \\mathbb{E}[k_i]\n> = 0$ and $\\Var(q_i) = \\Var(k_i) = 1$, then the dot\n> product $q \\cdot k = \\sum_{i=1}^{d_k} q_i k_i$ has mean $0$ and variance $d_k$.\n\n> **Proof.** Each term $q_i k_i$ has mean $\\mathbb{E}[q_i k_i] = \\mathbb{E}[q_i]\\,\n> \\mathbb{E}[k_i] = 0$ by independence, so $\\mathbb{E}[q \\cdot k] = 0$. For the\n> variance, independence across $i$ makes the terms uncorrelated, so variances\n> add:\n> $$\n> \\Var(q \\cdot k) = \\sum_{i=1}^{d_k} \\Var(q_i k_i).\n> $$\n> For a single term, $\\Var(q_i k_i) = \\mathbb{E}[q_i^2 k_i^2] -\n> (\\mathbb{E}[q_i k_i])^2 = \\mathbb{E}[q_i^2]\\,\\mathbb{E}[k_i^2] - 0 = 1 \\cdot 1 =\n> 1$, again using independence. Summing $d_k$ unit-variance terms gives\n> $\\Var(q \\cdot k) = d_k$. $\\qed$\n\nSo the raw scores have standard deviation $\\sqrt{d_k}$, which _grows_ with the\nmodel dimension. Feed scores of magnitude $\\sim\\!\\sqrt{d_k}$ into a softmax and,\nfor $d_k$ in the hundreds, the largest entry dwarfs the rest: the softmax\nsaturates onto one position, approaching a one-hot vector. In that regime its\nJacobian is nearly zero, and almost no gradient flows back to the queries and keys.\nDividing by $\\sqrt{d_k}$ rescales the scores to unit variance, $\\Var\n\\!\\parens{ (q \\cdot k)\u002F\\sqrt{d_k}} = 1$, keeping the softmax in its sensitive,\nhigh-gradient region regardless of dimension.\n\n$$\n% caption: Without the $1\u002F\\sqrt{d_k}$ correction the softmax saturates: large-variance\n% scores (black) collapse onto one key, while scaled scores (blue) stay usable.\n\\begin{tikzpicture}[>=stealth, font=\\footnotesize]\n  \\definecolor{acc}{HTML}{2348F2}\n  \\draw[->, thick] (-0.3,0) -- (6.6,0) node[right, font=\\scriptsize] {key index};\n  \\draw[->, thick] (0,-0.2) -- (0,3.0) node[above, font=\\scriptsize] {weight};\n  % saturated (neutral): one tall spike\n  \\foreach \\x\u002F\\h in {0.6\u002F0.15, 1.4\u002F0.2, 2.2\u002F2.55, 3.0\u002F0.25, 3.8\u002F0.15, 4.6\u002F0.18, 5.4\u002F0.12}\n    \\draw[black, thick, fill=black!12] (\\x-0.14,0) rectangle (\\x+0.0,\\h);\n  % scaled (blue): spread distribution, offset to the right of each neutral bar\n  \\foreach \\x\u002F\\h in {0.6\u002F0.55, 1.4\u002F0.75, 2.2\u002F1.35, 3.0\u002F1.0, 3.8\u002F0.7, 4.6\u002F0.5, 5.4\u002F0.35}\n    \\draw[acc, thick, fill=acc!18] (\\x+0.02,0) rectangle (\\x+0.16,\\h);\n  \\node[black, font=\\scriptsize, anchor=west] at (2.9,2.4) {unscaled (saturated)};\n  \\node[acc, font=\\scriptsize, anchor=west] at (3.2,1.55) {scaled};\n\\end{tikzpicture}\n$$\n\n```algorithm\ncaption: $\\textsc{ScaledDotProductAttention}(Q, K, V)$ — soft content-based lookup\n$S \\gets Q K^{T}$ \u002F\u002F $n \\times n$ raw scores\n$S \\gets S \u002F \\sqrt{d_k}$ \u002F\u002F unit-variance correction\nif masked then \u002F\u002F decoder self-attention\n  for each $i, j$ with $j > i$ do\n    $S_{ij} \\gets -\\infty$ \u002F\u002F forbid attending to the future\nfor each row $i$ do\n  $A_{i} \\gets \\softmax(S_{i})$ \u002F\u002F weights sum to $1$\nreturn $A V$ \u002F\u002F blend the values\n```\n\n### A worked example\n\nPush numbers through the formula. Take three\ntokens with $d_k = d_v = 2$, and, to keep the arithmetic legible, let the\nqueries and keys coincide, so this is self-attention:\n\n$$\nQ = K = \\begin{bmatrix} 1 & 0 \\\\ 0 & 1 \\\\ 1 & 1 \\end{bmatrix}\n\\;\\in\\; \\mathbb{R}^{3\\times 2},\n\\qquad\nV = \\begin{bmatrix} 1 & 0 \\\\ 0 & 1 \\\\ 0.5 & 0.5 \\end{bmatrix}\n\\;\\in\\; \\mathbb{R}^{3\\times 2}.\n$$\n\n**Step 1 — raw scores $S = QK^\\top$.** Each entry is $q_i \\cdot k_j$. Row $1$ is\nquery $(1,0)$ dotted against the three keys: $(1,0)\\cdot(1,0)=1$,\n$(1,0)\\cdot(0,1)=0$, $(1,0)\\cdot(1,1)=1$. Filling in all nine,\n\n$$\nS = QK^\\top = \\begin{bmatrix} 1 & 0 & 1 \\\\ 0 & 1 & 1 \\\\ 1 & 1 & 2 \\end{bmatrix}\n\\;\\in\\; \\mathbb{R}^{3\\times 3}.\n$$\n\n**Step 2 — scale by $\\sqrt{d_k} = \\sqrt 2 \\approx 1.414$.** Every score shrinks by\nthe same factor, pulling the logits toward zero before the softmax:\n\n$$\n\\frac{S}{\\sqrt 2} = \\begin{bmatrix} 0.707 & 0 & 0.707 \\\\ 0 & 0.707 & 0.707 \\\\ 0.707 & 0.707 & 1.414 \\end{bmatrix}.\n$$\n\n**Step 3 — row-wise softmax.** Work row $1$ explicitly. Exponentiate: $e^{0.707}\n\\approx 2.028$, $e^{0} = 1$, $e^{0.707} \\approx 2.028$, summing to $5.056$.\nDividing gives $2.028\u002F5.056 \\approx 0.401$ for the first and third keys and\n$1\u002F5.056 \\approx 0.198$ for the second. Doing the same to every row,\n\n$$\nA = \\softmax\\!\\parens{\\frac{S}{\\sqrt 2}}\n= \\begin{bmatrix} 0.401 & 0.198 & 0.401 \\\\ 0.198 & 0.401 & 0.401 \\\\ 0.248 & 0.248 & 0.503 \\end{bmatrix},\n$$\n\nand each row sums to $1$, as a distribution over the three keys must. Notice the\nthird token — whose query $(1,1)$ aligns with every key — spreads its weight most\nevenly but still favours itself (key $3$, weight $0.503$), because $(1,1)\\cdot(1,1)\n= 2$ is the largest score in that row.\n\n**Step 4 — blend the values, $O = AV$.** Row $1$ of the output is $0.401\\,v_1 +\n0.198\\,v_2 + 0.401\\,v_3 = 0.401(1,0) + 0.198(0,1) + 0.401(0.5,0.5) = (0.602,\n0.398)$. All three outputs:\n\n$$\nO = AV = \\begin{bmatrix} 0.602 & 0.398 \\\\ 0.398 & 0.602 \\\\ 0.500 & 0.500 \\end{bmatrix}\n\\;\\in\\; \\mathbb{R}^{3\\times 2}.\n$$\n\nEach output row is a point _inside_ the triangle spanned by the three value\nvectors — a convex combination, never an extrapolation beyond them. That is the\ngeometric signature of attention: the output can only ever be a weighted average\nof the values it was given.\n\n## Self-attention\n\nWhen the queries, keys, and values are all linear projections of the _same_\nsequence $X \\in \\mathbb{R}^{n \\times d}$, attention becomes **self-attention**:\neach position attends to every other position in its own input, including itself.\n\n> **Definition (Self-attention).** Self-attention is scaled dot-product attention\n> in which $Q = X W_Q$, $K = X W_K$, $V = X W_V$ all derive from one input\n> sequence $X$. The result re-expresses each position as a content-weighted\n> summary of the whole sequence.\n\nIts central object is the **attention matrix**\n\n$$\nA = \\softmax\\!\\parens{\\frac{QK^\\top}{\\sqrt{d_k}}} \\in \\mathbb{R}^{n \\times n},\n\\qquad\nA_{ij} = \\text{how much position } i \\text{ attends to position } j,\n$$\n\na row-stochastic matrix ($\\sum_j A_{ij} = 1$) that is a soft, content-based,\n**all-pairs routing table**: row $i$ says where token $i$ gathers its information.\nUnlike a convolution's fixed local stencil or a recurrence's strictly leftward\nchain, every entry is non-zero and _learned from content_: the word \"it\" can\nroute directly to the noun it refers to, however far back.\n\n$$\n% caption: A self-attention matrix. Each row is a token's query and darker cells carry more weight; here \"it\" attends strongly to \"animal\", resolving the reference in one hop.\n\\begin{tikzpicture}[font=\\scriptsize, x=0.95cm, y=0.62cm]\n  \\definecolor{acc}{HTML}{2348F2}\n  % column headers (top)\n  \\foreach \\c\u002F\\t in {0\u002Fthe, 1\u002Fanimal, 2\u002Fcrossed, 3\u002Fbecause, 4\u002Fit, 5\u002Fwas, 6\u002Ftired}\n    \\node[rotate=42, anchor=west, font=\\scriptsize] at (\\c+0.2, 7.15) {\\t};\n  % row headers (left)\n  \\foreach \\r\u002F\\t in {0\u002Fthe, 1\u002Fanimal, 2\u002Fcrossed, 3\u002Fbecause, 4\u002Fit, 5\u002Fwas, 6\u002Ftired}\n    \\node[anchor=east, font=\\footnotesize] at (-0.15, 6.4-\\r) {\\texttt{\\t}};\n  % each cell drawn explicitly: \\cell{col}{row}{shade 0-100} — light tint + crisp outline\n  \\def\\cell#1#2#3{\\draw[black, fill=acc!#3] (#1+0.07,6.15-#2) rectangle (#1+0.85,6.15-#2+0.78);}\n  % row 0  the\n  \\cell{0}{0}{62}\\cell{1}{0}{10}\\cell{2}{0}{6}\\cell{3}{0}{5}\\cell{4}{0}{5}\\cell{5}{0}{4}\\cell{6}{0}{4}\n  % row 1  animal\n  \\cell{0}{1}{12}\\cell{1}{1}{58}\\cell{2}{1}{10}\\cell{3}{1}{4}\\cell{4}{1}{8}\\cell{5}{1}{4}\\cell{6}{1}{4}\n  % row 2  crossed\n  \\cell{0}{2}{8}\\cell{1}{2}{32}\\cell{2}{2}{42}\\cell{3}{2}{4}\\cell{4}{2}{6}\\cell{5}{2}{4}\\cell{6}{2}{4}\n  % row 3  because\n  \\cell{0}{3}{5}\\cell{1}{3}{10}\\cell{2}{3}{12}\\cell{3}{3}{42}\\cell{4}{3}{16}\\cell{5}{3}{8}\\cell{6}{3}{5}\n  % row 4  it  -> attends to animal\n  \\cell{0}{4}{5}\\cell{1}{4}{72}\\cell{2}{4}{5}\\cell{3}{4}{5}\\cell{4}{4}{28}\\cell{5}{4}{5}\\cell{6}{4}{5}\n  % row 5  was\n  \\cell{0}{5}{5}\\cell{1}{5}{16}\\cell{2}{5}{6}\\cell{3}{5}{10}\\cell{4}{5}{20}\\cell{5}{5}{42}\\cell{6}{5}{10}\n  % row 6  tired\n  \\cell{0}{6}{5}\\cell{1}{6}{22}\\cell{2}{6}{5}\\cell{3}{6}{6}\\cell{4}{6}{16}\\cell{5}{6}{12}\\cell{6}{6}{56}\n\\end{tikzpicture}\n$$\n\n### Permutation equivariance\n\nSelf-attention has no notion of order. Permuting the rows of $X$ permutes the rows\nof the output identically and changes nothing else; the mechanism treats its\ninput as a _set_, not a sequence.\n\n> **Theorem (Permutation equivariance of self-attention).** Let $P$ be an $n\n> \\times n$ permutation matrix and let $\\SA(X) =\n> \\Attn(XW_Q, XW_K, XW_V)$. Then $\\SA(PX) = P\\,\n> \\SA(X)$.\n\n> **Proof.** Permuting the inputs sends $Q \\mapsto PQ$, $K \\mapsto PK$, $V \\mapsto\n> PV$, since each is $X$ times a fixed projection. The scores become $(PQ)(PK)^\\top\n> = P\\,(QK^\\top)\\,P^\\top$. The softmax acts row-wise and $P^\\top$ permutes columns\n> consistently with the row permutation $P$, so $\\softmax(P\\,S\\,\n> P^\\top) = P\\,\\softmax(S)\\,P^\\top$. Multiplying by the values,\n> $$\n> P\\,\\softmax(S)\\,P^\\top \\cdot PV = P\\,\\softmax(S)\\,V,\n> $$\n> because $P^\\top P = I$. The right side is $P\\,\\SA(X)$. $\\qed$\n\nEquivariance renders position invisible: \"dog bites\nman\" and \"man bites dog\" produce the same set of outputs, merely reordered. Order\nmust therefore be injected explicitly, the job of positional encoding below.\n\n## Multi-head attention\n\nA single attention function forces every position to mix all of its relationships\ninto one set of weights. **Multi-head attention** runs $h$ attention functions in\nparallel, each with its own projections, so different heads can specialize: one\ntracking syntactic agreement, another long-range coreference, another local\nadjacency.[^chollet-mha]\n\n> **Definition (Multi-head attention).** Project the inputs $h$ times with\n> independent matrices, attend in each subspace, then concatenate and project:\n> $$\n> \\head_i = \\Attn(X W_Q^{(i)}, X W_K^{(i)}, X W_V^{(i)}),\n> $$\n> $$\n> \\MHA(X) = \\Concat(\\head_1, \\dots, \\head_h)\\, W_O .\n> $$\n> Each head works in a $d_k = d \u002F h$ dimensional subspace, so the total compute\n> matches a single full-width head; $W_O \\in \\mathbb{R}^{d \\times d}$ recombines\n> the heads.\n\n$$\n% caption: Multi-head attention. The input is projected into $h$ subspaces, each running its own scaled dot-product attention; the outputs are concatenated and mixed by $W_O$.\n\\begin{tikzpicture}[>=stealth, font=\\footnotesize,\n  box\u002F.style={draw, minimum width=17mm, minimum height=9mm, align=center}]\n  \\definecolor{acc}{HTML}{2348F2}\n  \\node[box, draw=acc, text=acc] (in) at (0,0) {input X};\n  \\node[box] (h1) at (3.6,2.2)  {head 1};\n  \\node[box] (h2) at (3.6,0)    {head 2};\n  \\node[box] (h3) at (3.6,-2.2) {head h};\n  \\node[font=\\small] at (3.6,-1.1) {$\\vdots$};\n  \\node[box, draw=acc, text=acc] (cat) at (7.3,0) {concat};\n  \\node[box, draw=acc, text=acc] (wo) at (10.4,0) {$W_O$};\n  \\draw[->, acc, thick] (in) -- (h1);\n  \\draw[->, acc, thick] (in) -- (h2);\n  \\draw[->, acc, thick] (in) -- (h3);\n  \\draw[->, acc, thick] (h1) -- (cat);\n  \\draw[->, acc, thick] (h2) -- (cat);\n  \\draw[->, acc, thick] (h3) -- (cat);\n  \\draw[->, acc, thick] (cat) -- (wo);\n\\end{tikzpicture}\n$$\n\nThe shapes are worth tracing once.\nFix $d = d_{\\text{model}}$ and $h$ heads, and set the per-head width $d_k =\nd_v = d\u002Fh$ (so $h = 8$ heads of a $d = 512$ model each work in $64$ dimensions).\nFor head $i$ the projections are $W_Q^{(i)}, W_K^{(i)}, W_V^{(i)} \\in\n\\mathbb{R}^{d \\times d\u002Fh}$, so from an input $X \\in \\mathbb{R}^{n \\times d}$,\n\n$$\n\\underbrace{X W_Q^{(i)}}_{n \\times d\u002Fh} \\quad\n\\underbrace{X W_K^{(i)}}_{n \\times d\u002Fh} \\quad\n\\underbrace{X W_V^{(i)}}_{n \\times d\u002Fh}\n\\;\\Longrightarrow\\;\n\\underbrace{\\head_i}_{n \\times d\u002Fh},\n$$\n\neach head returns an $n \\times (d\u002Fh)$ block. Concatenating the $h$ blocks along\nthe width rebuilds a full $n \\times d$ matrix, which $W_O \\in \\mathbb{R}^{d \\times\nd}$ mixes back into the model width:\n\n$$\n\\underbrace{\\Concat(\\head_1, \\dots, \\head_h)}_{n \\times d}\n\\;\\cdot\\;\n\\underbrace{W_O}_{d \\times d}\n\\;=\\;\n\\underbrace{\\MHA(X)}_{n \\times d}.\n$$\n\nBecause $h \\cdot (d\u002Fh) = d$, the total projection and attention cost is the same\nas one full-width head; heads split the width rather than add to it. The\nexpressive gain is that the model holds $h$ distinct routing tables at once\ninstead of averaging them into one — one head can lock onto the previous token\nwhile another tracks long-range coreference, and $W_O$ learns how to combine\ntheir outputs.\n\n$$\n% caption: Multi-head width bookkeeping for $d = 512$, $h = 8$. Each head projects the $n \\times 512$ input down to $n \\times 64$, attends there, and the $8$ blocks concatenate back to $n \\times 512$ before $W_O$.\n\\begin{tikzpicture}[>=stealth, font=\\footnotesize,\n  wide\u002F.style={draw=acc, fill=acc!10, minimum width=34mm, minimum height=8mm, align=center},\n  slab\u002F.style={draw, minimum width=9mm, minimum height=8mm, align=center, fill=black!6}]\n  \\definecolor{acc}{HTML}{2348F2}\n  \\node[wide] (in) at (0,3.2) {\\texttt{X : n x 512}};\n  % 8 per-head slabs\n  \\foreach \\i in {0,...,7}\n    \\node[slab] (s\\i) at (\\i*1.05 - 3.7, 1.4) {\\texttt{64}};\n  \\node[font=\\footnotesize, text=black, anchor=west] at (4.0,1.4) {\\texttt{8 heads, n x 64}};\n  \\node[coordinate] (j) at (0,0.05) {};\n  \\node[wide] (cat) at (0,-1.0) {\\texttt{concat : n x 512}};\n  \\node[wide] (out) at (0,-2.5) {\\texttt{W\\_O out : n x 512}};\n  \\draw[->, acc, thick] (in) -- (s3.north);\n  \\draw[->, acc, thick] (in) -- (s4.north);\n  \\foreach \\i in {0,...,7} \\draw[black] (s\\i) -- (j);\n  \\draw[->, black] (j) -- (cat.north);\n  \\draw[->, acc, thick] (cat) -- (out);\n\\end{tikzpicture}\n$$\n\n## Positional encoding\n\nBecause self-attention is permutation-equivariant, the model cannot tell position\n$1$ from position $5$ unless we tell it. The fix is to _add_ a position-dependent\nvector $\\PE(p)$ to each input embedding before the first layer, so\nthat order rides along in the representation itself.[^chollet-pos]\n\n> **Definition (Sinusoidal positional encoding).** For position $p$ and embedding\n> dimension index $i$, with model width $d$,\n> $$\n> \\PE(p, 2i) = \\sin\\!\\parens{\\frac{p}{10000^{\\,2i\u002Fd}}},\n> \\qquad\n> \\PE(p, 2i+1) = \\cos\\!\\parens{\\frac{p}{10000^{\\,2i\u002Fd}}}.\n> $$\n> Each dimension is a sinusoid; the wavelength grows geometrically from $2\\pi$ to\n> $10000 \\cdot 2\\pi$ across dimensions, so the vector encodes position at many\n> resolutions at once.\n\nThe geometric spread of frequencies is the key property: low-index dimensions oscillate\nquickly to distinguish neighbors, high-index dimensions oscillate slowly to\ndistinguish distant regions. The encoding also admits _relative_ shifts, since\n$\\PE(p+\\Delta)$ is a fixed linear function of $\\PE(p)$,\nso the model can learn to attend by offset, not just absolute index.\n\n$$\n% caption: Sinusoidal positional encodings. Each curve is one embedding dimension over\n% position $p$: low dimensions (blue) oscillate fast, high ones (red) slowly.\n\\begin{tikzpicture}[>=stealth, font=\\footnotesize]\n  \\definecolor{acc}{HTML}{2348F2}\n  \\definecolor{red}{HTML}{C0392B}\n  \\draw[->, thick] (-0.2,0) -- (7.6,0) node[right, font=\\scriptsize] {position p};\n  \\draw[->, thick] (0,-1.4) -- (0,1.5) node[above, font=\\scriptsize] {value};\n  \\draw[black] (0,0) -- (7.4,0);\n  % high frequency (blue)\n  \\draw[acc, very thick] plot[domain=0:7.3, samples=120] (\\x, {1.1*sin(2.6*\\x r)});\n  % medium frequency (black)\n  \\draw[black, very thick] plot[domain=0:7.3, samples=120] (\\x, {1.1*sin(1.1*\\x r)});\n  % low frequency (red)\n  \\draw[red, very thick] plot[domain=0:7.3, samples=120] (\\x, {1.1*sin(0.45*\\x r)});\n  \\node[acc, font=\\scriptsize, anchor=west] at (6.2,1.25) {fast};\n  \\node[red, font=\\scriptsize, anchor=west] at (6.2,-0.55) {slow};\n\\end{tikzpicture}\n$$\n\nLearned positional embeddings — a trainable lookup table indexed by position —\nare the common alternative; they fit the data more tightly but, unlike the\nsinusoids, do not extrapolate to sequences longer than those seen in training.\n\n| Scheme | Form | Generalizes past training length? | Cost |\n| --- | --- | --- | --- |\n| Sinusoidal | fixed $\\sin\u002F\\cos$ of position | yes (deterministic for any $p$) | none |\n| Learned | trainable table $E_{\\text{pos}} \\in \\mathbb{R}^{n_{\\max} \\times d}$ | no (undefined for $p > n_{\\max}$) | $n_{\\max} \\cdot d$ params |\n| Relative | bias on scores by offset $i - j$ | yes | per-head bias |\n\n## The Transformer block\n\nSelf-attention mixes information _across_ positions; it does no nonlinear\nprocessing _within_ a position. The **Transformer block** pairs the two: a\nmulti-head self-attention sublayer followed by a position-wise feed-forward\nnetwork, each sublayer wrapped in a residual connection and [layer\nnormalization](\u002Fdeep-learning\u002Fregularization\u002Fnormalization).[^chollet-block]\n\n> **Definition (Position-wise feed-forward network).** A two-layer MLP applied\n> independently and identically to every position:\n> $$\n> \\FFN(x) = \\ReLU(x W_1 + b_1)\\,W_2 + b_2,\n> $$\n> with an inner width typically $4d$. It is the per-token nonlinearity that\n> attention lacks; sharing weights across positions makes it a $1 \\times 1$\n> convolution in disguise.\n\nEach sublayer is wrapped identically, $\\out = \\LayerNorm\n(x + \\Sublayer(x))$. The residual skip lets gradients reach early\nlayers undiminished — the same mechanism that powers [residual\nnetworks](\u002Fdeep-learning\u002Farchitectures\u002Fcnn-architectures) — and LayerNorm holds\neach position's activations at a stable scale across depth.\n\n$$\n% caption: A Transformer encoder block: multi-head self-attention then a position-wise FFN, each wrapped in an add-and-normalize. The two curved skips are the residual connections.\n\\begin{tikzpicture}[>=stealth, font=\\footnotesize,\n  box\u002F.style={draw, minimum width=42mm, minimum height=9mm, align=center},\n  add\u002F.style={draw, minimum width=42mm, minimum height=8mm, align=center, fill=black!8}]\n  \\definecolor{acc}{HTML}{2348F2}\n  \\node (in)  at (0,0)    {input};\n  \\node[box, draw=acc, text=acc, thick] (mha) at (0,1.3)  {\\texttt{multi-head self-attention}};\n  \\node[add] (an1) at (0,2.7)  {\\texttt{add \\& norm}};\n  \\node[box, draw=acc, text=acc, thick] (ffn) at (0,4.1)  {\\texttt{feed-forward}};\n  \\node[add] (an2) at (0,5.5)  {\\texttt{add \\& norm}};\n  \\node (out) at (0,6.7)  {output};\n  \\draw[->, thick] (in) -- (mha);\n  \\draw[->, thick] (mha) -- (an1);\n  \\draw[->, thick] (an1) -- (ffn);\n  \\draw[->, thick] (ffn) -- (an2);\n  \\draw[->, thick] (an2) -- (out);\n  % residual skips on the right\n  \\draw[->, acc, thick] (2.1,0.2) .. controls (3.4,0.9) and (3.4,2.0) .. (2.1,2.7)\n    node[pos=0.5, right, font=\\scriptsize, text=acc] {skip};\n  \\draw[->, acc, thick] (2.1,3.0) .. controls (3.4,3.7) and (3.4,4.8) .. (2.1,5.5)\n    node[pos=0.5, right, font=\\scriptsize, text=acc] {skip};\n\\end{tikzpicture}\n$$\n\n### Encoder and decoder\n\nThe original Transformer is an **encoder-decoder**. The encoder stacks identical\nblocks with unrestricted self-attention — every position sees every other. The\ndecoder adds two twists that make it a valid autoregressive generator.\n\n| Component | Self-attention | Cross-attention | Sees the future? |\n| --- | --- | --- | --- |\n| Encoder block | full (all-pairs) | — | yes (whole input visible) |\n| Decoder block | _masked_ | attends to encoder output | no (causal mask) |\n\n- **Masked self-attention.** To generate token $t$ using only tokens $\u003C t$, the\n  decoder sets $S_{ij} = -\\infty$ for $j > i$ before the softmax (the masking\n  branch of the algorithm above), forcing $A_{ij} = 0$ on the future. Generation\n  stays causal.\n- **Cross-attention.** A second attention sublayer takes its queries from the\n  decoder but its keys and values from the encoder's output, letting each\n  generated token read the entire source sequence — the channel through which a\n  translation model reads the sentence it is translating.\n\nCross-attention is the setting where the $n \\ne m$ shapes matter. In an\nencoder of length $m$ and a decoder generating length $n$, the decoder supplies\n$Q \\in \\mathbb{R}^{n \\times d_k}$ while the encoder supplies $K, V \\in\n\\mathbb{R}^{m \\times d_k}$, so the score matrix is a rectangular $n \\times m$\nlookup — every output position against every source position. Self-attention is\nthe special case $Q = K = V$ from one sequence, giving the square $n \\times n$\nmatrix; cross-attention keeps queries and keys from _different_ sequences.\n\n### The causal mask\n\nThe mask is a single additive matrix $M \\in \\mathbb{R}^{n \\times n}$ applied to\nthe scaled scores before the softmax: $M_{ij} = 0$ for $j \\le i$ and $M_{ij} =\n-\\infty$ for $j > i$. Adding $-\\infty$ to a logit sends $e^{-\\infty} = 0$, so the\nsoftmax assigns those future positions weight exactly $0$ — the reason for the\n$-\\infty$ rather than merely a large negative number. For four tokens,\n\n$$\nM = \\begin{bmatrix}\n0 & -\\infty & -\\infty & -\\infty \\\\\n0 & 0 & -\\infty & -\\infty \\\\\n0 & 0 & 0 & -\\infty \\\\\n0 & 0 & 0 & 0\n\\end{bmatrix},\n\\qquad\nA = \\softmax\\!\\parens{\\frac{QK^\\top}{\\sqrt{d_k}}} + M \\text{ pattern}\n= \\begin{bmatrix}\n\\bullet & 0 & 0 & 0 \\\\\n\\bullet & \\bullet & 0 & 0 \\\\\n\\bullet & \\bullet & \\bullet & 0 \\\\\n\\bullet & \\bullet & \\bullet & \\bullet\n\\end{bmatrix},\n$$\n\nleaving a lower-triangular weight matrix: token $i$ attends only to tokens $1\n\\dots i$. The mask costs nothing at training time (one addition), yet it is what\nlets the decoder train on a whole target sequence in parallel while still\npredicting each token from its prefix alone.\n\n$$\n% caption: The causal mask. Below and on the diagonal (blue) a token may attend; above it (crossed) the score is set to $-\\infty$ so the softmax weight is $0$. Row $i$ sees only tokens $1 \\dots i$.\n\\begin{tikzpicture}[font=\\scriptsize, x=0.9cm, y=0.9cm]\n  \\definecolor{acc}{HTML}{2348F2}\n  \\foreach \\c\u002F\\lab in {0\u002F1, 1\u002F2, 2\u002F3, 3\u002F4, 4\u002F5}\n    \\node[font=\\scriptsize, text=black] at (\\c+0.5, 5.2) {$k_{\\lab}$};\n  \\foreach \\r\u002F\\lab in {0\u002F1, 1\u002F2, 2\u002F3, 3\u002F4, 4\u002F5}\n    \\node[font=\\scriptsize, text=black, anchor=east] at (-0.05, 4.5-\\r) {$q_{\\lab}$};\n  \\foreach \\r in {0,...,4} \\foreach \\c in {0,...,4} {\n    \\ifnum\\c>\\r\n      \\draw[black] (\\c,4-\\r) rectangle (\\c+1,5-\\r);\n      \\draw[black] (\\c+0.18,4.18-\\r) -- (\\c+0.82,4.82-\\r);\n      \\draw[black] (\\c+0.18,4.82-\\r) -- (\\c+0.82,4.18-\\r);\n    \\else\n      \\draw[draw=acc, fill=acc!22] (\\c,4-\\r) rectangle (\\c+1,5-\\r);\n    \\fi\n  }\n\\end{tikzpicture}\n$$\n\n## Recurrence vs. convolution vs. self-attention\n\nWhy did attention displace both recurrence and convolution for sequences? Compare\nthe three on the costs that matter: how far information must travel to connect two\npositions (**maximum path length**, which bounds how easily long-range gradients\nflow), how much of the work is inherently **sequential**, and the **per-layer\ncompute**.\n\n| Layer type | Per-layer complexity | Sequential ops | Max path length |\n| --- | --- | --- | --- |\n| Recurrent | $O(n\\, d^2)$ | $O(n)$ | $O(n)$ |\n| Convolutional (kernel $w$) | $O(w\\, n\\, d^2)$ | $O(1)$ | $O(\\log_w n)$ |\n| Self-attention | $O(n^2 d)$ | $O(1)$ | $O(1)$ |\n\nTwo columns decide it. Self-attention's maximum path length is $O(1)$: _any_ two\npositions interact directly, so the gradient between them never has to pass through a\nlong chain. Its sequential depth is $O(1)$, so the whole layer parallelizes\nacross positions on a GPU, unlike the strictly stepwise recurrence. The price is\nthe $O(n^2 d)$ term: the $n \\times n$ attention matrix is quadratic in sequence\nlength, cheap when $n \u003C d$ (the usual regime) but the dominant cost for long\nsequences, the bottleneck that later efficient-attention variants attack.\n\n$$\n% caption: Connectivity, not just cost. Self-attention (left) links every token pair directly at path length $1$; a recurrence (right) leaves distant tokens $O(n)$ hops apart.\n\\begin{tikzpicture}[>=stealth, font=\\footnotesize,\n  tok\u002F.style={circle, draw, minimum size=7mm, inner sep=0pt}]\n  \\definecolor{acc}{HTML}{2348F2}\n  % --- self-attention: all-pairs ---\n  \\foreach \\i in {1,2,3,4} \\node[tok, draw=acc, text=acc] (a\\i) at (\\i*1.25,0) {\\i};\n  \\draw[acc] (a1) to[bend left=38] (a2);\n  \\draw[acc] (a1) to[bend left=42] (a3);\n  \\draw[acc] (a1) to[bend left=46] (a4);\n  \\draw[acc] (a2) to[bend left=38] (a3);\n  \\draw[acc] (a2) to[bend left=42] (a4);\n  \\draw[acc] (a3) to[bend left=38] (a4);\n  \\node[align=center, font=\\footnotesize] at (3.1,-1.1) {\\texttt{self-attention}\\\\(all-pairs, path $1$)};\n  % --- recurrence: chain ---\n  \\begin{scope}[xshift=7.4cm]\n    \\foreach \\i in {1,2,3,4} \\node[tok] (b\\i) at (\\i*1.25,0) {\\i};\n    \\draw[->, black, thick] (b1) -- (b2);\n    \\draw[->, black, thick] (b2) -- (b3);\n    \\draw[->, black, thick] (b3) -- (b4);\n    \\node[align=center, font=\\scriptsize] at (3.1,-1.1) {recurrence\\\\(chain, path $n$)};\n  \\end{scope}\n\\end{tikzpicture}\n$$\n\n## Where attention came from\n\nAttention postdates Goodfellow's 2016 text, so the primary sources carry the\ndetail Goodfellow cannot.\n\n- **Where attention came from.** The dot-product form here is the streamlined\n  version. The idea started as **additive attention** (Bahdanau et al., ICLR 2015),\n  which scored a query against each key with a small MLP, $\\score(q,k)\n  = v^\\top \\tanh(W_q q + W_k k)$, to let a translation decoder read every encoder\n  state instead of one fixed vector. Vaswani et al., \"Attention Is All You Need\"\n  (NeurIPS 2017), replaced the MLP with the cheap scaled dot product, added the\n  $1\u002F\\sqrt{d_k}$ correction derived above, and removed recurrence entirely.\n- **Reducing the $O(n^2)$ cost.** The quadratic score matrix is the target of a\n  whole line of work. Two directions matter. _Sparse and linear approximations_ —\n  Longformer (Beltagy et al., 2020) with sliding-window attention, and the\n  Performer\u002Flinear-attention family — reduce the asymptotic cost to $O(n\\log n)$ or\n  $O(n)$ by never forming the full matrix. _Exact but IO-aware kernels_ —\n  FlashAttention (Dao et al., NeurIPS 2022) — keep the exact softmax but tile the\n  computation so the $n\\times n$ matrix never leaves fast on-chip memory, cutting the\n  memory footprint from $O(n^2)$ to $O(n)$ with no change to the result. The second\n  approach is now the default in production Transformers.\n- **Relative and rotary position.** The additive sinusoidal encoding is only the\n  first of several schemes. Shaw et al. (2018) added learned relative-position biases\n  inside the scores; RoPE (Su et al., 2021) rotates the query and key vectors by an\n  angle proportional to position, so a dot product depends only on the _relative_\n  offset — the standard choice in modern decoder-only LLMs.\n\nThe full architecture built from this mechanism — the stacked encoder–decoder,\ncausal masking, and the parameter\u002Fcompute accounting — is the subject of\n[the next lesson](\u002Fdeep-learning\u002Farchitectures\u002Fthe-transformer-architecture).\n\n## Takeaways\n\n- **Attention** is a differentiable soft lookup: a query scores a set of keys, a\n  softmax turns the scores into weights, and the output is the weighted blend of\n  the values, $\\Attn(Q,K,V) = \\softmax(QK^\\top\u002F\\sqrt{d_k})\\,V$.\n- The **$\\sqrt{d_k}$** scaling cancels the dot product's variance growth\n  ($\\Var(q \\cdot k) = d_k$), keeping the softmax out of its\n  saturated, zero-gradient regime.\n- **Self-attention** ($Q,K,V$ from one sequence) yields an all-pairs routing\n  matrix $A$; being permutation-equivariant, it needs **positional encoding** —\n  sinusoidal or learned — to know order.\n- **Multi-head** attention runs $h$ parallel attentions in $d\u002Fh$-dimensional\n  subspaces, capturing several relationship patterns at once for the cost of one.\n- The **Transformer block** = multi-head self-attention + position-wise FFN, each\n  in a residual + LayerNorm wrapper; decoders add a causal mask and\n  cross-attention.\n- Against recurrence and convolution, self-attention wins on **path length**\n  ($O(1)$) and **parallelism** ($O(1)$ sequential ops), paying an $O(n^2 d)$\n  per-layer cost — the trade that made it the dominant sequence architecture.\n\n[^gf-attention]: **Goodfellow**, _Deep Learning_, §10.11 — Explicit Memory & attention over memory cells: content-based addressing as a differentiable soft lookup (the Transformer itself postdates the 2016 text).\n[^chollet-attn]: **Chollet**, _Deep Learning with Python_, Ch. 11 — Deep Learning for Text: scaled dot-product attention, the $1\u002F\\sqrt{d_k}$ correction, and the `softmax(QK^T\u002F\\sqrt{d_k})V` core of the Transformer.\n[^chollet-mha]: **Chollet**, _Deep Learning with Python_, Ch. 11 — Multi-Head Attention: $h$ parallel attention functions in $d\u002Fh$-dimensional subspaces, concatenated and projected by $W_O$.\n[^chollet-pos]: **Chollet**, _Deep Learning with Python_, Ch. 11 — Positional encoding: injecting order into a permutation-equivariant model, sinusoidal versus learned embeddings.\n[^chollet-block]: **Chollet**, _Deep Learning with Python_, Ch. 11 — The Transformer encoder block: multi-head self-attention plus a position-wise feed-forward network, each in a residual + normalization wrapper.\n",{"text":23409,"minutes":23410,"time":23411,"words":23412},"14 min read",13.01,780600,2602,{"title":5,"description":23398},[23415,23417],{"book":23111,"ref":23416},"§10.11 — Explicit Memory & Attention (the Transformer postdates the 2016 text)",{"book":23128,"ref":23418},"Ch. 11 — Deep Learning for Text; The Transformer Architecture","available","06.deep-learning\u002F05.architectures\u002F05.attention-and-transformers","Attention replaces fixed wiring with content-based routing: every position reads from every other through a soft, learned dot-product lookup. We derive scaled dot-product attention and its $\\sqrt{d_k}$ correction, build it into multi-head self-attention, inject order with positional encodings, and stack the whole thing into the Transformer block that displaced recurrence and convolution alike.\n",[23403],"Xu7jFXaUJfUoNaa3Pv-5O9hZh1qSJFHBseG3z-8Jzp4",{"\u002Falgorithms\u002Ffoundations\u002Fwhat-is-an-algorithm":23425,"\u002Falgorithms\u002Ffoundations\u002Fproof-techniques":23426,"\u002Falgorithms\u002Ffoundations\u002Fasymptotic-analysis":23427,"\u002Falgorithms\u002Ffoundations\u002Fgrowth-rates-and-loop-analysis":23428,"\u002Falgorithms\u002Ffoundations\u002Frecurrences":23429,"\u002Falgorithms\u002Ffoundations\u002Famortized-analysis":23430,"\u002Falgorithms\u002Fdivide-and-conquer\u002Fmergesort":23431,"\u002Falgorithms\u002Fdivide-and-conquer\u002Fquicksort":23432,"\u002Falgorithms\u002Fdivide-and-conquer\u002Fselection":23433,"\u002Falgorithms\u002Fdivide-and-conquer\u002Ffast-multiplication":23434,"\u002Falgorithms\u002Fsorting\u002Fheaps-and-heapsort":23435,"\u002Falgorithms\u002Fsorting\u002Fsorting-lower-bounds":23436,"\u002Falgorithms\u002Fsorting\u002Flinear-time-sorting":23437,"\u002Falgorithms\u002Fsorting\u002Fexternal-sorting":23438,"\u002Falgorithms\u002Fdata-structures\u002Felementary-structures":23439,"\u002Falgorithms\u002Fdata-structures\u002Fhash-tables":23440,"\u002Falgorithms\u002Fdata-structures\u002Fbinary-search-trees":23441,"\u002Falgorithms\u002Fdata-structures\u002Favl-trees":23442,"\u002Falgorithms\u002Fdata-structures\u002Fbalanced-trees":23443,"\u002Falgorithms\u002Fdata-structures\u002Funion-find":23444,"\u002Falgorithms\u002Fdata-structures\u002Ffenwick-and-segment-trees":23445,"\u002Falgorithms\u002Fdata-structures\u002Fspatial-data-structures":23446,"\u002Falgorithms\u002Fdata-structures\u002Fskip-lists-and-probabilistic-structures":23447,"\u002Falgorithms\u002Fdata-structures\u002Fb-trees":23448,"\u002Falgorithms\u002Fdata-structures\u002Fdata-stream-algorithms":23449,"\u002Falgorithms\u002Fdata-structures\u002Fstreaming-sketches":23450,"\u002Falgorithms\u002Fsequences\u002Ftwo-pointers-and-windows":23451,"\u002Falgorithms\u002Fsequences\u002Fprefix-sums":23452,"\u002Falgorithms\u002Fsequences\u002Fmonotonic-stacks":23453,"\u002Falgorithms\u002Fsequences\u002Fbinary-search-on-the-answer":23454,"\u002Falgorithms\u002Fsequences\u002Fstring-matching":23455,"\u002Falgorithms\u002Fsequences\u002Fkmp-and-z-function":23456,"\u002Falgorithms\u002Fsequences\u002Ftries":23457,"\u002Falgorithms\u002Fsequences\u002Fsuffix-arrays-and-aho-corasick":23458,"\u002Falgorithms\u002Fgraphs\u002Frepresentations-and-traversal":23459,"\u002Falgorithms\u002Fgraphs\u002Fdepth-first-search":23460,"\u002Falgorithms\u002Fgraphs\u002Ftopological-sort-and-scc":23461,"\u002Falgorithms\u002Fgraphs\u002Fminimum-spanning-trees":23462,"\u002Falgorithms\u002Fgraphs\u002Fkruskal-and-prim":23463,"\u002Falgorithms\u002Fgraphs\u002Fshortest-paths":23464,"\u002Falgorithms\u002Fgraphs\u002Fall-pairs-and-negative-weights":23465,"\u002Falgorithms\u002Fgraphs\u002Fnetwork-flow":23466,"\u002Falgorithms\u002Fgraphs\u002Fmax-flow-min-cut":23467,"\u002Falgorithms\u002Fgraphs\u002Fbridges-and-articulation-points":23468,"\u002Falgorithms\u002Fgraphs\u002Flowest-common-ancestor":23469,"\u002Falgorithms\u002Fgraphs\u002Ftwo-sat":23470,"\u002Falgorithms\u002Fgraphs\u002Feulerian-tours":23471,"\u002Falgorithms\u002Fgraphs\u002Fbipartite-matching":23472,"\u002Falgorithms\u002Fgreedy\u002Fthe-greedy-method":23473,"\u002Falgorithms\u002Fgreedy\u002Fscheduling-and-intervals":23474,"\u002Falgorithms\u002Fgreedy\u002Fhuffman-codes":23475,"\u002Falgorithms\u002Fgreedy\u002Fmatroids":23476,"\u002Falgorithms\u002Fgreedy\u002Fstable-matching":23477,"\u002Falgorithms\u002Fdynamic-programming\u002Fprinciples":23478,"\u002Falgorithms\u002Fdynamic-programming\u002Fsequence-dp":23479,"\u002Falgorithms\u002Fdynamic-programming\u002Flongest-increasing-subsequence":23480,"\u002Falgorithms\u002Fdynamic-programming\u002Fknapsack":23481,"\u002Falgorithms\u002Fdynamic-programming\u002Fcoin-change-and-unbounded":23482,"\u002Falgorithms\u002Fdynamic-programming\u002Finterval-dp":23483,"\u002Falgorithms\u002Fdynamic-programming\u002Ftree-dp":23484,"\u002Falgorithms\u002Fdynamic-programming\u002Fbitmask-dp":23485,"\u002Falgorithms\u002Fdynamic-programming\u002Fdp-optimizations":23486,"\u002Falgorithms\u002Fdynamic-programming\u002Fdp-on-graphs":23487,"\u002Falgorithms\u002Fdynamic-programming\u002Fdigit-and-probability-dp":23488,"\u002Falgorithms\u002Fbacktracking\u002Fbacktracking-fundamentals":23489,"\u002Falgorithms\u002Fbacktracking\u002Fconstraint-search":23490,"\u002Falgorithms\u002Fbacktracking\u002Fbranch-and-bound":23491,"\u002Falgorithms\u002Fbacktracking\u002Fgraph-backtracking":23492,"\u002Falgorithms\u002Fmathematical-algorithms\u002Fnumber-theory-basics":23493,"\u002Falgorithms\u002Fmathematical-algorithms\u002Fmodular-exponentiation-and-primality":23494,"\u002Falgorithms\u002Fmathematical-algorithms\u002Fsieve-and-factorization":23495,"\u002Falgorithms\u002Fmathematical-algorithms\u002Fcombinatorics":23496,"\u002Falgorithms\u002Fmathematical-algorithms\u002Fmatrix-exponentiation":23497,"\u002Falgorithms\u002Fmathematical-algorithms\u002Ffast-fourier-transform":23498,"\u002Falgorithms\u002Fmathematical-algorithms\u002Fgradient-descent":23499,"\u002Falgorithms\u002Fcomputational-geometry\u002Fgeometric-primitives":23500,"\u002Falgorithms\u002Fcomputational-geometry\u002Fconvex-hull":23501,"\u002Falgorithms\u002Fcomputational-geometry\u002Fsweep-line":23502,"\u002Falgorithms\u002Fcomputational-geometry\u002Fpolygons-and-proximity":23503,"\u002Falgorithms\u002Fintractability\u002Fp-np-reductions":23504,"\u002Falgorithms\u002Fintractability\u002Fnp-completeness":23505,"\u002Falgorithms\u002Fintractability\u002Fcoping-with-hardness":23506,"\u002Falgorithms\u002Fintractability\u002Fapproximation-algorithms":23507,"\u002Falgorithms":23508,"\u002Fcalculus\u002Flimits-and-continuity\u002Ffunctions-and-models":23509,"\u002Fcalculus\u002Flimits-and-continuity\u002Fthe-limit-of-a-function":23510,"\u002Fcalculus\u002Flimits-and-continuity\u002Flimit-laws-and-the-precise-definition":23511,"\u002Fcalculus\u002Flimits-and-continuity\u002Fcontinuity":23512,"\u002Fcalculus\u002Fderivatives\u002Fthe-derivative-and-rates-of-change":23513,"\u002Fcalculus\u002Fderivatives\u002Fdifferentiation-rules-and-the-chain-rule":23514,"\u002Fcalculus\u002Fderivatives\u002Fimplicit-differentiation-and-related-rates":23515,"\u002Fcalculus\u002Fderivatives\u002Flinear-approximations-and-differentials":23516,"\u002Fcalculus\u002Fapplications-of-derivatives\u002Fextrema-and-the-mean-value-theorem":23517,"\u002Fcalculus\u002Fapplications-of-derivatives\u002Fhow-derivatives-shape-a-graph":23518,"\u002Fcalculus\u002Fapplications-of-derivatives\u002Fcurve-sketching-and-optimization":23519,"\u002Fcalculus\u002Fapplications-of-derivatives\u002Fnewtons-method-and-antiderivatives":23520,"\u002Fcalculus\u002Fintegrals\u002Farea-and-the-definite-integral":23521,"\u002Fcalculus\u002Fintegrals\u002Fthe-fundamental-theorem-of-calculus":23522,"\u002Fcalculus\u002Fintegrals\u002Fthe-substitution-rule":23523,"\u002Fcalculus\u002Fapplications-of-integration\u002Fareas-and-volumes":23524,"\u002Fcalculus\u002Fapplications-of-integration\u002Fwork-average-value-and-arc-length":23525,"\u002Fcalculus\u002Fapplications-of-integration\u002Fphysics-economics-and-probability":23526,"\u002Fcalculus\u002Fexponential-logarithmic-and-inverse-functions\u002Finverse-functions-logarithms-and-exponentials":23527,"\u002Fcalculus\u002Fexponential-logarithmic-and-inverse-functions\u002Fgrowth-decay-inverse-trig-and-hyperbolic-functions":23528,"\u002Fcalculus\u002Fexponential-logarithmic-and-inverse-functions\u002Flhospitals-rule":23529,"\u002Fcalculus\u002Ftechniques-of-integration\u002Fintegration-by-parts":23530,"\u002Fcalculus\u002Ftechniques-of-integration\u002Ftrigonometric-integrals-and-substitution":23531,"\u002Fcalculus\u002Ftechniques-of-integration\u002Fpartial-fractions-and-integration-strategy":23532,"\u002Fcalculus\u002Ftechniques-of-integration\u002Fapproximate-and-improper-integrals":23533,"\u002Fcalculus\u002Fparametric-and-polar\u002Fparametric-curves-and-their-calculus":23534,"\u002Fcalculus\u002Fparametric-and-polar\u002Fpolar-coordinates":23535,"\u002Fcalculus\u002Fparametric-and-polar\u002Fconic-sections":23536,"\u002Fcalculus\u002Fsequences-and-series\u002Fsequences":23537,"\u002Fcalculus\u002Fsequences-and-series\u002Fseries-and-the-integral-test":23538,"\u002Fcalculus\u002Fsequences-and-series\u002Fthe-convergence-tests":23539,"\u002Fcalculus\u002Fsequences-and-series\u002Fpower-series":23540,"\u002Fcalculus\u002Fsequences-and-series\u002Ftaylor-and-maclaurin-series":23541,"\u002Fcalculus\u002Fvectors-and-space-curves\u002Fvectors-and-the-dot-product":23542,"\u002Fcalculus\u002Fvectors-and-space-curves\u002Fthe-cross-product-lines-and-planes":23523,"\u002Fcalculus\u002Fvectors-and-space-curves\u002Fcylinders-and-quadric-surfaces":23543,"\u002Fcalculus\u002Fvectors-and-space-curves\u002Fvector-functions-and-space-curves":23544,"\u002Fcalculus\u002Fvectors-and-space-curves\u002Farc-length-curvature-and-motion":23545,"\u002Fcalculus\u002Fpartial-derivatives\u002Ffunctions-of-several-variables":23513,"\u002Fcalculus\u002Fpartial-derivatives\u002Fpartial-derivatives":23546,"\u002Fcalculus\u002Fpartial-derivatives\u002Ftangent-planes-and-the-chain-rule":23547,"\u002Fcalculus\u002Fpartial-derivatives\u002Fdirectional-derivatives-and-the-gradient":23548,"\u002Fcalculus\u002Fpartial-derivatives\u002Foptimization-and-lagrange-multipliers":23549,"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fdouble-integrals":23550,"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Ftriple-integrals-and-coordinate-systems":23551,"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fvector-fields-and-line-integrals":23552,"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fgreens-theorem-curl-and-divergence":23553,"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fsurface-integrals":23554,"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fstokes-and-the-divergence-theorem":23555,"\u002Fcalculus":23556,"\u002Fmechanics\u002Ffoundations\u002Fmeasurement-and-dimensions":23557,"\u002Fmechanics\u002Ffoundations\u002Fvector-algebra":23558,"\u002Fmechanics\u002Fkinematics\u002Fone-dimensional-motion":23559,"\u002Fmechanics\u002Fkinematics\u002Fmotion-graphs":23560,"\u002Fmechanics\u002Fkinematics\u002Fprojectile-motion":23561,"\u002Fmechanics\u002Fkinematics\u002Frelative-motion":23562,"\u002Fmechanics\u002Fkinematics\u002Fcircular-motion":23563,"\u002Fmechanics\u002Fdynamics\u002Fnewtons-laws":23564,"\u002Fmechanics\u002Fdynamics\u002Ffree-body-diagrams":23565,"\u002Fmechanics\u002Fdynamics\u002Ffriction-and-curved-motion":23566,"\u002Fmechanics\u002Fdynamics\u002Fnumerical-dynamics":23567,"\u002Fmechanics\u002Fdynamics\u002Fcenter-of-mass-systems":23568,"\u002Fmechanics\u002Fenergy\u002Fwork-and-kinetic-energy":23569,"\u002Fmechanics\u002Fenergy\u002Fpotential-energy":23570,"\u002Fmechanics\u002Fenergy\u002Fmultiparticle-work":23571,"\u002Fmechanics\u002Fenergy\u002Fmass-energy-and-binding":23572,"\u002Fmechanics\u002Fenergy\u002Fphotons-and-quantization":23573,"\u002Fmechanics\u002Fmomentum\u002Fmomentum-and-collisions":23574,"\u002Fmechanics\u002Fmomentum\u002Fcenter-of-mass-collisions":23575,"\u002Fmechanics\u002Fmomentum\u002Frocket-propulsion":23576,"\u002Fmechanics\u002Frotation\u002Frotational-inertia":23577,"\u002Fmechanics\u002Frotation\u002Frotational-dynamics":23578,"\u002Fmechanics\u002Frotation\u002Frolling-motion":23579,"\u002Fmechanics\u002Frotation\u002Fangular-momentum":23580,"\u002Fmechanics\u002Frotation\u002Frolling-resistance":23581,"\u002Fmechanics\u002Frotation\u002Fgyroscopic-precession":23582,"\u002Fmechanics\u002Fgravity-and-matter\u002Fkeplerian-orbits":23583,"\u002Fmechanics\u002Fgravity-and-matter\u002Fgravitational-fields":23584,"\u002Fmechanics\u002Fgravity-and-matter\u002Fstatic-equilibrium":23585,"\u002Fmechanics\u002Fgravity-and-matter\u002Ffluid-statics":23586,"\u002Fmechanics\u002Fgravity-and-matter\u002Ffluid-flow":23587,"\u002Fmechanics\u002Fgravity-and-matter\u002Forbital-motion":23588,"\u002Fmechanics\u002Fgravity-and-matter\u002Fstress-and-elasticity":23589,"\u002Fmechanics\u002Foscillations-waves\u002Fdamped-oscillators":23590,"\u002Fmechanics\u002Foscillations-waves\u002Ftravelling-waves":23591,"\u002Fmechanics\u002Foscillations-waves\u002Fwave-superposition":23592,"\u002Fmechanics\u002Foscillations-waves\u002Fstanding-waves":23593,"\u002Fmechanics\u002Foscillations-waves\u002Fsound-waves":23594,"\u002Fmechanics\u002Foscillations-waves\u002Fdoppler-effect":23595,"\u002Fmechanics\u002Foscillations-waves\u002Fwave-packets":23596,"\u002Fmechanics\u002Foscillations-waves\u002Fbeats-and-coupling":23597,"\u002Fmechanics\u002Foscillations-waves\u002Fsimple-harmonic-motion":23598,"\u002Fmechanics\u002Foscillations-waves\u002Fpendulum-motion":23599,"\u002Fmechanics\u002Foscillations-waves\u002Fdriven-oscillators":23600,"\u002Fmechanics\u002Foscillations-waves\u002Fwave-boundaries":23601,"\u002Fmechanics\u002Fthermodynamics\u002Fkinetic-theory-of-ideal-gases":23602,"\u002Fmechanics\u002Fthermodynamics\u002Ffirst-law-of-thermodynamics":23603,"\u002Fmechanics\u002Fthermodynamics\u002Fentropy-and-the-second-law":23604,"\u002Fmechanics\u002Fthermodynamics\u002Fthermal-processes":23605,"\u002Fmechanics\u002Fthermodynamics\u002Fphase-changes":23606,"\u002Fmechanics\u002Fthermodynamics\u002Fthermal-machines":23607,"\u002Fmechanics":23608,"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Fcharge-and-conductors":23609,"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Fcoulombs-law":23610,"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Felectric-field-and-force":23611,"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Felectric-field-maps":23612,"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Felectric-dipoles":23613,"\u002Felectricity-and-magnetism\u002Fcontinuous-charge-distributions\u002Fcontinuous-charge-fields":23614,"\u002Felectricity-and-magnetism\u002Fcontinuous-charge-distributions\u002Fgauss-law-and-conductors":23615,"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Fpoint-charge-potential":23616,"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Fpotential-gradients-and-equipotentials":23617,"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Felectrostatic-energy-and-pressure":23618,"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Flaplace-boundary-problems":23619,"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Fcontinuous-charge-potentials":23620,"\u002Felectricity-and-magnetism\u002Fcapacitance\u002Fcapacitance-fundamentals":23597,"\u002Felectricity-and-magnetism\u002Fcapacitance\u002Fcapacitor-networks":23621,"\u002Felectricity-and-magnetism\u002Fcapacitance\u002Fcapacitor-energy-and-force":23622,"\u002Felectricity-and-magnetism\u002Fcapacitance\u002Fdielectric-polarization-and-breakdown":23623,"\u002Felectricity-and-magnetism\u002Fdirect-current-circuits\u002Fcurrent-and-resistance":23593,"\u002Felectricity-and-magnetism\u002Fdirect-current-circuits\u002Fkirchhoff-network-analysis":23458,"\u002Felectricity-and-magnetism\u002Fdirect-current-circuits\u002Frc-transients":23624,"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fmagnetic-trajectories":23584,"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fhall-effect":23625,"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fmagnetic-force-on-conductors":23626,"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fmagnetic-dipoles":23627,"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fmass-spectrometry":23628,"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fmoving-charge-fields":23629,"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fbiot-savart-law":23630,"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fcircular-current-loops":23631,"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Famperes-law":23632,"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fgauss-law-for-magnetism":23633,"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fmagnetic-materials":23558,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Fmagnetic-flux":23634,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Ffaradays-law":23635,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Flenzs-law":23636,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Fmotional-emf":23637,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Feddy-currents":23638,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Fself-inductance":23639,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Fmagnetic-energy":23640,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Frl-circuits":23641,"\u002Felectricity-and-magnetism\u002Falternating-current\u002Fac-fundamentals":23576,"\u002Felectricity-and-magnetism\u002Falternating-current\u002Freactance":23575,"\u002Felectricity-and-magnetism\u002Falternating-current\u002Frlc-resonance":23642,"\u002Felectricity-and-magnetism\u002Falternating-current\u002Fac-power":23643,"\u002Felectricity-and-magnetism\u002Falternating-current\u002Ftransformers":23644,"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Fdisplacement-current":23645,"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Felectromagnetic-waves":23646,"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Felectromagnetic-momentum":23647,"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Fdipole-radiation":23648,"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Fpolarization":23649,"\u002Felectricity-and-magnetism\u002Foptics\u002Freflection-and-refraction":23650,"\u002Felectricity-and-magnetism\u002Foptics\u002Fthin-lenses":23602,"\u002Felectricity-and-magnetism\u002Foptics\u002Fspherical-mirrors":23600,"\u002Felectricity-and-magnetism":23651,"\u002Flinear-algebra\u002Flinear-systems\u002Fsystems-and-echelon-forms":23652,"\u002Flinear-algebra\u002Flinear-systems\u002Fvector-and-matrix-equations":23653,"\u002Flinear-algebra\u002Flinear-systems\u002Fsolution-sets-and-applications":23654,"\u002Flinear-algebra\u002Flinear-systems\u002Flinear-independence":23655,"\u002Flinear-algebra\u002Flinear-systems\u002Flinear-transformations":23656,"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fmatrix-operations":23657,"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fmatrix-inverse-and-invertibility":23658,"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fpartitioned-matrices-and-lu":23659,"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fsubspaces-dimension-rank":23660,"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fapplications-leontief-and-graphics":23510,"\u002Flinear-algebra\u002Fdeterminants\u002Fdeterminants-and-cofactors":23661,"\u002Flinear-algebra\u002Fdeterminants\u002Fproperties-of-determinants":23662,"\u002Flinear-algebra\u002Fdeterminants\u002Fcramer-volume-and-area":23514,"\u002Flinear-algebra\u002Fvector-spaces\u002Fvector-spaces-and-subspaces":23663,"\u002Flinear-algebra\u002Fvector-spaces\u002Fnull-and-column-spaces":23664,"\u002Flinear-algebra\u002Fvector-spaces\u002Fbases-and-independent-sets":23665,"\u002Flinear-algebra\u002Fvector-spaces\u002Fcoordinate-systems":23666,"\u002Flinear-algebra\u002Fvector-spaces\u002Fdimension-and-rank":23667,"\u002Flinear-algebra\u002Fvector-spaces\u002Fchange-of-basis":23668,"\u002Flinear-algebra\u002Fvector-spaces\u002Fdifference-equations-and-markov":23669,"\u002Flinear-algebra\u002Feigenvalues\u002Feigenvectors-and-eigenvalues":23670,"\u002Flinear-algebra\u002Feigenvalues\u002Fthe-characteristic-equation":23671,"\u002Flinear-algebra\u002Feigenvalues\u002Fdiagonalization":23672,"\u002Flinear-algebra\u002Feigenvalues\u002Feigenvectors-and-linear-transformations":23673,"\u002Flinear-algebra\u002Feigenvalues\u002Fcomplex-eigenvalues":23674,"\u002Flinear-algebra\u002Feigenvalues\u002Fdynamical-systems":23675,"\u002Flinear-algebra\u002Feigenvalues\u002Fpower-method":23676,"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Finner-product-length-orthogonality":23677,"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Forthogonal-sets-and-projections":23678,"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Fgram-schmidt-and-qr":23679,"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Fleast-squares-problems":23680,"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Fleast-squares-applications":23681,"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Finner-product-spaces":23682,"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fdiagonalizing-symmetric-matrices":23549,"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fquadratic-forms":23683,"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fconstrained-optimization":23684,"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fsingular-value-decomposition":23685,"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fsvd-applications-pca-imaging":23686,"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fnumerical-thinking-and-matrix-computation":23687,"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Flu-and-cholesky":23688,"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fconditioning-and-floating-point":23689,"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fstability-and-error-analysis":23690,"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fqr-and-numerical-least-squares":23691,"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fnumerical-eigenvalues-and-svd":23692,"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Faffine-combinations":23693,"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Faffine-independence-and-barycentric-coordinates":23694,"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Fconvex-combinations-and-convex-sets":23695,"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Fhyperplanes-and-polytopes":23696,"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Fcurves-and-surfaces":23697,"\u002Flinear-algebra":23698,"\u002Ftheory-of-computation":23699,"\u002Fcomputer-architecture\u002Ffoundations\u002Fbits-bytes-and-words":23700,"\u002Fcomputer-architecture\u002Ffoundations\u002Finteger-representation":23701,"\u002Fcomputer-architecture\u002Ffoundations\u002Finteger-arithmetic":23702,"\u002Fcomputer-architecture\u002Ffoundations\u002Ffloating-point":23703,"\u002Fcomputer-architecture\u002Ffoundations\u002Fboolean-algebra-and-bit-manipulation":23704,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fthe-machines-view":23705,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fdata-movement":23706,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Farithmetic-and-logic":23707,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fcontrol-flow":23708,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fprocedures":23709,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Farrays-structs-and-alignment":23710,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fmemory-layout-and-buffer-overflows":23711,"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Fwhat-an-isa-is":23712,"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Finstruction-formats-and-operands":23713,"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Faddressing-modes":23714,"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Fthe-y86-64-instruction-set":23715,"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Fy86-64-programming":23716,"\u002Fcomputer-architecture\u002Fdigital-logic\u002Ftransistors-gates-and-boolean-functions":23717,"\u002Fcomputer-architecture\u002Fdigital-logic\u002Fcombinational-logic-and-hcl":23718,"\u002Fcomputer-architecture\u002Fdigital-logic\u002Fmultiplexers-decoders-and-the-alu":23719,"\u002Fcomputer-architecture\u002Fdigital-logic\u002Fmemory-elements-latches-flip-flops-and-clocking":23720,"\u002Fcomputer-architecture\u002Fdigital-logic\u002Fregister-files-and-random-access-memory":23721,"\u002Fcomputer-architecture\u002Fprocessor-design\u002Fthe-fetch-decode-execute-cycle":23722,"\u002Fcomputer-architecture\u002Fprocessor-design\u002Fthe-seq-stages":23723,"\u002Fcomputer-architecture\u002Fprocessor-design\u002Fcontrol-logic-and-sequencing":23724,"\u002Fcomputer-architecture\u002Fprocessor-design\u002Fassembling-seq":23725,"\u002Fcomputer-architecture\u002Fprocessor-design\u002Ftracing-a-program":23726,"\u002Fcomputer-architecture\u002Fpipelining\u002Fpipelining-principles":23727,"\u002Fcomputer-architecture\u002Fpipelining\u002Ffrom-seq-to-pipe":23728,"\u002Fcomputer-architecture\u002Fpipelining\u002Fdata-hazards-stalling-and-forwarding":23729,"\u002Fcomputer-architecture\u002Fpipelining\u002Fcontrol-hazards-and-branch-prediction":23730,"\u002Fcomputer-architecture\u002Fpipelining\u002Fthe-complete-pipe-processor":23731,"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Fstorage-technologies-and-the-latency-gap":23732,"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Flocality":23733,"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Fcache-memories-direct-mapped":23734,"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Fset-associative-and-write-policies":23735,"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Fcache-performance-and-cache-friendly-code":23736,"\u002Fcomputer-architecture\u002Fvirtual-memory\u002Faddress-spaces-and-translation":23737,"\u002Fcomputer-architecture\u002Fvirtual-memory\u002Fpage-tables-and-page-faults":23738,"\u002Fcomputer-architecture\u002Fvirtual-memory\u002Fthe-tlb-and-multi-level-page-tables":23739,"\u002Fcomputer-architecture\u002Fexceptions-and-io\u002Fexceptional-control-flow":23740,"\u002Fcomputer-architecture\u002Fexceptions-and-io\u002Finterrupts-and-the-kernel":23741,"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fprocesses-threads-and-parallelism":23742,"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fhardware-multithreading":23743,"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fcache-coherence":23744,"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fmemory-consistency-and-synchronization":23745,"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fmulticore-organization":23746,"\u002Fcomputer-architecture\u002Fcapstone\u002Fthe-whole-machine":23747,"\u002Fcomputer-architecture\u002Fcapstone\u002Fassembling-a-complete-cpu":23748,"\u002Fcomputer-architecture":23699,"\u002Fdifferential-equations\u002Ffoundations\u002Fmodels-and-direction-fields":23749,"\u002Fdifferential-equations\u002Ffoundations\u002Fclassification-and-terminology":23750,"\u002Fdifferential-equations\u002Ffirst-order\u002Flinear-first-order-integrating-factors":23751,"\u002Fdifferential-equations\u002Ffirst-order\u002Fseparable-and-exact":23514,"\u002Fdifferential-equations\u002Ffirst-order\u002Fmodeling-first-order":23752,"\u002Fdifferential-equations\u002Ffirst-order\u002Fautonomous-and-population-dynamics":23513,"\u002Fdifferential-equations\u002Ffirst-order\u002Fexistence-uniqueness-euler":23520,"\u002Fdifferential-equations\u002Ffirst-order\u002Ffirst-order-difference-equations":23753,"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fhomogeneous-constant-coefficients":23754,"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fcomplex-and-repeated-roots":23554,"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fnonhomogeneous-undetermined-coefficients":23755,"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fvariation-of-parameters":23756,"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fmechanical-electrical-vibrations":23757,"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fhigher-order-linear":23758,"\u002Fdifferential-equations\u002Fseries-solutions\u002Fpower-series-ordinary-points":23759,"\u002Fdifferential-equations\u002Fseries-solutions\u002Fregular-singular-frobenius":23760,"\u002Fdifferential-equations\u002Fseries-solutions\u002Fbessel-and-special-functions":23761,"\u002Fdifferential-equations\u002Flaplace\u002Flaplace-definition-ivps":23762,"\u002Fdifferential-equations\u002Flaplace\u002Fstep-impulse-convolution":23763,"\u002Fdifferential-equations\u002Fsystems\u002Fmatrices-eigenvalues-review":23764,"\u002Fdifferential-equations\u002Fsystems\u002Fconstant-coefficient-systems-phase-portraits":23765,"\u002Fdifferential-equations\u002Fsystems\u002Frepeated-eigenvalues-fundamental-matrices":23766,"\u002Fdifferential-equations\u002Fnumerical\u002Feuler-and-runge-kutta":23761,"\u002Fdifferential-equations\u002Fnumerical\u002Fmultistep-systems-stability":23767,"\u002Fdifferential-equations\u002Fnonlinear\u002Fphase-plane-autonomous-stability":23768,"\u002Fdifferential-equations\u002Fnonlinear\u002Flocally-linear-and-liapunov":23769,"\u002Fdifferential-equations\u002Fnonlinear\u002Fcompeting-species-predator-prey-limit-cycles":23770,"\u002Fdifferential-equations\u002Fpdes-fourier-bvp\u002Ffourier-series":23771,"\u002Fdifferential-equations\u002Fpdes-fourier-bvp\u002Fheat-wave-laplace-equations":23772,"\u002Fdifferential-equations\u002Fpdes-fourier-bvp\u002Fsturm-liouville":23773,"\u002Fdifferential-equations\u002Fhistory-variations\u002Fcalculus-of-variations":23774,"\u002Fdifferential-equations\u002Fhistory-variations\u002Fhistorical-notes":23775,"\u002Fdifferential-equations":23776,"\u002Frelativity\u002Ffoundations\u002Fspecial-relativity-postulates":23777,"\u002Frelativity\u002Ffoundations\u002Florentz-transformation-spacetime":23778,"\u002Frelativity\u002Ffoundations\u002Ftime-dilation-length-contraction":23779,"\u002Frelativity\u002Ffoundations\u002Frelativistic-momentum-energy":23780,"\u002Frelativity\u002Ffoundations\u002Fgeneral-relativity":23659,"\u002Frelativity\u002Fspacetime-and-the-lorentz-group\u002Fminkowski-spacetime-and-the-interval":23781,"\u002Frelativity\u002Fspacetime-and-the-lorentz-group\u002Ffour-vectors-and-index-notation":23782,"\u002Frelativity\u002Fspacetime-and-the-lorentz-group\u002Fthe-lorentz-group-and-rapidity":23783,"\u002Frelativity\u002Fspacetime-and-the-lorentz-group\u002Fdoppler-aberration-and-appearance":23784,"\u002Frelativity\u002Frelativistic-dynamics\u002Ffour-momentum-force-and-accelerated-motion":23785,"\u002Frelativity\u002Frelativistic-dynamics\u002Fparticle-decays-and-two-body-kinematics":23786,"\u002Frelativity\u002Frelativistic-dynamics\u002Fcollisions-thresholds-and-the-cm-frame":23529,"\u002Frelativity\u002Frelativistic-dynamics\u002Fmandelstam-variables-and-invariants":23787,"\u002Frelativity\u002Fcovariant-electrodynamics\u002Ffour-current-and-the-four-potential":23788,"\u002Frelativity\u002Fcovariant-electrodynamics\u002Fthe-electromagnetic-field-tensor":23789,"\u002Frelativity\u002Fcovariant-electrodynamics\u002Ftransformation-of-electric-and-magnetic-fields":23790,"\u002Frelativity\u002Fcovariant-electrodynamics\u002Fcovariant-maxwell-and-the-stress-energy-tensor":23791,"\u002Frelativity\u002Fcurved-spacetime\u002Fthe-equivalence-principle-formalized":23792,"\u002Frelativity\u002Fcurved-spacetime\u002Fmanifolds-vectors-and-the-metric":23793,"\u002Frelativity\u002Fcurved-spacetime\u002Fcovariant-derivative-and-christoffel-symbols":23794,"\u002Frelativity\u002Fcurved-spacetime\u002Fgeodesics-and-the-geodesic-equation":23795,"\u002Frelativity\u002Fcurved-spacetime\u002Fcurvature-riemann-and-geodesic-deviation":23796,"\u002Frelativity\u002Fcurved-spacetime\u002Fthe-einstein-field-equations":23751,"\u002Frelativity\u002Fthe-schwarzschild-solution\u002Fthe-schwarzschild-metric":23797,"\u002Frelativity\u002Fthe-schwarzschild-solution\u002Fgeodesics-and-orbits-in-schwarzschild":23798,"\u002Frelativity\u002Fthe-schwarzschild-solution\u002Flight-bending-and-null-geodesics":23799,"\u002Frelativity\u002Ftests-of-general-relativity\u002Fperihelion-precession-of-mercury":23800,"\u002Frelativity\u002Ftests-of-general-relativity\u002Fdeflection-of-light-and-gravitational-lensing":23801,"\u002Frelativity\u002Ftests-of-general-relativity\u002Fgravitational-redshift-and-shapiro-delay":23680,"\u002Frelativity\u002Ftests-of-general-relativity\u002Frelativity-in-technology-gps":23802,"\u002Frelativity\u002Fblack-holes\u002Fhorizons-and-coordinate-singularities":23803,"\u002Frelativity\u002Fblack-holes\u002Frotating-and-charged-black-holes":23690,"\u002Frelativity\u002Fblack-holes\u002Fblack-hole-thermodynamics":23804,"\u002Frelativity\u002Fgravitational-waves\u002Flinearized-gravity-and-wave-solutions":23805,"\u002Frelativity\u002Fgravitational-waves\u002Fgeneration-and-the-quadrupole-formula":23806,"\u002Frelativity\u002Fgravitational-waves\u002Fdetection-ligo-and-the-first-events":23807,"\u002Frelativity\u002Fcosmological-bridge\u002Fthe-cosmological-principle-and-flrw-metric":23808,"\u002Frelativity\u002Fcosmological-bridge\u002Ffriedmann-equations-and-cosmic-dynamics":23809,"\u002Frelativity":23810,"\u002Fphysical-computing":23699,"\u002Fquantum-mechanics\u002Fold-quantum-theory\u002Fblackbody-radiation-and-the-planck-quantum":23811,"\u002Fquantum-mechanics\u002Fold-quantum-theory\u002Fthe-photoelectric-effect-and-the-photon":23790,"\u002Fquantum-mechanics\u002Fold-quantum-theory\u002Fx-rays-and-the-compton-effect":23812,"\u002Fquantum-mechanics\u002Fold-quantum-theory\u002Fthe-old-quantum-theory-bohr-and-sommerfeld":23813,"\u002Fquantum-mechanics\u002Fmatter-waves\u002Fde-broglie-waves-and-electron-diffraction":23814,"\u002Fquantum-mechanics\u002Fmatter-waves\u002Fwave-packets-and-the-probability-interpretation":23815,"\u002Fquantum-mechanics\u002Fmatter-waves\u002Fthe-uncertainty-principle":23816,"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fthe-schrodinger-equation-in-one-dimension":23817,"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fthe-free-particle-and-wave-packet-dynamics":23818,"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fparticle-in-infinite-and-finite-square-wells":23766,"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Foperators-expectation-values-and-the-harmonic-oscillator":23690,"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fthe-dirac-delta-potential":23819,"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fbarrier-penetration-and-quantum-tunneling":23820,"\u002Fquantum-mechanics\u002Fformalism\u002Fhilbert-space-and-dirac-notation":23821,"\u002Fquantum-mechanics\u002Fformalism\u002Fobservables-hermitian-operators-and-eigenvalues":23822,"\u002Fquantum-mechanics\u002Fformalism\u002Fthe-postulates-and-quantum-measurement":23818,"\u002Fquantum-mechanics\u002Fformalism\u002Fposition-momentum-and-continuous-spectra":23823,"\u002Fquantum-mechanics\u002Fformalism\u002Fcommutators-and-the-generalized-uncertainty-principle":23798,"\u002Fquantum-mechanics\u002Fformalism\u002Ftime-evolution-schrodinger-and-heisenberg-pictures":23551,"\u002Fquantum-mechanics\u002Foscillator-and-symmetry\u002Fladder-operators-and-the-number-states":23824,"\u002Fquantum-mechanics\u002Foscillator-and-symmetry\u002Fcoherent-and-squeezed-states":23825,"\u002Fquantum-mechanics\u002Foscillator-and-symmetry\u002Fsymmetries-generators-and-conservation-laws":23531,"\u002Fquantum-mechanics\u002Foscillator-and-symmetry\u002Fparity-time-reversal-and-discrete-symmetries":23826,"\u002Fquantum-mechanics\u002Fangular-momentum\u002Forbital-angular-momentum-and-spherical-harmonics":23827,"\u002Fquantum-mechanics\u002Fangular-momentum\u002Fthe-angular-momentum-algebra":23828,"\u002Fquantum-mechanics\u002Fangular-momentum\u002Faddition-of-angular-momenta-and-clebsch-gordan":23829,"\u002Fquantum-mechanics\u002Fcentral-potentials\u002Fthe-schrodinger-equation-in-three-dimensions":23830,"\u002Fquantum-mechanics\u002Fcentral-potentials\u002Fthe-hydrogen-atom":23831,"\u002Fquantum-mechanics\u002Fcentral-potentials\u002Fthe-isotropic-oscillator-and-hidden-symmetry":23832,"\u002Fquantum-mechanics\u002Fspin\u002Fspin-half-pauli-matrices-and-stern-gerlach":23833,"\u002Fquantum-mechanics\u002Fspin\u002Fspin-in-a-magnetic-field-precession-and-resonance":23834,"\u002Fquantum-mechanics\u002Fspin\u002Ftwo-level-systems-and-the-bloch-sphere":23790,"\u002Fquantum-mechanics\u002Fidentical-particles\u002Fidentical-particles-and-exchange-symmetry":23835,"\u002Fquantum-mechanics\u002Fidentical-particles\u002Fthe-pauli-principle-atoms-and-the-periodic-table":23836,"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Ftime-independent-perturbation-theory":23837,"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Ffine-structure-and-the-real-hydrogen-atom":23824,"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Fthe-zeeman-and-stark-effects":23512,"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Fthe-variational-method":23838,"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Fthe-wkb-approximation":23839,"\u002Fquantum-mechanics":23840,"\u002Freal-analysis\u002Ffoundations\u002Fsets-logic-functions":23774,"\u002Freal-analysis\u002Ffoundations\u002Fordered-fields-completeness":23841,"\u002Freal-analysis\u002Ffoundations\u002Fabsolute-value-bounds":23842,"\u002Freal-analysis\u002Ffoundations\u002Fintervals-uncountability":23663,"\u002Freal-analysis\u002Fsequences-series\u002Fsequences-limits":23843,"\u002Freal-analysis\u002Fsequences-series\u002Flimit-laws-monotone":23549,"\u002Freal-analysis\u002Fsequences-series\u002Flimsup-bolzano-weierstrass":23844,"\u002Freal-analysis\u002Fsequences-series\u002Fcauchy-completeness":23845,"\u002Freal-analysis\u002Fsequences-series\u002Fseries-convergence":23686,"\u002Freal-analysis\u002Fsequences-series\u002Fabsolute-conditional-rearrangement":23796,"\u002Freal-analysis\u002Fmetric-spaces\u002Fmetric-spaces-norms":23846,"\u002Freal-analysis\u002Fmetric-spaces\u002Fopen-closed-sets":23847,"\u002Freal-analysis\u002Fmetric-spaces\u002Fconvergence-completeness":23848,"\u002Freal-analysis\u002Fmetric-spaces\u002Fcompactness":23849,"\u002Freal-analysis\u002Fmetric-spaces\u002Fconnectedness":23850,"\u002Freal-analysis\u002Fcontinuity\u002Flimits-of-functions":23823,"\u002Freal-analysis\u002Fcontinuity\u002Fcontinuous-functions":23851,"\u002Freal-analysis\u002Fcontinuity\u002Fevt-ivt":23656,"\u002Freal-analysis\u002Fcontinuity\u002Funiform-continuity":23852,"\u002Freal-analysis\u002Fcontinuity\u002Fcontinuity-metric-spaces":23853,"\u002Freal-analysis\u002Fcontinuity\u002Flimits-infinity-monotone":23510,"\u002Freal-analysis\u002Fdifferentiation\u002Fthe-derivative":23854,"\u002Freal-analysis\u002Fdifferentiation\u002Fmean-value-theorem":23855,"\u002Freal-analysis\u002Fdifferentiation\u002Ftaylors-theorem":23812,"\u002Freal-analysis\u002Fdifferentiation\u002Finverse-function-1d":23539,"\u002Freal-analysis\u002Friemann-integration\u002Fdarboux-integral":23690,"\u002Freal-analysis\u002Friemann-integration\u002Fintegrability-classes":23856,"\u002Freal-analysis\u002Friemann-integration\u002Fproperties-of-the-integral":23857,"\u002Freal-analysis\u002Friemann-integration\u002Ffundamental-theorem":23675,"\u002Freal-analysis\u002Friemann-integration\u002Flog-exp-improper":23800,"\u002Freal-analysis\u002Ffunction-sequences\u002Fpointwise-uniform-convergence":23858,"\u002Freal-analysis\u002Ffunction-sequences\u002Finterchange-of-limits":23859,"\u002Freal-analysis\u002Ffunction-sequences\u002Fpower-series-weierstrass":23860,"\u002Freal-analysis\u002Ffunction-sequences\u002Fpicard-ode":23695,"\u002Freal-analysis\u002Fseveral-variables\u002Fdifferentiability-rn":23861,"\u002Freal-analysis\u002Fseveral-variables\u002Fgradient-chain-rule":23862,"\u002Freal-analysis\u002Fseveral-variables\u002Fhigher-derivatives-taylor-extrema":23863,"\u002Freal-analysis\u002Fseveral-variables\u002Finverse-implicit-theorems":23863,"\u002Freal-analysis\u002Fseveral-variables\u002Fmultiple-integrals":23864,"\u002Freal-analysis":23865,"\u002Fabstract-algebra\u002Ffoundations\u002Fsets-functions-relations":23866,"\u002Fabstract-algebra\u002Ffoundations\u002Fintegers-and-modular-arithmetic":23867,"\u002Fabstract-algebra\u002Fgroups-and-symmetry\u002Fgroup-axioms-and-first-examples":23868,"\u002Fabstract-algebra\u002Fgroups-and-symmetry\u002Fdihedral-and-symmetric-groups":23869,"\u002Fabstract-algebra\u002Fgroups-and-symmetry\u002Fmatrix-and-quaternion-groups":23870,"\u002Fabstract-algebra\u002Fgroups-and-symmetry\u002Fhomomorphisms-and-group-actions":23871,"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fsubgroups-and-substructures":23872,"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fcyclic-groups":23873,"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fgeneration-and-subgroup-lattices":23874,"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fcosets-lagrange-and-normal-subgroups":23875,"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fisomorphism-theorems":23839,"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fcomposition-series-and-the-alternating-group":23876,"\u002Fabstract-algebra\u002Fgroup-actions-and-sylow\u002Factions-and-cayleys-theorem":23868,"\u002Fabstract-algebra\u002Fgroup-actions-and-sylow\u002Fconjugation-and-the-class-equation":23785,"\u002Fabstract-algebra\u002Fgroup-actions-and-sylow\u002Fsylow-theorems":23877,"\u002Fabstract-algebra\u002Fgroup-actions-and-sylow\u002Fautomorphisms-and-simple-groups":23878,"\u002Fabstract-algebra\u002Fproducts-and-group-structure\u002Fdirect-products-and-finite-abelian-groups":23879,"\u002Fabstract-algebra\u002Fproducts-and-group-structure\u002Fsemidirect-products":23880,"\u002Fabstract-algebra\u002Fproducts-and-group-structure\u002Fnilpotent-and-solvable-groups":23881,"\u002Fabstract-algebra\u002Fproducts-and-group-structure\u002Fclassifying-small-groups":23882,"\u002Fabstract-algebra\u002Fring-theory\u002Frings-definitions-and-examples":23883,"\u002Fabstract-algebra\u002Fring-theory\u002Fideals-quotients-and-homomorphisms":23884,"\u002Fabstract-algebra\u002Fring-theory\u002Ffractions-and-the-chinese-remainder-theorem":23878,"\u002Fabstract-algebra\u002Ffactorization-and-polynomials\u002Feuclidean-domains-pids-ufds":23885,"\u002Fabstract-algebra\u002Ffactorization-and-polynomials\u002Fpolynomial-rings-over-fields":23854,"\u002Fabstract-algebra\u002Ffactorization-and-polynomials\u002Fgauss-lemma-and-unique-factorization":23886,"\u002Fabstract-algebra\u002Ffactorization-and-polynomials\u002Firreducibility-criteria-and-groebner":23887,"\u002Fabstract-algebra\u002Fmodule-theory\u002Fintroduction-to-modules":23888,"\u002Fabstract-algebra\u002Fmodule-theory\u002Ffree-modules-and-direct-sums":23889,"\u002Fabstract-algebra\u002Fmodule-theory\u002Ftensor-products-and-exact-sequences":23890,"\u002Fabstract-algebra\u002Fmodule-theory\u002Fvector-spaces-and-linear-maps":23891,"\u002Fabstract-algebra\u002Fmodules-over-pids\u002Fstructure-theorem-over-pids":23892,"\u002Fabstract-algebra\u002Fmodules-over-pids\u002Frational-canonical-form":23893,"\u002Fabstract-algebra\u002Fmodules-over-pids\u002Fjordan-canonical-form":23894,"\u002Fabstract-algebra\u002Ffield-theory\u002Ffield-extensions-and-algebraic-elements":23895,"\u002Fabstract-algebra\u002Ffield-theory\u002Fstraightedge-and-compass-constructions":23528,"\u002Fabstract-algebra\u002Ffield-theory\u002Fsplitting-fields-and-algebraic-closure":23896,"\u002Fabstract-algebra\u002Ffield-theory\u002Fseparable-and-cyclotomic-extensions":23897,"\u002Fabstract-algebra\u002Fgalois-theory\u002Fthe-galois-correspondence":23770,"\u002Fabstract-algebra\u002Fgalois-theory\u002Ffinite-fields":23898,"\u002Fabstract-algebra\u002Fgalois-theory\u002Fcyclotomic-and-abelian-extensions":23899,"\u002Fabstract-algebra\u002Fgalois-theory\u002Fgalois-groups-of-polynomials":23827,"\u002Fabstract-algebra\u002Fgalois-theory\u002Fsolvability-by-radicals-and-the-quintic":23899,"\u002Fabstract-algebra\u002Fcapstone\u002Fcommutative-algebra-and-algebraic-geometry":23900,"\u002Fabstract-algebra\u002Fcapstone\u002Frepresentation-and-character-theory":23901,"\u002Fabstract-algebra":23902,"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fatomic-spectra-rutherford":23903,"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fbohr-model-hydrogen":23904,"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fx-ray-spectra-franck-hertz":23905,"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fbohr-sommerfeld-old-quantum-theory":23906,"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fold-quantum-theory-limits-wkb":23907,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fschrodinger-3d-hydrogen":23819,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fhydrogen-wave-functions":23908,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fradial-equation-in-full":23909,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fsymmetry-degeneracy-runge-lenz":23910,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fexpectation-values-virial":23911,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fquantum-defects-alkali-spectra":23912,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Frydberg-atoms":23913,"\u002Fatomic-physics\u002Ffine-structure-and-the-dirac-atom\u002Frelativistic-kinetic-correction":23914,"\u002Fatomic-physics\u002Ffine-structure-and-the-dirac-atom\u002Fspin-orbit-thomas-precession":23653,"\u002Fatomic-physics\u002Ffine-structure-and-the-dirac-atom\u002Fdarwin-term-fine-structure-formula":23782,"\u002Fatomic-physics\u002Ffine-structure-and-the-dirac-atom\u002Fdirac-equation-hydrogen":23528,"\u002Fatomic-physics\u002Fqed-corrections-and-hyperfine-structure\u002Flamb-shift-qed":23915,"\u002Fatomic-physics\u002Fqed-corrections-and-hyperfine-structure\u002Fhyperfine-structure-21cm":23512,"\u002Fatomic-physics\u002Fqed-corrections-and-hyperfine-structure\u002Fnuclear-effects-isotope-shift":23916,"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fperiodic-table-atomic-spectra":23917,"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fcentral-field-self-consistent":23550,"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fidentical-particles-hartree-fock":23844,"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fhelium-two-electron-atom":23918,"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fls-jj-coupling-term-symbols":23919,"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fhund-rules-ground-terms":23920,"\u002Fatomic-physics\u002Fatoms-in-external-fields\u002Fzeeman-effect":23921,"\u002Fatomic-physics\u002Fatoms-in-external-fields\u002Fpaschen-back-intermediate":23922,"\u002Fatomic-physics\u002Fatoms-in-external-fields\u002Fstark-effect-polarizability":23923,"\u002Fatomic-physics\u002Fradiative-transitions-and-line-shapes\u002Ftime-dependent-perturbation-golden-rule":23924,"\u002Fatomic-physics\u002Fradiative-transitions-and-line-shapes\u002Fdipole-approximation-einstein-coefficients":23925,"\u002Fatomic-physics\u002Fradiative-transitions-and-line-shapes\u002Fselection-rules-forbidden-transitions":23926,"\u002Fatomic-physics\u002Fradiative-transitions-and-line-shapes\u002Flifetimes-and-line-shapes":23927,"\u002Fatomic-physics\u002Flasers-and-spectroscopy\u002Flaser-principles":23928,"\u002Fatomic-physics\u002Flasers-and-spectroscopy\u002Fspectroscopy-techniques":23929,"\u002Fatomic-physics\u002Flasers-and-spectroscopy\u002Fline-catalog-nist-asd":23930,"\u002Fatomic-physics\u002Fmodern-atomic-physics\u002Flaser-cooling-doppler":23931,"\u002Fatomic-physics\u002Fmodern-atomic-physics\u002Fsub-doppler-trapping":23875,"\u002Fatomic-physics\u002Fmodern-atomic-physics\u002Fbose-einstein-condensation":23932,"\u002Fatomic-physics\u002Fmodern-atomic-physics\u002Foptical-clocks-precision":23933,"\u002Fatomic-physics":23934,"\u002Fdatabases":23699,"\u002Fcategory-theory\u002Ffoundations\u002Fwhat-is-a-category":23935,"\u002Fcategory-theory\u002Ffoundations\u002Fexamples-of-categories":23936,"\u002Fcategory-theory\u002Ffoundations\u002Fspecial-morphisms":23937,"\u002Fcategory-theory\u002Ffoundations\u002Ffunctors":23866,"\u002Fcategory-theory\u002Ffoundations\u002Fnatural-transformations":23938,"\u002Fcategory-theory\u002Ffoundations\u002Fsize-and-set-theory":23939,"\u002Fcategory-theory\u002Funiversal-properties\u002Funiversal-properties":23940,"\u002Fcategory-theory\u002Funiversal-properties\u002Fproducts-and-coproducts":23941,"\u002Fcategory-theory\u002Funiversal-properties\u002Fconstructions-on-categories":23942,"\u002Fcategory-theory\u002Frepresentables-yoneda\u002Frepresentable-functors":23943,"\u002Fcategory-theory\u002Frepresentables-yoneda\u002Fyoneda-lemma":23944,"\u002Fcategory-theory\u002Frepresentables-yoneda\u002Fyoneda-consequences":23945,"\u002Fcategory-theory\u002Flimits-colimits\u002Flimits":23946,"\u002Fcategory-theory\u002Flimits-colimits\u002Fproducts-equalizers-pullbacks":23947,"\u002Fcategory-theory\u002Flimits-colimits\u002Fcolimits":23948,"\u002Fcategory-theory\u002Flimits-colimits\u002Fcomputing-limits":23949,"\u002Fcategory-theory\u002Flimits-colimits\u002Flimits-and-functors":23950,"\u002Fcategory-theory\u002Fadjunctions\u002Fadjunctions":23951,"\u002Fcategory-theory\u002Fadjunctions\u002Funits-and-counits":23952,"\u002Fcategory-theory\u002Fadjunctions\u002Fadjunctions-via-universal-arrows":23953,"\u002Fcategory-theory\u002Fadjunctions\u002Ffree-forgetful-adjunctions":23954,"\u002Fcategory-theory\u002Fadjoints-limits\u002Flimits-via-adjoints":23955,"\u002Fcategory-theory\u002Fadjoints-limits\u002Fpresheaf-limits-colimits":23956,"\u002Fcategory-theory\u002Fadjoints-limits\u002Fadjoints-preserve-limits":23957,"\u002Fcategory-theory\u002Fadjoints-limits\u002Fadjoint-functor-theorem":23948,"\u002Fcategory-theory\u002Fmonads-algebras\u002Fmonads":23958,"\u002Fcategory-theory\u002Fmonads-algebras\u002Falgebras-eilenberg-moore":23959,"\u002Fcategory-theory\u002Fmonads-algebras\u002Fkleisli-and-programming":23960,"\u002Fcategory-theory\u002Fmonads-algebras\u002Falgebras-for-endofunctors":23961,"\u002Fcategory-theory\u002Fcartesian-closed-lambda\u002Fcartesian-closed-categories":23962,"\u002Fcategory-theory\u002Fcartesian-closed-lambda\u002Flambda-calculus-correspondence":23907,"\u002Fcategory-theory\u002Fcartesian-closed-lambda\u002Ffixed-points-and-recursion":23963,"\u002Fcategory-theory":23964,"\u002Fdeep-learning\u002Fmathematical-background\u002Flinear-algebra-for-deep-learning":23965,"\u002Fdeep-learning\u002Fmathematical-background\u002Fprobability-and-information-theory":23966,"\u002Fdeep-learning\u002Fmathematical-background\u002Fnumerical-computation":23967,"\u002Fdeep-learning\u002Fmathematical-background\u002Fcalculus":23968,"\u002Fdeep-learning\u002Ffoundations\u002Fwhat-is-deep-learning":23969,"\u002Fdeep-learning\u002Ffoundations\u002Fmachine-learning-refresher":23970,"\u002Fdeep-learning\u002Ffoundations\u002Flinear-models-and-the-perceptron":23929,"\u002Fdeep-learning\u002Fneural-networks\u002Fthe-multilayer-perceptron":23971,"\u002Fdeep-learning\u002Fneural-networks\u002Factivation-functions":23972,"\u002Fdeep-learning\u002Fneural-networks\u002Funiversal-approximation":23973,"\u002Fdeep-learning\u002Fneural-networks\u002Fbackpropagation":23974,"\u002Fdeep-learning\u002Fneural-networks\u002Floss-functions-and-output-units":23975,"\u002Fdeep-learning\u002Foptimization\u002Fgradient-descent-and-sgd":23976,"\u002Fdeep-learning\u002Foptimization\u002Fmomentum-and-adaptive-methods":23977,"\u002Fdeep-learning\u002Foptimization\u002Finitialization":23978,"\u002Fdeep-learning\u002Foptimization\u002Fthe-optimization-landscape":23979,"\u002Fdeep-learning\u002Foptimization\u002Fsecond-order-and-approximate-methods":23980,"\u002Fdeep-learning\u002Fregularization\u002Fregularization-overview":23981,"\u002Fdeep-learning\u002Fregularization\u002Fdropout-and-data-augmentation":23982,"\u002Fdeep-learning\u002Fregularization\u002Fearly-stopping-and-parameter-sharing":23983,"\u002Fdeep-learning\u002Fregularization\u002Fnormalization":23984,"\u002Fdeep-learning\u002Farchitectures\u002Fconvolutional-networks":23985,"\u002Fdeep-learning\u002Farchitectures\u002Fcnn-architectures":23986,"\u002Fdeep-learning\u002Farchitectures\u002Frecurrent-networks":23987,"\u002Fdeep-learning\u002Farchitectures\u002Flstm-and-gru":23988,"\u002Fdeep-learning\u002Farchitectures\u002Fattention-and-transformers":23412,"\u002Fdeep-learning\u002Farchitectures\u002Fthe-transformer-architecture":23989,"\u002Fdeep-learning\u002Farchitectures\u002Ftransformers-in-practice":23990,"\u002Fdeep-learning\u002Farchitectures\u002Fgraph-neural-networks":23991,"\u002Fdeep-learning\u002Farchitectures\u002Fstate-space-models":23992,"\u002Fdeep-learning\u002Ftheory\u002Fgeneralization-theory":23993,"\u002Fdeep-learning\u002Ftheory\u002Fadversarial-robustness":23994,"\u002Fdeep-learning\u002Ftheory\u002Fadversarial-defenses":23995,"\u002Fdeep-learning\u002Ftheory\u002Fbayesian-and-ensemble-methods":23996,"\u002Fdeep-learning\u002Ftheory\u002Fdeep-equilibrium-models":23941,"\u002Fdeep-learning\u002Fgenerative-models\u002Flinear-factor-models":23997,"\u002Fdeep-learning\u002Fgenerative-models\u002Fautoencoders":23998,"\u002Fdeep-learning\u002Fgenerative-models\u002Fvariational-autoencoders":23999,"\u002Fdeep-learning\u002Fgenerative-models\u002Fgenerative-adversarial-networks":24000,"\u002Fdeep-learning\u002Fgenerative-models\u002Fautoregressive-and-normalizing-flows":24001,"\u002Fdeep-learning\u002Fgenerative-models\u002Fenergy-based-and-boltzmann-machines":24002,"\u002Fdeep-learning\u002Fgenerative-models\u002Fdiffusion-and-score-based-models":24003,"\u002Fdeep-learning\u002Fprobabilistic-methods\u002Fstructured-probabilistic-models":23493,"\u002Fdeep-learning\u002Fprobabilistic-methods\u002Fmonte-carlo-and-mcmc":24004,"\u002Fdeep-learning\u002Fprobabilistic-methods\u002Fapproximate-inference":24005,"\u002Fdeep-learning\u002Fpractical\u002Fpractical-methodology":23729,"\u002Fdeep-learning\u002Fpractical\u002Fhyperparameters-and-debugging":24006,"\u002Fdeep-learning\u002Fpractical\u002Frepresentation-learning":24007,"\u002Fdeep-learning\u002Fpractical\u002Ftransfer-learning":24008,"\u002Fdeep-learning\u002Fpractical\u002Fapplications":24009,"\u002Fdeep-learning\u002Fpractical\u002Fmodel-compression-and-distillation":24010,"\u002Fdeep-learning\u002Fpractical\u002Fmeta-learning-and-few-shot":24011,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Flarge-language-models":24012,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fscaling-inference-and-alignment":24013,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fseq2seq-pretraining-and-bart":24014,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Ftext-to-text-transfer-and-conditional-generation":24015,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fspeech-and-audio-models":24016,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fself-supervised-speech-and-synthesis":24017,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fai-agents":23702,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fagent-memory-retrieval-and-orchestration":24018,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fmixture-of-experts":24019,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fmultimodal-models":24020,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Ffusion-and-vision-language-models":24021,"\u002Fdeep-learning\u002Freinforcement-learning\u002Ffoundations-of-reinforcement-learning":23739,"\u002Fdeep-learning\u002Freinforcement-learning\u002Fmodel-free-prediction-and-control":24022,"\u002Fdeep-learning\u002Freinforcement-learning\u002Fdeep-q-networks":24023,"\u002Fdeep-learning\u002Freinforcement-learning\u002Fpolicy-gradients-and-actor-critic":24024,"\u002Fdeep-learning\u002Freinforcement-learning\u002Frl-from-human-feedback":24025,"\u002Fdeep-learning":23699,"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Fequilibrium-state-variables-zeroth-law":24026,"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Ffirst-law-heat-and-work":23824,"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Fsecond-law-entropy-and-the-carnot-bound":24027,"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Fthermodynamic-potentials-and-maxwell-relations":24028,"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Fstability-response-functions-and-the-third-law":24029,"\u002Fstatistical-mechanics\u002Ffoundations\u002Fclassical-statistics-and-equipartition":24030,"\u002Fstatistical-mechanics\u002Ffoundations\u002Fphase-space-and-liouvilles-theorem":24031,"\u002Fstatistical-mechanics\u002Ffoundations\u002Fensembles-and-the-equal-probability-postulate":24032,"\u002Fstatistical-mechanics\u002Ffoundations\u002Fstatistical-entropy-boltzmann-and-gibbs":24033,"\u002Fstatistical-mechanics\u002Fmicrocanonical\u002Fmicrocanonical-ensemble-and-entropy":24034,"\u002Fstatistical-mechanics\u002Fmicrocanonical\u002Fequilibrium-conditions-temperature-pressure-chemical-potential":23871,"\u002Fstatistical-mechanics\u002Fmicrocanonical\u002Fideal-gas-phase-space-and-the-sackur-tetrode-entropy":24035,"\u002Fstatistical-mechanics\u002Fmicrocanonical\u002Ftwo-state-systems-paramagnets-and-negative-temperature":24036,"\u002Fstatistical-mechanics\u002Fcanonical\u002Fcanonical-ensemble-and-the-boltzmann-distribution":24037,"\u002Fstatistical-mechanics\u002Fcanonical\u002Fpartition-function-and-the-helmholtz-free-energy":23758,"\u002Fstatistical-mechanics\u002Fcanonical\u002Fenergy-fluctuations-and-ensemble-equivalence":23551,"\u002Fstatistical-mechanics\u002Fcanonical\u002Fthe-einstein-solid-and-harmonic-systems":24038,"\u002Fstatistical-mechanics\u002Fcanonical\u002Fparamagnetism-and-the-schottky-anomaly":24039,"\u002Fstatistical-mechanics\u002Fclassical-gas\u002Fideal-gas-partition-function-and-the-gibbs-paradox":24040,"\u002Fstatistical-mechanics\u002Fclassical-gas\u002Fequipartition-and-the-virial-theorem":23685,"\u002Fstatistical-mechanics\u002Fclassical-gas\u002Fmolecular-gases-rotation-and-vibration":24041,"\u002Fstatistical-mechanics\u002Fgrand-canonical\u002Fgrand-canonical-ensemble-and-the-grand-partition-function":24042,"\u002Fstatistical-mechanics\u002Fgrand-canonical\u002Fchemical-potential-fugacity-and-number-fluctuations":23763,"\u002Fstatistical-mechanics\u002Fgrand-canonical\u002Fensemble-summary-and-the-thermodynamic-web":23690,"\u002Fstatistical-mechanics\u002Fquantum-statistics\u002Fquantum-statistics-bose-einstein-and-fermi-dirac":24043,"\u002Fstatistical-mechanics\u002Fquantum-statistics\u002Fderiving-the-quantum-distributions":23659,"\u002Fstatistical-mechanics\u002Fquantum-statistics\u002Fthe-classical-limit-and-quantum-concentration":23529,"\u002Fstatistical-mechanics\u002Fquantum-statistics\u002Fideal-quantum-gases-general-framework":23545,"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fbose-einstein-condensation-and-the-fermion-gas":24044,"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fthe-photon-gas-and-plancks-radiation-law":24045,"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fblackbody-thermodynamics-and-radiation-pressure":24046,"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fphonons-and-the-debye-model":23677,"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fbose-einstein-condensation-derived":24047,"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fthermodynamics-of-the-bose-gas-and-superfluidity":23539,"\u002Fstatistical-mechanics\u002Ffermi-gas\u002Fthe-ideal-fermi-gas-at-zero-temperature":24048,"\u002Fstatistical-mechanics\u002Ffermi-gas\u002Fsommerfeld-expansion-and-electrons-in-metals":24049,"\u002Fstatistical-mechanics\u002Ffermi-gas\u002Fwhite-dwarfs-and-the-chandrasekhar-limit":24050,"\u002Fstatistical-mechanics\u002Ffermi-gas\u002Fneutron-stars-and-nuclear-matter":24051,"\u002Fstatistical-mechanics\u002Finteractions\u002Fthe-cluster-expansion-and-virial-coefficients":23873,"\u002Fstatistical-mechanics\u002Finteractions\u002Fthe-van-der-waals-gas-and-liquid-gas-coexistence":24052,"\u002Fstatistical-mechanics\u002Finteractions\u002Fquantum-gases-with-interactions-and-exchange":24053,"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fphases-coexistence-and-classification":24054,"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fthe-ising-model-and-exact-solutions":24055,"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fmean-field-theory-and-the-weiss-model":23819,"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fcritical-exponents-and-landau-theory":24056,"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fthe-renormalization-group-idea":23797,"\u002Fstatistical-mechanics\u002Ffluctuations\u002Fthermodynamic-fluctuations-and-response":24057,"\u002Fstatistical-mechanics\u002Ffluctuations\u002Fbrownian-motion-and-the-langevin-equation":23658,"\u002Fstatistical-mechanics\u002Ffluctuations\u002Flinear-response-and-the-fluctuation-dissipation-theorem":24058,"\u002Fstatistical-mechanics":24059,"\u002Fcondensed-matter\u002Fmolecules-and-bonding\u002Fbonding-mechanisms":24060,"\u002Fcondensed-matter\u002Fmolecules-and-bonding\u002Fmolecular-orbitals-and-h2-plus":23524,"\u002Fcondensed-matter\u002Fmolecules-and-bonding\u002Fhydrogen-molecule-and-exchange":23784,"\u002Fcondensed-matter\u002Fmolecules-and-bonding\u002Fvan-der-waals-forces":24061,"\u002Fcondensed-matter\u002Fmolecular-spectra\u002Frotational-vibrational-spectra":24062,"\u002Fcondensed-matter\u002Fmolecular-spectra\u002Fanharmonicity-and-rovibrational-structure":24063,"\u002Fcondensed-matter\u002Fmolecular-spectra\u002Framan-and-electronic-bands":24064,"\u002Fcondensed-matter\u002Fmolecular-spectra\u002Flasers-and-masers":24065,"\u002Fcondensed-matter\u002Fcrystal-structure\u002Fstructure-of-solids":24066,"\u002Fcondensed-matter\u002Fcrystal-structure\u002Fbravais-lattices-and-crystal-systems":23657,"\u002Fcondensed-matter\u002Fcrystal-structure\u002Freciprocal-lattice-and-brillouin-zones":24067,"\u002Fcondensed-matter\u002Fcrystal-structure\u002Fdiffraction-and-structure-factors":24068,"\u002Fcondensed-matter\u002Flattice-dynamics\u002Fphonon-dispersion":24069,"\u002Fcondensed-matter\u002Flattice-dynamics\u002Fphonons-quantization-and-dos":24070,"\u002Fcondensed-matter\u002Flattice-dynamics\u002Fdebye-einstein-heat-capacity":23800,"\u002Fcondensed-matter\u002Flattice-dynamics\u002Fanharmonicity-and-thermal-transport":24071,"\u002Fcondensed-matter\u002Ffree-electron-fermi-gas\u002Ffree-electron-gas-and-conduction":24072,"\u002Fcondensed-matter\u002Ffree-electron-fermi-gas\u002Fsommerfeld-model-and-heat-capacity":24073,"\u002Fcondensed-matter\u002Ffree-electron-fermi-gas\u002Ftransport-and-the-hall-effect":24074,"\u002Fcondensed-matter\u002Ffree-electron-fermi-gas\u002Fscreening-and-plasmons":24075,"\u002Fcondensed-matter\u002Fband-theory\u002Fblochs-theorem-and-energy-bands":23855,"\u002Fcondensed-matter\u002Fband-theory\u002Fnearly-free-electron-model":23799,"\u002Fcondensed-matter\u002Fband-theory\u002Ftight-binding-method":24076,"\u002Fcondensed-matter\u002Fband-theory\u002Ffermi-surfaces-and-semiclassical-dynamics":24077,"\u002Fcondensed-matter\u002Fsemiconductors\u002Fsemiconductor-bands-and-junctions":24078,"\u002Fcondensed-matter\u002Fsemiconductors\u002Fintrinsic-and-extrinsic-semiconductors":24079,"\u002Fcondensed-matter\u002Fsemiconductors\u002Fcarrier-transport-and-recombination":23790,"\u002Fcondensed-matter\u002Fsemiconductors\u002Fthe-pn-junction":24080,"\u002Fcondensed-matter\u002Fsemiconductors\u002Ftransistors-and-optoelectronics":24081,"\u002Fcondensed-matter\u002Fdielectrics-and-ferroelectrics\u002Fdielectrics-and-polarization":24026,"\u002Fcondensed-matter\u002Fdielectrics-and-ferroelectrics\u002Fferroelectrics-and-piezoelectrics":23927,"\u002Fcondensed-matter\u002Fmagnetism\u002Fdiamagnetism-and-paramagnetism":23654,"\u002Fcondensed-matter\u002Fmagnetism\u002Fexchange-and-ferromagnetism":24082,"\u002Fcondensed-matter\u002Fmagnetism\u002Fantiferromagnetism-and-domains":23520,"\u002Fcondensed-matter\u002Fmagnetism\u002Fspin-waves-and-magnons":24083,"\u002Fcondensed-matter\u002Fsuperconductivity\u002Fsuperconductivity-phenomenology":24084,"\u002Fcondensed-matter\u002Fsuperconductivity\u002Flondon-theory-and-the-meissner-effect":23665,"\u002Fcondensed-matter\u002Fsuperconductivity\u002Fginzburg-landau-theory":24085,"\u002Fcondensed-matter\u002Fsuperconductivity\u002Fbcs-theory":23920,"\u002Fcondensed-matter\u002Fsuperconductivity\u002Fjosephson-and-high-tc":24086,"\u002Fcondensed-matter\u002Fnanostructures\u002Fquantum-wells-wires-and-dots":23512,"\u002Fcondensed-matter\u002Fnanostructures\u002Finteger-quantum-hall-effect":24087,"\u002Fcondensed-matter\u002Fnanostructures\u002Ffractional-quantum-hall-and-topology":23523,"\u002Fcondensed-matter\u002Fnanostructures\u002Fgraphene-and-dirac-materials":24088,"\u002Fcondensed-matter":23840,"\u002Flogic\u002Ffoundations\u002Flogic-as-a-mathematical-model":24089,"\u002Flogic\u002Fsentential-logic\u002Fformal-languages-and-well-formed-formulas":24090,"\u002Flogic\u002Fsentential-logic\u002Ftruth-assignments-and-tautologies":24091,"\u002Flogic\u002Fsentential-logic\u002Funique-readability-and-parsing":24092,"\u002Flogic\u002Fsentential-logic\u002Finduction-and-recursion":23537,"\u002Flogic\u002Fsentential-logic\u002Fexpressive-completeness-and-normal-forms":24093,"\u002Flogic\u002Fsentential-logic\u002Fboolean-circuits":24094,"\u002Flogic\u002Fsentential-logic\u002Fcompactness-and-effectiveness":23537,"\u002Flogic\u002Ffirst-order-languages\u002Ffirst-order-languages":24095,"\u002Flogic\u002Ffirst-order-languages\u002Fstructures-truth-and-satisfaction":23950,"\u002Flogic\u002Ffirst-order-languages\u002Fdefinability-and-elementary-equivalence":24096,"\u002Flogic\u002Ffirst-order-languages\u002Fterms-substitution-and-parsing":24097,"\u002Flogic\u002Fdeductive-calculus\u002Fa-deductive-calculus":24098,"\u002Flogic\u002Fdeductive-calculus\u002Fdeduction-theorem-and-derived-rules":24096,"\u002Flogic\u002Fdeductive-calculus\u002Fsoundness":24099,"\u002Flogic\u002Fdeductive-calculus\u002Fcompleteness-and-consistency":24100,"\u002Flogic\u002Fmodels-and-theories\u002Fcompactness-and-lowenheim-skolem":24101,"\u002Flogic\u002Fmodels-and-theories\u002Ftheories-elementary-classes-and-categoricity":24102,"\u002Flogic\u002Fmodels-and-theories\u002Finterpretations-between-theories":24103,"\u002Flogic\u002Fmodels-and-theories\u002Fnonstandard-analysis":24104,"\u002Flogic\u002Farithmetic-and-definability\u002Fdefinability-in-arithmetic":24105,"\u002Flogic\u002Farithmetic-and-definability\u002Fnatural-numbers-with-successor":24106,"\u002Flogic\u002Farithmetic-and-definability\u002Fpresburger-and-reducts":24027,"\u002Flogic\u002Farithmetic-and-definability\u002Fa-subtheory-and-representability":24107,"\u002Flogic\u002Fincompleteness\u002Farithmetization-of-syntax":24100,"\u002Flogic\u002Fincompleteness\u002Fincompleteness-and-undecidability":24108,"\u002Flogic\u002Fincompleteness\u002Fsecond-incompleteness-theorem":24109,"\u002Flogic\u002Fcomputability-and-representability\u002Frecursive-functions":23743,"\u002Flogic\u002Fcomputability-and-representability\u002Frepresenting-exponentiation":24110,"\u002Flogic\u002Fsecond-order-logic\u002Fsecond-order-languages":23928,"\u002Flogic\u002Fsecond-order-logic\u002Fskolem-functions-and-many-sorted-logic":24111,"\u002Flogic\u002Fsecond-order-logic\u002Fgeneral-structures":24112,"\u002Flogic":24113,"\u002Freinforcement-learning\u002Ffoundations\u002Fwhat-is-reinforcement-learning":24114,"\u002Freinforcement-learning\u002Ffoundations\u002Fa-brief-history-of-rl":24115,"\u002Freinforcement-learning\u002Ffoundations\u002Fmulti-armed-bandits":23729,"\u002Freinforcement-learning\u002Ffoundations\u002Fbandit-exploration-algorithms":24116,"\u002Freinforcement-learning\u002Ffoundations\u002Fmarkov-decision-processes":24117,"\u002Freinforcement-learning\u002Ffoundations\u002Fvalue-functions-and-optimality":24118,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fdynamic-programming":24119,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fdp-async-and-gpi":24109,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fmonte-carlo-methods":24120,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fmonte-carlo-off-policy":24121,"\u002Freinforcement-learning\u002Ftabular-methods\u002Ftemporal-difference-learning":24122,"\u002Freinforcement-learning\u002Ftabular-methods\u002Ftd-control-sarsa-and-q-learning":24020,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fn-step-bootstrapping":24123,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fn-step-off-policy-methods":24124,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fplanning-and-learning":24125,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fplanning-focusing-and-decision-time":24126,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fdecision-time-planning":24127,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fmonte-carlo-tree-search":23746,"\u002Freinforcement-learning\u002Fapproximation\u002Fon-policy-prediction":24128,"\u002Freinforcement-learning\u002Fapproximation\u002Ffeature-construction-and-nonlinear":24129,"\u002Freinforcement-learning\u002Fapproximation\u002Fon-policy-control":24130,"\u002Freinforcement-learning\u002Fapproximation\u002Faverage-reward-control":24131,"\u002Freinforcement-learning\u002Fapproximation\u002Foff-policy-and-the-deadly-triad":23937,"\u002Freinforcement-learning\u002Fapproximation\u002Fbellman-error-and-gradient-td":24132,"\u002Freinforcement-learning\u002Fapproximation\u002Feligibility-traces":24108,"\u002Freinforcement-learning\u002Fapproximation\u002Ftrue-online-and-sarsa-lambda":24133,"\u002Freinforcement-learning\u002Fapproximation\u002Fpolicy-gradient-methods":24134,"\u002Freinforcement-learning\u002Fapproximation\u002Factor-critic-and-continuous-actions":24135,"\u002Freinforcement-learning\u002Fapproximation\u002Fleast-squares-and-memory-based-methods":23757,"\u002Freinforcement-learning\u002Fapproximation\u002Fmemory-and-kernel-methods":24136,"\u002Freinforcement-learning\u002Fapproximation\u002Foff-policy-eligibility-traces":23887,"\u002Freinforcement-learning\u002Fapproximation\u002Fstable-off-policy-traces":24137,"\u002Freinforcement-learning\u002Fdeep-rl\u002Fdeep-q-networks":24138,"\u002Freinforcement-learning\u002Fdeep-rl\u002Fdqn-improvements":23719,"\u002Freinforcement-learning\u002Fdeep-rl\u002Factor-critic-and-ppo":24139,"\u002Freinforcement-learning\u002Fdeep-rl\u002Fppo-and-continuous-control":24140,"\u002Freinforcement-learning\u002Fdeep-rl\u002Fcase-studies":24141,"\u002Freinforcement-learning\u002Fdeep-rl\u002Frl-beyond-games":24142,"\u002Freinforcement-learning\u002Fdeep-rl\u002Ffrontiers":24143,"\u002Freinforcement-learning\u002Fdeep-rl\u002Freward-design-and-open-problems":23996,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fdistributional-and-rainbow":24144,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fdistributional-and-rainbow-part-2":24145,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fcontinuous-control":24146,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fcontinuous-control-part-2":24031,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmodel-based-rl":24147,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmodel-based-rl-part-2":24148,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fexploration":24149,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fexploration-part-2":23746,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Foffline-rl":23813,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Foffline-rl-part-2":24150,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fimitation-and-inverse-rl":24151,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fimitation-and-inverse-rl-part-2":24152,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmulti-agent-rl":24153,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmulti-agent-rl-part-2":24154,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fhierarchical-rl":24155,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fhierarchical-rl-part-2":24156,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Frlhf-and-language-models":24157,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fpartial-observability-pomdps":24158,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fpartial-observability-pomdps-part-2":24159,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fsafe-and-constrained-rl":24160,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fsafe-and-constrained-rl-part-2":24161,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmeta-rl-and-generalization":24162,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fpsychology-of-reinforcement":24163,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Finstrumental-conditioning-and-control":24164,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fdopamine-and-td-error":24165,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fdopamine-in-the-brain":24166,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fanimal-learning-and-cognition":24167,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fcognitive-maps-and-planning":24168,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fneuroscience-of-reinforcement":24169,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fseveral-learning-systems":24170,"\u002Freinforcement-learning":23699,"\u002Fartificial-intelligence\u002Ffoundations\u002Fwhat-is-ai":24171,"\u002Fartificial-intelligence\u002Ffoundations\u002Ffoundations-of-ai":24172,"\u002Fartificial-intelligence\u002Ffoundations\u002Fintelligent-agents":24173,"\u002Fartificial-intelligence\u002Ffoundations\u002Fagent-architectures":24174,"\u002Fartificial-intelligence\u002Fsearch\u002Funinformed-search":24175,"\u002Fartificial-intelligence\u002Fsearch\u002Fsearch-strategies-compared":24176,"\u002Fartificial-intelligence\u002Fsearch\u002Finformed-search":24177,"\u002Fartificial-intelligence\u002Fsearch\u002Fheuristic-functions":24178,"\u002Fartificial-intelligence\u002Fsearch\u002Flocal-search":24179,"\u002Fartificial-intelligence\u002Fsearch\u002Fpopulation-and-continuous-search":24180,"\u002Fartificial-intelligence\u002Fsearch\u002Fadversarial-search":24181,"\u002Fartificial-intelligence\u002Fsearch\u002Fgames-of-chance-and-imperfect-information":24182,"\u002Fartificial-intelligence\u002Fsearch\u002Fconstraint-satisfaction":24183,"\u002Fartificial-intelligence\u002Fsearch\u002Fcsp-search-and-structure":24025,"\u002Fartificial-intelligence\u002Fsearch\u002Fsearch-under-uncertainty":23881,"\u002Fartificial-intelligence\u002Fsearch\u002Fbelief-state-and-online-search":24184,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fpropositional-logic":24185,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fpropositional-inference":24186,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Ffirst-order-logic":24187,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Ffirst-order-logic-in-use":24188,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Finference-and-resolution":24189,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Ffirst-order-resolution":24007,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fclassical-planning":24190,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fplanning-graphs-and-graphplan":24191,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fplanning-in-the-real-world":24192,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fplanning-under-uncertainty":24193,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fknowledge-representation":24194,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Freasoning-systems-and-defaults":24195,"\u002Fartificial-intelligence\u002Funcertainty\u002Fprobability-and-bayes":24196,"\u002Fartificial-intelligence\u002Funcertainty\u002Fbayes-rule-and-naive-bayes":24197,"\u002Fartificial-intelligence\u002Funcertainty\u002Fbayesian-networks":24198,"\u002Fartificial-intelligence\u002Funcertainty\u002Finference-in-bayesian-networks":24199,"\u002Fartificial-intelligence\u002Funcertainty\u002Freasoning-over-time":24200,"\u002Fartificial-intelligence\u002Funcertainty\u002Ftracking-and-data-association":24201,"\u002Fartificial-intelligence\u002Funcertainty\u002Fmaking-decisions":23940,"\u002Fartificial-intelligence\u002Funcertainty\u002Fmarkov-decision-processes":24191,"\u002Fartificial-intelligence\u002Funcertainty\u002Fdecision-networks-and-game-theory":24202,"\u002Fartificial-intelligence\u002Funcertainty\u002Fgame-theory-and-mechanism-design":23480,"\u002Fartificial-intelligence\u002Flearning\u002Flearning-from-examples":24203,"\u002Fartificial-intelligence\u002Flearning\u002Ftheory-and-model-families":24204,"\u002Fartificial-intelligence\u002Flearning\u002Fprobabilistic-learning":24205,"\u002Fartificial-intelligence\u002Flearning\u002Fexpectation-maximization":24206,"\u002Fartificial-intelligence\u002Flearning\u002Freinforcement-learning":24207,"\u002Fartificial-intelligence\u002Flearning\u002Fgeneralization-and-policy-search":23412,"\u002Fartificial-intelligence\u002Flearning\u002Fknowledge-in-learning":24208,"\u002Fartificial-intelligence\u002Flearning\u002Fknowledge-based-learning-methods":24209,"\u002Fartificial-intelligence\u002Ffrontiers\u002Fvision-and-perception":24210,"\u002Fartificial-intelligence\u002Ffrontiers\u002Freconstructing-the-3d-world":24211,"\u002Fartificial-intelligence\u002Ffrontiers\u002Frobotics":24212,"\u002Fartificial-intelligence\u002Ffrontiers\u002Frobot-planning-and-control":24213,"\u002Fartificial-intelligence\u002Ffrontiers\u002Fnatural-language-in-ai":24214,"\u002Fartificial-intelligence\u002Ffrontiers\u002Fnlp-grammar-translation-and-speech":24215,"\u002Fartificial-intelligence\u002Ffrontiers\u002Fphilosophy-and-future":24216,"\u002Fartificial-intelligence\u002Ffrontiers\u002Fai-ethics-and-future":24217,"\u002Fartificial-intelligence":23699,"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fnuclear-constituents-nuclide-chart":23956,"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fnuclear-size-charge-distributions":24218,"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fnuclear-masses-binding-energy":23516,"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fsemi-empirical-mass-formula":23514,"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fnuclear-moments-multipoles":24049,"\u002Fnuclear-physics\u002Fnuclear-force-deuteron\u002Fnuclear-force-shell-overview":24219,"\u002Fnuclear-physics\u002Fnuclear-force-deuteron\u002Fthe-deuteron":23542,"\u002Fnuclear-physics\u002Fnuclear-force-deuteron\u002Fnucleon-nucleon-scattering":23852,"\u002Fnuclear-physics\u002Fnuclear-force-deuteron\u002Fmeson-theory-isospin":24220,"\u002Fnuclear-physics\u002Fnuclear-models\u002Ffermi-gas-model":24221,"\u002Fnuclear-physics\u002Fnuclear-models\u002Fliquid-drop-collective-coordinates":24222,"\u002Fnuclear-physics\u002Fnuclear-models\u002Fshell-model-single-particle":23912,"\u002Fnuclear-physics\u002Fnuclear-models\u002Fcollective-model-rotations-vibrations":24223,"\u002Fnuclear-physics\u002Fradioactive-decay\u002Fdecay-law-modes":24224,"\u002Fnuclear-physics\u002Fradioactive-decay\u002Fdecay-kinetics-equilibrium":24225,"\u002Fnuclear-physics\u002Falpha-decay\u002Falpha-decay-gamow-theory":24110,"\u002Fnuclear-physics\u002Falpha-decay\u002Falpha-fine-structure-hindrance":24226,"\u002Fnuclear-physics\u002Fbeta-decay\u002Fbeta-decay-energetics-neutrino":24227,"\u002Fnuclear-physics\u002Fbeta-decay\u002Ffermi-theory-beta-decay":23526,"\u002Fnuclear-physics\u002Fbeta-decay\u002Fweak-interaction-parity-violation":23755,"\u002Fnuclear-physics\u002Fbeta-decay\u002Fdouble-beta-decay-neutrino-mass":24228,"\u002Fnuclear-physics\u002Fgamma-decay\u002Fgamma-multipole-radiation":23856,"\u002Fnuclear-physics\u002Fgamma-decay\u002Finternal-conversion-isomers":24229,"\u002Fnuclear-physics\u002Fgamma-decay\u002Fangular-correlations-mossbauer":24230,"\u002Fnuclear-physics\u002Fnuclear-reactions\u002Freaction-kinematics-cross-sections":23751,"\u002Fnuclear-physics\u002Fnuclear-reactions\u002Fcompound-nucleus-resonances":23839,"\u002Fnuclear-physics\u002Fnuclear-reactions\u002Fdirect-reactions-optical-model":24231,"\u002Fnuclear-physics\u002Ffission\u002Ffission-barrier-dynamics":24232,"\u002Fnuclear-physics\u002Ffission\u002Fchain-reactions-reactor-physics":24233,"\u002Fnuclear-physics\u002Ffusion-nucleosynthesis\u002Ffusion-reactions-confinement":23521,"\u002Fnuclear-physics\u002Ffusion-nucleosynthesis\u002Fstellar-nucleosynthesis":23896,"\u002Fnuclear-physics\u002Ffusion-nucleosynthesis\u002Fbig-bang-nucleosynthesis":23758,"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fcharged-particle-stopping-power":24234,"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fphoton-neutron-interactions":23806,"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fradiation-detectors":23870,"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fdosimetry-radiation-biology":24235,"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fnuclear-applications-dating-medicine":24236,"\u002Fnuclear-physics":24237,"\u002Fnatural-language-processing\u002Ffoundations\u002Fwhat-is-nlp":24238,"\u002Fnatural-language-processing\u002Ffoundations\u002Fregex-and-text-normalization":24239,"\u002Fnatural-language-processing\u002Ffoundations\u002Fminimum-edit-distance":23877,"\u002Fnatural-language-processing\u002Ffoundations\u002Fn-gram-language-models":24240,"\u002Fnatural-language-processing\u002Ffoundations\u002Fsmoothing-and-backoff":24241,"\u002Fnatural-language-processing\u002Fclassification\u002Fnaive-bayes-and-sentiment":24242,"\u002Fnatural-language-processing\u002Fclassification\u002Fevaluating-classifiers":23725,"\u002Fnatural-language-processing\u002Fclassification\u002Flogistic-regression":24243,"\u002Fnatural-language-processing\u002Fclassification\u002Fsentiment-and-affect-lexicons":24244,"\u002Fnatural-language-processing\u002Fsemantics\u002Fvector-semantics-and-embeddings":24021,"\u002Fnatural-language-processing\u002Fsemantics\u002Fstatic-word-embeddings":24245,"\u002Fnatural-language-processing\u002Fsemantics\u002Fneural-language-models":24190,"\u002Fnatural-language-processing\u002Fsequences\u002Fsequence-labeling":24246,"\u002Fnatural-language-processing\u002Fsequences\u002Fcrfs-and-neural-taggers":24247,"\u002Fnatural-language-processing\u002Fsequences\u002Frnns-and-lstms":24248,"\u002Fnatural-language-processing\u002Ftransformers\u002Ftransformers-and-attention":24249,"\u002Fnatural-language-processing\u002Ftransformers\u002Fthe-transformer-architecture":24250,"\u002Fnatural-language-processing\u002Ftransformers\u002Flarge-language-models":24251,"\u002Fnatural-language-processing\u002Ftransformers\u002Fllm-pretraining-and-scaling":24252,"\u002Fnatural-language-processing\u002Ftransformers\u002Ffine-tuning-and-prompting":23703,"\u002Fnatural-language-processing\u002Ftransformers\u002Fprompting-and-alignment":24253,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fconstituency-parsing":24254,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fcky-scoring-and-evaluation":24196,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fdependency-parsing":24255,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fgraph-based-and-neural-dependency-parsing":24256,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fword-senses-and-wsd":24257,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fwsd-in-practice-and-induction":24258,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fsemantic-roles-and-information-extraction":24259,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Frelations-events-and-templates":24260,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fcoreference-and-discourse":24261,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fcoherence-and-discourse-structure":24262,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Flogical-semantics":24120,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fcompositional-semantics-and-description-logics":24263,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fsemantic-parsing":24264,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fneural-semantic-parsing":24265,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Finformation-extraction":24266,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Ftimes-events-and-templates":24267,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fdiscourse-coherence":24268,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fentity-based-and-global-coherence":24269,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fconstituency-grammars":24270,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Ftreebanks-and-lexicalized-grammars":24271,"\u002Fnatural-language-processing\u002Fapplications\u002Fmachine-translation":24272,"\u002Fnatural-language-processing\u002Fapplications\u002Fmachine-translation-decoding-and-evaluation":24273,"\u002Fnatural-language-processing\u002Fapplications\u002Fquestion-answering":24274,"\u002Fnatural-language-processing\u002Fapplications\u002Fquestion-answering-knowledge-and-llms":23975,"\u002Fnatural-language-processing\u002Fapplications\u002Fdialogue-and-chatbots":24141,"\u002Fnatural-language-processing\u002Fapplications\u002Fdialogue-systems-and-assistants":23731,"\u002Fnatural-language-processing\u002Fapplications\u002Ftext-summarization":24275,"\u002Fnatural-language-processing\u002Fapplications\u002Fabstractive-summarization-and-evaluation":24276,"\u002Fnatural-language-processing\u002Fspeech\u002Fphonetics":24277,"\u002Fnatural-language-processing\u002Fspeech\u002Facoustic-phonetics":24278,"\u002Fnatural-language-processing\u002Fspeech\u002Fautomatic-speech-recognition":23412,"\u002Fnatural-language-processing\u002Fspeech\u002Fasr-evaluation-and-applications":24279,"\u002Fnatural-language-processing":23699,"\u002Fparticle-physics\u002Ffoundations\u002Fhistorical-overview-particle-zoo":24280,"\u002Fparticle-physics\u002Ffoundations\u002Fparticle-physics-basic-concepts":23897,"\u002Fparticle-physics\u002Ffoundations\u002Ffundamental-interactions-force-carriers":24281,"\u002Fparticle-physics\u002Funits-kinematics\u002Fnatural-units-and-scales":24041,"\u002Fparticle-physics\u002Funits-kinematics\u002Ffour-vectors-invariant-mass":24282,"\u002Fparticle-physics\u002Funits-kinematics\u002Fdecay-scattering-kinematics-mandelstam":24283,"\u002Fparticle-physics\u002Funits-kinematics\u002Fcross-sections-golden-rule":24231,"\u002Fparticle-physics\u002Fsymmetries\u002Fconservation-laws-symmetries":24284,"\u002Fparticle-physics\u002Fsymmetries\u002Fdiscrete-symmetries-cpt":24285,"\u002Fparticle-physics\u002Fsymmetries\u002Fparity-violation-weak":23847,"\u002Fparticle-physics\u002Fsymmetries\u002Fsu2-su3-flavor-symmetry":23549,"\u002Fparticle-physics\u002Fquark-model\u002Feightfold-way-su3":24286,"\u002Fparticle-physics\u002Fquark-model\u002Fmeson-spectroscopy":23801,"\u002Fparticle-physics\u002Fquark-model\u002Fbaryon-spectroscopy":24287,"\u002Fparticle-physics\u002Fquark-model\u002Fcolor-confinement-exotics":23834,"\u002Fparticle-physics\u002Frelativistic-wave-equations\u002Fklein-gordon-equation":24288,"\u002Fparticle-physics\u002Frelativistic-wave-equations\u002Fdirac-equation-spinors":24087,"\u002Fparticle-physics\u002Frelativistic-wave-equations\u002Fantiparticles-hole-theory":24289,"\u002Fparticle-physics\u002Fqed\u002Ffeynman-rules-qed":24290,"\u002Fparticle-physics\u002Fqed\u002Fqed-tree-processes":24056,"\u002Fparticle-physics\u002Fqed\u002Frenormalization-running-coupling":24291,"\u002Fparticle-physics\u002Fqed\u002Felectron-g-2":23520,"\u002Fparticle-physics\u002Fweak-interaction\u002Fva-structure-weak":24292,"\u002Fparticle-physics\u002Fweak-interaction\u002Fw-z-bosons-decays":24293,"\u002Fparticle-physics\u002Fweak-interaction\u002Fckm-matrix":24294,"\u002Fparticle-physics\u002Fweak-interaction\u002Fcp-violation-kaons-b-mesons":23838,"\u002Fparticle-physics\u002Fqcd\u002Fcolor-su3-gluons":24053,"\u002Fparticle-physics\u002Fqcd\u002Fasymptotic-freedom-confinement":24295,"\u002Fparticle-physics\u002Fqcd\u002Fdeep-inelastic-scattering-partons":24296,"\u002Fparticle-physics\u002Fqcd\u002Fjets-hadronization":23913,"\u002Fparticle-physics\u002Felectroweak-higgs\u002Felectroweak-su2-u1":24297,"\u002Fparticle-physics\u002Felectroweak-higgs\u002Fspontaneous-symmetry-breaking":23958,"\u002Fparticle-physics\u002Felectroweak-higgs\u002Fhiggs-mechanism":24298,"\u002Fparticle-physics\u002Felectroweak-higgs\u002Fhiggs-boson-discovery":24299,"\u002Fparticle-physics\u002Felectroweak-higgs\u002Fstandard-model":23847,"\u002Fparticle-physics\u002Fneutrinos\u002Fneutrino-oscillations":23880,"\u002Fparticle-physics\u002Fneutrinos\u002Fneutrino-mass-pmns":24300,"\u002Fparticle-physics\u002Fneutrinos\u002Fdirac-majorana-experiments":24301,"\u002Fparticle-physics\u002Fexperiment\u002Faccelerators-luminosity":24302,"\u002Fparticle-physics\u002Fexperiment\u002Fdetectors-subsystems":23936,"\u002Fparticle-physics\u002Fexperiment\u002Fhow-discoveries-are-made":24303,"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fbeyond-standard-model":23677,"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fgrand-unified-theories":24304,"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fsupersymmetry":23780,"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fhierarchy-problem-naturalness":24305,"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fdark-matter-candidates":24306,"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fmatter-antimatter-open-questions":24103,"\u002Fparticle-physics":24307,"\u002Fastrophysics-cosmology\u002Forientation\u002Fthe-sun-and-stars":23898,"\u002Fastrophysics-cosmology\u002Forientation\u002Fstellar-death-final-states":24051,"\u002Fastrophysics-cosmology\u002Forientation\u002Fgalaxies-and-cosmology":24308,"\u002Fastrophysics-cosmology\u002Fobservational-foundations\u002Fmagnitudes-fluxes-and-the-distance-modulus":23837,"\u002Fastrophysics-cosmology\u002Fobservational-foundations\u002Fstellar-spectra-and-spectral-classification":24309,"\u002Fastrophysics-cosmology\u002Fobservational-foundations\u002Ftelescopes-and-detectors-across-the-spectrum":23897,"\u002Fastrophysics-cosmology\u002Fobservational-foundations\u002Fthe-cosmic-distance-ladder":23809,"\u002Fastrophysics-cosmology\u002Fradiation-and-matter\u002Fblackbody-radiation-and-specific-intensity":24310,"\u002Fastrophysics-cosmology\u002Fradiation-and-matter\u002Fradiative-transfer-and-the-transfer-equation":24311,"\u002Fastrophysics-cosmology\u002Fradiation-and-matter\u002Fspectral-line-formation-and-broadening":24312,"\u002Fastrophysics-cosmology\u002Fradiation-and-matter\u002Fopacity-and-the-rosseland-mean":24313,"\u002Fastrophysics-cosmology\u002Fstellar-structure\u002Fhydrostatic-equilibrium-and-the-virial-theorem":24314,"\u002Fastrophysics-cosmology\u002Fstellar-structure\u002Fthe-equations-of-stellar-structure":24102,"\u002Fastrophysics-cosmology\u002Fstellar-structure\u002Fthe-equation-of-state-and-polytropes":24033,"\u002Fastrophysics-cosmology\u002Fstellar-structure\u002Fthe-standard-solar-model":24315,"\u002Fastrophysics-cosmology\u002Fnuclear-astrophysics\u002Fthermonuclear-reaction-rates-and-the-gamow-peak":24316,"\u002Fastrophysics-cosmology\u002Fnuclear-astrophysics\u002Fhydrogen-burning-pp-chains-and-cno":24317,"\u002Fastrophysics-cosmology\u002Fnuclear-astrophysics\u002Fhelium-burning-and-the-triple-alpha-process":24318,"\u002Fastrophysics-cosmology\u002Fnuclear-astrophysics\u002Fadvanced-burning-and-neutron-capture-nucleosynthesis":24230,"\u002Fastrophysics-cosmology\u002Fism-and-star-formation\u002Fphases-of-the-interstellar-medium":24319,"\u002Fastrophysics-cosmology\u002Fism-and-star-formation\u002Fmolecular-clouds-and-gravitational-collapse":23793,"\u002Fastrophysics-cosmology\u002Fism-and-star-formation\u002Fprotostars-and-the-pre-main-sequence":23856,"\u002Fastrophysics-cosmology\u002Fstellar-evolution\u002Fthe-main-sequence-and-its-structure":23838,"\u002Fastrophysics-cosmology\u002Fstellar-evolution\u002Fpost-main-sequence-low-mass-evolution":24320,"\u002Fastrophysics-cosmology\u002Fstellar-evolution\u002Fthe-evolution-of-massive-stars":23922,"\u002Fastrophysics-cosmology\u002Fstellar-evolution\u002Fstellar-pulsation-and-the-instability-strip":23538,"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fwhite-dwarfs-and-the-chandrasekhar-limit":24321,"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fcore-collapse-supernovae":24322,"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fthermonuclear-supernovae-type-ia":24052,"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fneutron-stars-and-pulsars":24323,"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fblack-holes-schwarzschild-and-kerr":24324,"\u002Fastrophysics-cosmology\u002Fbinaries-and-gravitational-waves\u002Fbinary-systems-and-mass-transfer":24325,"\u002Fastrophysics-cosmology\u002Fbinaries-and-gravitational-waves\u002Faccreting-compact-objects":24326,"\u002Fastrophysics-cosmology\u002Fbinaries-and-gravitational-waves\u002Fgravitational-waves-from-inspiraling-binaries":23519,"\u002Fastrophysics-cosmology\u002Fbinaries-and-gravitational-waves\u002Fmultimessenger-astronomy-and-gamma-ray-bursts":24327,"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Fthe-milky-way":24328,"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Fgalaxy-morphology-and-classification":23780,"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Fgalaxy-rotation-curves-and-dark-matter":24329,"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Factive-galactic-nuclei-and-supermassive-black-holes":24330,"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Fgalaxy-clusters-and-large-scale-structure":24045,"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fthe-expanding-universe-and-hubbles-law":23701,"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fthe-frw-metric-and-cosmological-redshift":24331,"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fthe-friedmann-equations-and-cosmic-dynamics":24332,"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fcosmological-models-and-distances":23958,"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fdark-energy-and-the-accelerating-universe":23944,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fthe-thermal-history-of-the-universe":23956,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fbig-bang-nucleosynthesis":24052,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Frecombination-and-the-cosmic-microwave-background":24333,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fcmb-anisotropies-and-cosmological-parameters":23518,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fcosmic-inflation":23710,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fstructure-formation-and-the-growth-of-perturbations":24334,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fdark-matter-dark-energy-and-open-questions":23797,"\u002Fastrophysics-cosmology":23902,"\u002Fcolophon":24335,"\u002F":23699},4250,4808,3626,2682,4109,4786,3878,3875,3751,3415,4067,3153,3000,4042,5461,5808,3961,3749,4327,5067,4246,4655,4154,5436,2640,4003,3601,2158,4331,4189,2273,3252,4633,4964,4172,3131,5524,3160,4031,2309,4207,3226,2648,4842,5340,3307,5701,4977,4039,2615,3472,4460,3848,4075,4400,3382,3010,3602,3737,3740,3707,3922,5191,4043,3804,4542,4214,5062,2850,4361,3443,3627,4044,3766,4140,3860,4006,5199,4334,5234,3651,5509,5680,153,1375,1073,1093,1125,1146,1014,1132,876,1541,1189,1173,984,1402,1301,950,1268,1063,1107,1408,1161,925,1012,866,964,1090,1142,1085,1020,1207,973,980,728,764,1225,1329,796,929,801,878,774,1044,1488,1175,1130,890,814,870,154,4073,5140,4961,5127,4870,5382,5195,4955,5369,4501,5576,3824,4132,4289,4307,4570,3403,5084,5105,5201,5116,5341,5175,5368,5188,5211,5499,5155,4981,5125,5415,5255,5304,5130,5167,5552,5164,5094,5239,5036,5190,5004,5099,5035,5159,5088,5026,4937,5023,5264,5244,133,5114,5078,5043,5312,5170,5342,5139,5151,5049,5212,5013,5068,5079,5102,5121,5081,5029,5379,5854,5110,2139,3798,5055,5364,4984,4935,4895,4972,5289,5112,5156,4987,5031,5025,5149,5302,5042,5002,4979,4922,4960,5279,126,1877,1180,1129,907,958,1112,1300,1053,1250,1181,1241,1234,966,1050,734,1190,484,1082,926,733,761,571,607,798,804,952,977,731,784,645,771,1017,742,1004,1000,1562,1254,1288,1101,1011,1486,1061,856,992,1169,988,137,0,2037,1782,2384,2254,2123,2332,1643,1714,2089,1751,1367,1660,2511,1998,1892,1854,1791,2438,2487,1917,2375,2525,2266,1845,2275,1810,1631,2310,2166,2233,2113,2505,2347,2672,2112,2473,2592,2380,3013,2513,3256,3218,2194,2173,2205,2326,2081,3342,3152,1799,1670,1027,960,1095,1291,986,897,1209,1055,1817,1801,1593,1465,1196,1464,1201,1230,1435,1684,1461,1926,1500,1409,1284,1774,1869,162,1487,1122,1188,1001,1351,982,1005,979,1325,1046,943,1279,824,1008,989,1798,1277,1025,987,1043,1211,1074,981,939,1002,739,1139,1108,1013,1070,978,1458,1317,157,1357,1077,2355,1116,1037,1178,1637,1314,1109,1056,1702,1474,1071,1158,832,993,1404,1024,1068,1339,1106,1264,1248,913,1848,1328,1633,1224,1143,135,1378,959,1028,998,911,1527,1203,1266,1483,1165,990,938,965,1257,1418,1099,942,1352,956,1035,1398,1003,1094,1292,138,1721,1827,1449,1354,1148,1184,1285,1281,1213,1290,1271,1252,1274,1778,1591,1503,1437,1571,1584,1957,1117,1781,1648,1342,1667,1510,1965,1607,1365,1849,1259,1303,1356,1238,2208,1564,173,1671,1286,1227,1638,1529,668,1078,918,709,865,880,940,1534,1015,874,922,841,794,1194,822,1105,1658,1359,1296,1438,1921,1844,1570,1429,1324,1400,140,1787,1558,1654,1492,1747,2224,2002,2009,1323,1349,1785,1573,1722,1829,1353,1548,1552,1583,1624,1585,1245,1364,1514,1343,1397,1355,2211,1481,1770,160,2388,2293,2256,2552,2569,2478,2039,2496,2578,2814,2519,2461,2587,2492,2714,3278,2654,3050,2447,2849,2238,2369,2061,2214,2563,2186,2985,2749,3364,2038,2282,2409,2126,2573,2206,2176,2268,2182,2402,2705,2633,2414,2213,2801,3313,3410,3195,1952,2017,1509,2537,2645,2027,2415,2838,2356,1906,3184,2950,2807,2954,1683,1316,1034,1138,1763,1822,1705,1246,1701,1097,1104,1187,1032,1083,1228,916,1489,1033,1652,997,692,837,1023,888,864,1089,1231,1214,1675,1156,1075,1520,1309,139,1205,1051,735,1123,1072,915,567,768,825,1253,983,1007,762,1058,861,862,971,1208,1149,1145,1029,1084,927,810,838,857,807,936,949,2321,1622,1069,1113,1057,854,1958,1528,1618,2049,1432,1679,1796,1685,1346,1275,1476,1505,1610,2018,1599,1215,1838,1909,132,3902,2215,2240,3266,3208,3073,2454,2969,2451,1875,2728,1884,2371,2516,2842,1690,1904,2346,3146,1386,2607,1966,2668,1665,2885,1606,2577,3074,2869,2403,2433,2082,1939,1587,2460,2747,2032,2642,1619,3123,1993,2090,2339,3829,1737,2622,2340,2322,3828,4409,2305,3411,2510,4527,3030,3569,3043,2457,1946,2277,2044,2909,1693,1945,2093,2399,2115,2898,2742,2242,3895,3378,3376,2769,2223,3062,3262,2651,2949,2768,3128,2423,1977,2087,2866,3388,2830,2210,2489,2884,3945,2099,2713,3402,1692,2931,4195,3989,3206,4391,3004,3704,3494,2902,999,881,901,919,748,869,1018,1045,1049,1333,954,1092,1019,976,1771,1480,1396,953,1026,161,3533,2495,1818,3007,2595,3427,3537,2216,1895,2304,3396,1739,2073,1962,2203,1767,2666,2264,2276,2852,1807,3735,1560,4144,1669,1676,1972,2418,3291,1525,2040,2766,2337,2220,2800,3001,2078,1759,2836,1896,2026,1758,1543,1047,896,946,1060,1384,1482,815,1414,1322,1440,1240,1468,1098,1133,847,1009,1381,1052,1191,1258,1370,1712,1441,1199,957,1079,150,1262,1417,1368,1219,1136,1064,1463,1636,1059,931,1115,1736,1174,1376,1363,1411,1247,1746,1313,1299,1617,1102,1076,1495,1265,1193,1263,80,{"\u002Falgorithms\u002Ffoundations\u002Fwhat-is-an-algorithm":24337,"\u002Falgorithms\u002Ffoundations\u002Fproof-techniques":24342,"\u002Falgorithms\u002Ffoundations\u002Fasymptotic-analysis":24346,"\u002Falgorithms\u002Ffoundations\u002Fgrowth-rates-and-loop-analysis":24350,"\u002Falgorithms\u002Ffoundations\u002Frecurrences":24354,"\u002Falgorithms\u002Ffoundations\u002Famortized-analysis":24358,"\u002Falgorithms\u002Fdivide-and-conquer\u002Fmergesort":24362,"\u002Falgorithms\u002Fdivide-and-conquer\u002Fquicksort":24367,"\u002Falgorithms\u002Fdivide-and-conquer\u002Fselection":24371,"\u002Falgorithms\u002Fdivide-and-conquer\u002Ffast-multiplication":24375,"\u002Falgorithms\u002Fsorting\u002Fheaps-and-heapsort":24379,"\u002Falgorithms\u002Fsorting\u002Fsorting-lower-bounds":24384,"\u002Falgorithms\u002Fsorting\u002Flinear-time-sorting":24388,"\u002Falgorithms\u002Fsorting\u002Fexternal-sorting":24392,"\u002Falgorithms\u002Fdata-structures\u002Felementary-structures":24396,"\u002Falgorithms\u002Fdata-structures\u002Fhash-tables":24401,"\u002Falgorithms\u002Fdata-structures\u002Fbinary-search-trees":24405,"\u002Falgorithms\u002Fdata-structures\u002Favl-trees":24409,"\u002Falgorithms\u002Fdata-structures\u002Fbalanced-trees":24413,"\u002Falgorithms\u002Fdata-structures\u002Funion-find":24417,"\u002Falgorithms\u002Fdata-structures\u002Ffenwick-and-segment-trees":24421,"\u002Falgorithms\u002Fdata-structures\u002Fspatial-data-structures":24425,"\u002Falgorithms\u002Fdata-structures\u002Fskip-lists-and-probabilistic-structures":24429,"\u002Falgorithms\u002Fdata-structures\u002Fb-trees":24433,"\u002Falgorithms\u002Fdata-structures\u002Fdata-stream-algorithms":24437,"\u002Falgorithms\u002Fdata-structures\u002Fstreaming-sketches":24441,"\u002Falgorithms\u002Fsequences\u002Ftwo-pointers-and-windows":24445,"\u002Falgorithms\u002Fsequences\u002Fprefix-sums":24450,"\u002Falgorithms\u002Fsequences\u002Fmonotonic-stacks":24454,"\u002Falgorithms\u002Fsequences\u002Fbinary-search-on-the-answer":24458,"\u002Falgorithms\u002Fsequences\u002Fstring-matching":24462,"\u002Falgorithms\u002Fsequences\u002Fkmp-and-z-function":24466,"\u002Falgorithms\u002Fsequences\u002Ftries":24470,"\u002Falgorithms\u002Fsequences\u002Fsuffix-arrays-and-aho-corasick":24474,"\u002Falgorithms\u002Fgraphs\u002Frepresentations-and-traversal":24478,"\u002Falgorithms\u002Fgraphs\u002Fdepth-first-search":24483,"\u002Falgorithms\u002Fgraphs\u002Ftopological-sort-and-scc":24487,"\u002Falgorithms\u002Fgraphs\u002Fminimum-spanning-trees":24491,"\u002Falgorithms\u002Fgraphs\u002Fkruskal-and-prim":24495,"\u002Falgorithms\u002Fgraphs\u002Fshortest-paths":24499,"\u002Falgorithms\u002Fgraphs\u002Fall-pairs-and-negative-weights":24503,"\u002Falgorithms\u002Fgraphs\u002Fnetwork-flow":24507,"\u002Falgorithms\u002Fgraphs\u002Fmax-flow-min-cut":24511,"\u002Falgorithms\u002Fgraphs\u002Fbridges-and-articulation-points":24515,"\u002Falgorithms\u002Fgraphs\u002Flowest-common-ancestor":24519,"\u002Falgorithms\u002Fgraphs\u002Ftwo-sat":24523,"\u002Falgorithms\u002Fgraphs\u002Feulerian-tours":24527,"\u002Falgorithms\u002Fgraphs\u002Fbipartite-matching":24531,"\u002Falgorithms\u002Fgreedy\u002Fthe-greedy-method":24535,"\u002Falgorithms\u002Fgreedy\u002Fscheduling-and-intervals":24540,"\u002Falgorithms\u002Fgreedy\u002Fhuffman-codes":24544,"\u002Falgorithms\u002Fgreedy\u002Fmatroids":24548,"\u002Falgorithms\u002Fgreedy\u002Fstable-matching":24552,"\u002Falgorithms\u002Fdynamic-programming\u002Fprinciples":24556,"\u002Falgorithms\u002Fdynamic-programming\u002Fsequence-dp":24561,"\u002Falgorithms\u002Fdynamic-programming\u002Flongest-increasing-subsequence":24565,"\u002Falgorithms\u002Fdynamic-programming\u002Fknapsack":24569,"\u002Falgorithms\u002Fdynamic-programming\u002Fcoin-change-and-unbounded":24573,"\u002Falgorithms\u002Fdynamic-programming\u002Finterval-dp":24577,"\u002Falgorithms\u002Fdynamic-programming\u002Ftree-dp":24581,"\u002Falgorithms\u002Fdynamic-programming\u002Fbitmask-dp":24585,"\u002Falgorithms\u002Fdynamic-programming\u002Fdp-optimizations":24589,"\u002Falgorithms\u002Fdynamic-programming\u002Fdp-on-graphs":24593,"\u002Falgorithms\u002Fdynamic-programming\u002Fdigit-and-probability-dp":24597,"\u002Falgorithms\u002Fbacktracking\u002Fbacktracking-fundamentals":24601,"\u002Falgorithms\u002Fbacktracking\u002Fconstraint-search":24606,"\u002Falgorithms\u002Fbacktracking\u002Fbranch-and-bound":24610,"\u002Falgorithms\u002Fbacktracking\u002Fgraph-backtracking":24614,"\u002Falgorithms\u002Fmathematical-algorithms\u002Fnumber-theory-basics":24618,"\u002Falgorithms\u002Fmathematical-algorithms\u002Fmodular-exponentiation-and-primality":24623,"\u002Falgorithms\u002Fmathematical-algorithms\u002Fsieve-and-factorization":24627,"\u002Falgorithms\u002Fmathematical-algorithms\u002Fcombinatorics":24631,"\u002Falgorithms\u002Fmathematical-algorithms\u002Fmatrix-exponentiation":24635,"\u002Falgorithms\u002Fmathematical-algorithms\u002Ffast-fourier-transform":24639,"\u002Falgorithms\u002Fmathematical-algorithms\u002Fgradient-descent":24643,"\u002Falgorithms\u002Fcomputational-geometry\u002Fgeometric-primitives":24647,"\u002Falgorithms\u002Fcomputational-geometry\u002Fconvex-hull":24652,"\u002Falgorithms\u002Fcomputational-geometry\u002Fsweep-line":24656,"\u002Falgorithms\u002Fcomputational-geometry\u002Fpolygons-and-proximity":24660,"\u002Falgorithms\u002Fintractability\u002Fp-np-reductions":24664,"\u002Falgorithms\u002Fintractability\u002Fnp-completeness":24669,"\u002Falgorithms\u002Fintractability\u002Fcoping-with-hardness":24673,"\u002Falgorithms\u002Fintractability\u002Fapproximation-algorithms":24677,"\u002Falgorithms":24681,"\u002Fcalculus\u002Flimits-and-continuity\u002Ffunctions-and-models":24684,"\u002Fcalculus\u002Flimits-and-continuity\u002Fthe-limit-of-a-function":24689,"\u002Fcalculus\u002Flimits-and-continuity\u002Flimit-laws-and-the-precise-definition":24693,"\u002Fcalculus\u002Flimits-and-continuity\u002Fcontinuity":24697,"\u002Fcalculus\u002Fderivatives\u002Fthe-derivative-and-rates-of-change":24701,"\u002Fcalculus\u002Fderivatives\u002Fdifferentiation-rules-and-the-chain-rule":24706,"\u002Fcalculus\u002Fderivatives\u002Fimplicit-differentiation-and-related-rates":24710,"\u002Fcalculus\u002Fderivatives\u002Flinear-approximations-and-differentials":24714,"\u002Fcalculus\u002Fapplications-of-derivatives\u002Fextrema-and-the-mean-value-theorem":24718,"\u002Fcalculus\u002Fapplications-of-derivatives\u002Fhow-derivatives-shape-a-graph":24723,"\u002Fcalculus\u002Fapplications-of-derivatives\u002Fcurve-sketching-and-optimization":24727,"\u002Fcalculus\u002Fapplications-of-derivatives\u002Fnewtons-method-and-antiderivatives":24731,"\u002Fcalculus\u002Fintegrals\u002Farea-and-the-definite-integral":24735,"\u002Fcalculus\u002Fintegrals\u002Fthe-fundamental-theorem-of-calculus":24740,"\u002Fcalculus\u002Fintegrals\u002Fthe-substitution-rule":24744,"\u002Fcalculus\u002Fapplications-of-integration\u002Fareas-and-volumes":24748,"\u002Fcalculus\u002Fapplications-of-integration\u002Fwork-average-value-and-arc-length":24753,"\u002Fcalculus\u002Fapplications-of-integration\u002Fphysics-economics-and-probability":24757,"\u002Fcalculus\u002Fexponential-logarithmic-and-inverse-functions\u002Finverse-functions-logarithms-and-exponentials":24761,"\u002Fcalculus\u002Fexponential-logarithmic-and-inverse-functions\u002Fgrowth-decay-inverse-trig-and-hyperbolic-functions":24766,"\u002Fcalculus\u002Fexponential-logarithmic-and-inverse-functions\u002Flhospitals-rule":24770,"\u002Fcalculus\u002Ftechniques-of-integration\u002Fintegration-by-parts":24774,"\u002Fcalculus\u002Ftechniques-of-integration\u002Ftrigonometric-integrals-and-substitution":24779,"\u002Fcalculus\u002Ftechniques-of-integration\u002Fpartial-fractions-and-integration-strategy":24783,"\u002Fcalculus\u002Ftechniques-of-integration\u002Fapproximate-and-improper-integrals":24787,"\u002Fcalculus\u002Fparametric-and-polar\u002Fparametric-curves-and-their-calculus":24791,"\u002Fcalculus\u002Fparametric-and-polar\u002Fpolar-coordinates":24796,"\u002Fcalculus\u002Fparametric-and-polar\u002Fconic-sections":24800,"\u002Fcalculus\u002Fsequences-and-series\u002Fsequences":24804,"\u002Fcalculus\u002Fsequences-and-series\u002Fseries-and-the-integral-test":24809,"\u002Fcalculus\u002Fsequences-and-series\u002Fthe-convergence-tests":24813,"\u002Fcalculus\u002Fsequences-and-series\u002Fpower-series":24817,"\u002Fcalculus\u002Fsequences-and-series\u002Ftaylor-and-maclaurin-series":24821,"\u002Fcalculus\u002Fvectors-and-space-curves\u002Fvectors-and-the-dot-product":24825,"\u002Fcalculus\u002Fvectors-and-space-curves\u002Fthe-cross-product-lines-and-planes":24830,"\u002Fcalculus\u002Fvectors-and-space-curves\u002Fcylinders-and-quadric-surfaces":24834,"\u002Fcalculus\u002Fvectors-and-space-curves\u002Fvector-functions-and-space-curves":24838,"\u002Fcalculus\u002Fvectors-and-space-curves\u002Farc-length-curvature-and-motion":24842,"\u002Fcalculus\u002Fpartial-derivatives\u002Ffunctions-of-several-variables":24846,"\u002Fcalculus\u002Fpartial-derivatives\u002Fpartial-derivatives":24851,"\u002Fcalculus\u002Fpartial-derivatives\u002Ftangent-planes-and-the-chain-rule":24854,"\u002Fcalculus\u002Fpartial-derivatives\u002Fdirectional-derivatives-and-the-gradient":24858,"\u002Fcalculus\u002Fpartial-derivatives\u002Foptimization-and-lagrange-multipliers":24862,"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fdouble-integrals":24866,"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Ftriple-integrals-and-coordinate-systems":24871,"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fvector-fields-and-line-integrals":24875,"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fgreens-theorem-curl-and-divergence":24879,"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fsurface-integrals":24883,"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fstokes-and-the-divergence-theorem":24887,"\u002Fcalculus":24891,"\u002Fmechanics\u002Ffoundations\u002Fmeasurement-and-dimensions":24894,"\u002Fmechanics\u002Ffoundations\u002Fvector-algebra":24898,"\u002Fmechanics\u002Fkinematics\u002Fone-dimensional-motion":24902,"\u002Fmechanics\u002Fkinematics\u002Fmotion-graphs":24907,"\u002Fmechanics\u002Fkinematics\u002Fprojectile-motion":24911,"\u002Fmechanics\u002Fkinematics\u002Frelative-motion":24915,"\u002Fmechanics\u002Fkinematics\u002Fcircular-motion":24919,"\u002Fmechanics\u002Fdynamics\u002Fnewtons-laws":24923,"\u002Fmechanics\u002Fdynamics\u002Ffree-body-diagrams":24928,"\u002Fmechanics\u002Fdynamics\u002Ffriction-and-curved-motion":24932,"\u002Fmechanics\u002Fdynamics\u002Fnumerical-dynamics":24936,"\u002Fmechanics\u002Fdynamics\u002Fcenter-of-mass-systems":24940,"\u002Fmechanics\u002Fenergy\u002Fwork-and-kinetic-energy":24944,"\u002Fmechanics\u002Fenergy\u002Fpotential-energy":24949,"\u002Fmechanics\u002Fenergy\u002Fmultiparticle-work":24953,"\u002Fmechanics\u002Fenergy\u002Fmass-energy-and-binding":24957,"\u002Fmechanics\u002Fenergy\u002Fphotons-and-quantization":24961,"\u002Fmechanics\u002Fmomentum\u002Fmomentum-and-collisions":24965,"\u002Fmechanics\u002Fmomentum\u002Fcenter-of-mass-collisions":24970,"\u002Fmechanics\u002Fmomentum\u002Frocket-propulsion":24974,"\u002Fmechanics\u002Frotation\u002Frotational-inertia":24978,"\u002Fmechanics\u002Frotation\u002Frotational-dynamics":24983,"\u002Fmechanics\u002Frotation\u002Frolling-motion":24987,"\u002Fmechanics\u002Frotation\u002Fangular-momentum":24991,"\u002Fmechanics\u002Frotation\u002Frolling-resistance":24995,"\u002Fmechanics\u002Frotation\u002Fgyroscopic-precession":24999,"\u002Fmechanics\u002Fgravity-and-matter\u002Fkeplerian-orbits":25003,"\u002Fmechanics\u002Fgravity-and-matter\u002Fgravitational-fields":25008,"\u002Fmechanics\u002Fgravity-and-matter\u002Fstatic-equilibrium":25012,"\u002Fmechanics\u002Fgravity-and-matter\u002Ffluid-statics":25016,"\u002Fmechanics\u002Fgravity-and-matter\u002Ffluid-flow":25020,"\u002Fmechanics\u002Fgravity-and-matter\u002Forbital-motion":25024,"\u002Fmechanics\u002Fgravity-and-matter\u002Fstress-and-elasticity":25028,"\u002Fmechanics\u002Foscillations-waves\u002Fdamped-oscillators":25032,"\u002Fmechanics\u002Foscillations-waves\u002Ftravelling-waves":25037,"\u002Fmechanics\u002Foscillations-waves\u002Fwave-superposition":25041,"\u002Fmechanics\u002Foscillations-waves\u002Fstanding-waves":25045,"\u002Fmechanics\u002Foscillations-waves\u002Fsound-waves":25049,"\u002Fmechanics\u002Foscillations-waves\u002Fdoppler-effect":25053,"\u002Fmechanics\u002Foscillations-waves\u002Fwave-packets":25057,"\u002Fmechanics\u002Foscillations-waves\u002Fbeats-and-coupling":25061,"\u002Fmechanics\u002Foscillations-waves\u002Fsimple-harmonic-motion":25065,"\u002Fmechanics\u002Foscillations-waves\u002Fpendulum-motion":25069,"\u002Fmechanics\u002Foscillations-waves\u002Fdriven-oscillators":25073,"\u002Fmechanics\u002Foscillations-waves\u002Fwave-boundaries":25077,"\u002Fmechanics\u002Fthermodynamics\u002Fkinetic-theory-of-ideal-gases":25081,"\u002Fmechanics\u002Fthermodynamics\u002Ffirst-law-of-thermodynamics":25086,"\u002Fmechanics\u002Fthermodynamics\u002Fentropy-and-the-second-law":25090,"\u002Fmechanics\u002Fthermodynamics\u002Fthermal-processes":25094,"\u002Fmechanics\u002Fthermodynamics\u002Fphase-changes":25098,"\u002Fmechanics\u002Fthermodynamics\u002Fthermal-machines":25102,"\u002Fmechanics":25106,"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Fcharge-and-conductors":25109,"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Fcoulombs-law":25114,"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Felectric-field-and-force":25118,"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Felectric-field-maps":25122,"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Felectric-dipoles":25126,"\u002Felectricity-and-magnetism\u002Fcontinuous-charge-distributions\u002Fcontinuous-charge-fields":25130,"\u002Felectricity-and-magnetism\u002Fcontinuous-charge-distributions\u002Fgauss-law-and-conductors":25135,"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Fpoint-charge-potential":25139,"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Fpotential-gradients-and-equipotentials":25144,"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Felectrostatic-energy-and-pressure":25148,"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Flaplace-boundary-problems":25152,"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Fcontinuous-charge-potentials":25156,"\u002Felectricity-and-magnetism\u002Fcapacitance\u002Fcapacitance-fundamentals":25160,"\u002Felectricity-and-magnetism\u002Fcapacitance\u002Fcapacitor-networks":25165,"\u002Felectricity-and-magnetism\u002Fcapacitance\u002Fcapacitor-energy-and-force":25169,"\u002Felectricity-and-magnetism\u002Fcapacitance\u002Fdielectric-polarization-and-breakdown":25173,"\u002Felectricity-and-magnetism\u002Fdirect-current-circuits\u002Fcurrent-and-resistance":25177,"\u002Felectricity-and-magnetism\u002Fdirect-current-circuits\u002Fkirchhoff-network-analysis":25182,"\u002Felectricity-and-magnetism\u002Fdirect-current-circuits\u002Frc-transients":25186,"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fmagnetic-trajectories":25190,"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fhall-effect":25195,"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fmagnetic-force-on-conductors":25199,"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fmagnetic-dipoles":25203,"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fmass-spectrometry":25207,"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fmoving-charge-fields":25211,"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fbiot-savart-law":25216,"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fcircular-current-loops":25220,"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Famperes-law":25224,"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fgauss-law-for-magnetism":25228,"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fmagnetic-materials":25232,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Fmagnetic-flux":25236,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Ffaradays-law":25241,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Flenzs-law":25245,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Fmotional-emf":25249,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Feddy-currents":25253,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Fself-inductance":25257,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Fmagnetic-energy":25261,"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Frl-circuits":25265,"\u002Felectricity-and-magnetism\u002Falternating-current\u002Fac-fundamentals":25269,"\u002Felectricity-and-magnetism\u002Falternating-current\u002Freactance":25274,"\u002Felectricity-and-magnetism\u002Falternating-current\u002Frlc-resonance":25278,"\u002Felectricity-and-magnetism\u002Falternating-current\u002Fac-power":25282,"\u002Felectricity-and-magnetism\u002Falternating-current\u002Ftransformers":25286,"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Fdisplacement-current":25290,"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Felectromagnetic-waves":25295,"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Felectromagnetic-momentum":25299,"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Fdipole-radiation":25303,"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Fpolarization":25307,"\u002Felectricity-and-magnetism\u002Foptics\u002Freflection-and-refraction":25311,"\u002Felectricity-and-magnetism\u002Foptics\u002Fthin-lenses":25316,"\u002Felectricity-and-magnetism\u002Foptics\u002Fspherical-mirrors":25320,"\u002Felectricity-and-magnetism":25324,"\u002Flinear-algebra\u002Flinear-systems\u002Fsystems-and-echelon-forms":25327,"\u002Flinear-algebra\u002Flinear-systems\u002Fvector-and-matrix-equations":25332,"\u002Flinear-algebra\u002Flinear-systems\u002Fsolution-sets-and-applications":25336,"\u002Flinear-algebra\u002Flinear-systems\u002Flinear-independence":25340,"\u002Flinear-algebra\u002Flinear-systems\u002Flinear-transformations":25344,"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fmatrix-operations":25348,"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fmatrix-inverse-and-invertibility":25353,"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fpartitioned-matrices-and-lu":25357,"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fsubspaces-dimension-rank":25361,"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fapplications-leontief-and-graphics":25365,"\u002Flinear-algebra\u002Fdeterminants\u002Fdeterminants-and-cofactors":25369,"\u002Flinear-algebra\u002Fdeterminants\u002Fproperties-of-determinants":25374,"\u002Flinear-algebra\u002Fdeterminants\u002Fcramer-volume-and-area":25378,"\u002Flinear-algebra\u002Fvector-spaces\u002Fvector-spaces-and-subspaces":25382,"\u002Flinear-algebra\u002Fvector-spaces\u002Fnull-and-column-spaces":25387,"\u002Flinear-algebra\u002Fvector-spaces\u002Fbases-and-independent-sets":25391,"\u002Flinear-algebra\u002Fvector-spaces\u002Fcoordinate-systems":25395,"\u002Flinear-algebra\u002Fvector-spaces\u002Fdimension-and-rank":25399,"\u002Flinear-algebra\u002Fvector-spaces\u002Fchange-of-basis":25403,"\u002Flinear-algebra\u002Fvector-spaces\u002Fdifference-equations-and-markov":25407,"\u002Flinear-algebra\u002Feigenvalues\u002Feigenvectors-and-eigenvalues":25411,"\u002Flinear-algebra\u002Feigenvalues\u002Fthe-characteristic-equation":25416,"\u002Flinear-algebra\u002Feigenvalues\u002Fdiagonalization":25420,"\u002Flinear-algebra\u002Feigenvalues\u002Feigenvectors-and-linear-transformations":25424,"\u002Flinear-algebra\u002Feigenvalues\u002Fcomplex-eigenvalues":25428,"\u002Flinear-algebra\u002Feigenvalues\u002Fdynamical-systems":25432,"\u002Flinear-algebra\u002Feigenvalues\u002Fpower-method":25436,"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Finner-product-length-orthogonality":25440,"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Forthogonal-sets-and-projections":25445,"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Fgram-schmidt-and-qr":25449,"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Fleast-squares-problems":25453,"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Fleast-squares-applications":25457,"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Finner-product-spaces":25461,"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fdiagonalizing-symmetric-matrices":25465,"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fquadratic-forms":25470,"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fconstrained-optimization":25474,"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fsingular-value-decomposition":25478,"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fsvd-applications-pca-imaging":25482,"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fnumerical-thinking-and-matrix-computation":25486,"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Flu-and-cholesky":25491,"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fconditioning-and-floating-point":25495,"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fstability-and-error-analysis":25499,"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fqr-and-numerical-least-squares":25503,"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fnumerical-eigenvalues-and-svd":25507,"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Faffine-combinations":25511,"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Faffine-independence-and-barycentric-coordinates":25516,"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Fconvex-combinations-and-convex-sets":25520,"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Fhyperplanes-and-polytopes":25524,"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Fcurves-and-surfaces":25528,"\u002Flinear-algebra":25532,"\u002Ftheory-of-computation":25535,"\u002Fcomputer-architecture\u002Ffoundations\u002Fbits-bytes-and-words":25538,"\u002Fcomputer-architecture\u002Ffoundations\u002Finteger-representation":25542,"\u002Fcomputer-architecture\u002Ffoundations\u002Finteger-arithmetic":25546,"\u002Fcomputer-architecture\u002Ffoundations\u002Ffloating-point":25550,"\u002Fcomputer-architecture\u002Ffoundations\u002Fboolean-algebra-and-bit-manipulation":25554,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fthe-machines-view":25558,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fdata-movement":25563,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Farithmetic-and-logic":25567,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fcontrol-flow":25571,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fprocedures":25575,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Farrays-structs-and-alignment":25579,"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fmemory-layout-and-buffer-overflows":25583,"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Fwhat-an-isa-is":25587,"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Finstruction-formats-and-operands":25592,"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Faddressing-modes":25596,"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Fthe-y86-64-instruction-set":25600,"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Fy86-64-programming":25604,"\u002Fcomputer-architecture\u002Fdigital-logic\u002Ftransistors-gates-and-boolean-functions":25608,"\u002Fcomputer-architecture\u002Fdigital-logic\u002Fcombinational-logic-and-hcl":25613,"\u002Fcomputer-architecture\u002Fdigital-logic\u002Fmultiplexers-decoders-and-the-alu":25617,"\u002Fcomputer-architecture\u002Fdigital-logic\u002Fmemory-elements-latches-flip-flops-and-clocking":25621,"\u002Fcomputer-architecture\u002Fdigital-logic\u002Fregister-files-and-random-access-memory":25625,"\u002Fcomputer-architecture\u002Fprocessor-design\u002Fthe-fetch-decode-execute-cycle":25629,"\u002Fcomputer-architecture\u002Fprocessor-design\u002Fthe-seq-stages":25634,"\u002Fcomputer-architecture\u002Fprocessor-design\u002Fcontrol-logic-and-sequencing":25638,"\u002Fcomputer-architecture\u002Fprocessor-design\u002Fassembling-seq":25642,"\u002Fcomputer-architecture\u002Fprocessor-design\u002Ftracing-a-program":25646,"\u002Fcomputer-architecture\u002Fpipelining\u002Fpipelining-principles":25650,"\u002Fcomputer-architecture\u002Fpipelining\u002Ffrom-seq-to-pipe":25655,"\u002Fcomputer-architecture\u002Fpipelining\u002Fdata-hazards-stalling-and-forwarding":25659,"\u002Fcomputer-architecture\u002Fpipelining\u002Fcontrol-hazards-and-branch-prediction":25663,"\u002Fcomputer-architecture\u002Fpipelining\u002Fthe-complete-pipe-processor":25667,"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Fstorage-technologies-and-the-latency-gap":25671,"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Flocality":25676,"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Fcache-memories-direct-mapped":25680,"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Fset-associative-and-write-policies":25684,"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Fcache-performance-and-cache-friendly-code":25688,"\u002Fcomputer-architecture\u002Fvirtual-memory\u002Faddress-spaces-and-translation":25692,"\u002Fcomputer-architecture\u002Fvirtual-memory\u002Fpage-tables-and-page-faults":25697,"\u002Fcomputer-architecture\u002Fvirtual-memory\u002Fthe-tlb-and-multi-level-page-tables":25701,"\u002Fcomputer-architecture\u002Fexceptions-and-io\u002Fexceptional-control-flow":25705,"\u002Fcomputer-architecture\u002Fexceptions-and-io\u002Finterrupts-and-the-kernel":25710,"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fprocesses-threads-and-parallelism":25714,"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fhardware-multithreading":25719,"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fcache-coherence":25723,"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fmemory-consistency-and-synchronization":25727,"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fmulticore-organization":25731,"\u002Fcomputer-architecture\u002Fcapstone\u002Fthe-whole-machine":25735,"\u002Fcomputer-architecture\u002Fcapstone\u002Fassembling-a-complete-cpu":25740,"\u002Fcomputer-architecture":25744,"\u002Fdifferential-equations\u002Ffoundations\u002Fmodels-and-direction-fields":25747,"\u002Fdifferential-equations\u002Ffoundations\u002Fclassification-and-terminology":25751,"\u002Fdifferential-equations\u002Ffirst-order\u002Flinear-first-order-integrating-factors":25755,"\u002Fdifferential-equations\u002Ffirst-order\u002Fseparable-and-exact":25760,"\u002Fdifferential-equations\u002Ffirst-order\u002Fmodeling-first-order":25764,"\u002Fdifferential-equations\u002Ffirst-order\u002Fautonomous-and-population-dynamics":25768,"\u002Fdifferential-equations\u002Ffirst-order\u002Fexistence-uniqueness-euler":25772,"\u002Fdifferential-equations\u002Ffirst-order\u002Ffirst-order-difference-equations":25776,"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fhomogeneous-constant-coefficients":25780,"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fcomplex-and-repeated-roots":25785,"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fnonhomogeneous-undetermined-coefficients":25789,"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fvariation-of-parameters":25793,"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fmechanical-electrical-vibrations":25797,"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fhigher-order-linear":25801,"\u002Fdifferential-equations\u002Fseries-solutions\u002Fpower-series-ordinary-points":25805,"\u002Fdifferential-equations\u002Fseries-solutions\u002Fregular-singular-frobenius":25810,"\u002Fdifferential-equations\u002Fseries-solutions\u002Fbessel-and-special-functions":25814,"\u002Fdifferential-equations\u002Flaplace\u002Flaplace-definition-ivps":25818,"\u002Fdifferential-equations\u002Flaplace\u002Fstep-impulse-convolution":25823,"\u002Fdifferential-equations\u002Fsystems\u002Fmatrices-eigenvalues-review":25827,"\u002Fdifferential-equations\u002Fsystems\u002Fconstant-coefficient-systems-phase-portraits":25832,"\u002Fdifferential-equations\u002Fsystems\u002Frepeated-eigenvalues-fundamental-matrices":25836,"\u002Fdifferential-equations\u002Fnumerical\u002Feuler-and-runge-kutta":25840,"\u002Fdifferential-equations\u002Fnumerical\u002Fmultistep-systems-stability":25845,"\u002Fdifferential-equations\u002Fnonlinear\u002Fphase-plane-autonomous-stability":25849,"\u002Fdifferential-equations\u002Fnonlinear\u002Flocally-linear-and-liapunov":25854,"\u002Fdifferential-equations\u002Fnonlinear\u002Fcompeting-species-predator-prey-limit-cycles":25858,"\u002Fdifferential-equations\u002Fpdes-fourier-bvp\u002Ffourier-series":25862,"\u002Fdifferential-equations\u002Fpdes-fourier-bvp\u002Fheat-wave-laplace-equations":25867,"\u002Fdifferential-equations\u002Fpdes-fourier-bvp\u002Fsturm-liouville":25871,"\u002Fdifferential-equations\u002Fhistory-variations\u002Fcalculus-of-variations":25875,"\u002Fdifferential-equations\u002Fhistory-variations\u002Fhistorical-notes":25880,"\u002Fdifferential-equations":25884,"\u002Frelativity\u002Ffoundations\u002Fspecial-relativity-postulates":25887,"\u002Frelativity\u002Ffoundations\u002Florentz-transformation-spacetime":25892,"\u002Frelativity\u002Ffoundations\u002Ftime-dilation-length-contraction":25896,"\u002Frelativity\u002Ffoundations\u002Frelativistic-momentum-energy":25900,"\u002Frelativity\u002Ffoundations\u002Fgeneral-relativity":25904,"\u002Frelativity\u002Fspacetime-and-the-lorentz-group\u002Fminkowski-spacetime-and-the-interval":25908,"\u002Frelativity\u002Fspacetime-and-the-lorentz-group\u002Ffour-vectors-and-index-notation":25913,"\u002Frelativity\u002Fspacetime-and-the-lorentz-group\u002Fthe-lorentz-group-and-rapidity":25917,"\u002Frelativity\u002Fspacetime-and-the-lorentz-group\u002Fdoppler-aberration-and-appearance":25921,"\u002Frelativity\u002Frelativistic-dynamics\u002Ffour-momentum-force-and-accelerated-motion":25925,"\u002Frelativity\u002Frelativistic-dynamics\u002Fparticle-decays-and-two-body-kinematics":25930,"\u002Frelativity\u002Frelativistic-dynamics\u002Fcollisions-thresholds-and-the-cm-frame":25934,"\u002Frelativity\u002Frelativistic-dynamics\u002Fmandelstam-variables-and-invariants":25938,"\u002Frelativity\u002Fcovariant-electrodynamics\u002Ffour-current-and-the-four-potential":25942,"\u002Frelativity\u002Fcovariant-electrodynamics\u002Fthe-electromagnetic-field-tensor":25947,"\u002Frelativity\u002Fcovariant-electrodynamics\u002Ftransformation-of-electric-and-magnetic-fields":25951,"\u002Frelativity\u002Fcovariant-electrodynamics\u002Fcovariant-maxwell-and-the-stress-energy-tensor":25955,"\u002Frelativity\u002Fcurved-spacetime\u002Fthe-equivalence-principle-formalized":25959,"\u002Frelativity\u002Fcurved-spacetime\u002Fmanifolds-vectors-and-the-metric":25964,"\u002Frelativity\u002Fcurved-spacetime\u002Fcovariant-derivative-and-christoffel-symbols":25968,"\u002Frelativity\u002Fcurved-spacetime\u002Fgeodesics-and-the-geodesic-equation":25972,"\u002Frelativity\u002Fcurved-spacetime\u002Fcurvature-riemann-and-geodesic-deviation":25976,"\u002Frelativity\u002Fcurved-spacetime\u002Fthe-einstein-field-equations":25980,"\u002Frelativity\u002Fthe-schwarzschild-solution\u002Fthe-schwarzschild-metric":25984,"\u002Frelativity\u002Fthe-schwarzschild-solution\u002Fgeodesics-and-orbits-in-schwarzschild":25989,"\u002Frelativity\u002Fthe-schwarzschild-solution\u002Flight-bending-and-null-geodesics":25993,"\u002Frelativity\u002Ftests-of-general-relativity\u002Fperihelion-precession-of-mercury":25997,"\u002Frelativity\u002Ftests-of-general-relativity\u002Fdeflection-of-light-and-gravitational-lensing":26002,"\u002Frelativity\u002Ftests-of-general-relativity\u002Fgravitational-redshift-and-shapiro-delay":26006,"\u002Frelativity\u002Ftests-of-general-relativity\u002Frelativity-in-technology-gps":26010,"\u002Frelativity\u002Fblack-holes\u002Fhorizons-and-coordinate-singularities":26014,"\u002Frelativity\u002Fblack-holes\u002Frotating-and-charged-black-holes":26019,"\u002Frelativity\u002Fblack-holes\u002Fblack-hole-thermodynamics":26023,"\u002Frelativity\u002Fgravitational-waves\u002Flinearized-gravity-and-wave-solutions":26027,"\u002Frelativity\u002Fgravitational-waves\u002Fgeneration-and-the-quadrupole-formula":26032,"\u002Frelativity\u002Fgravitational-waves\u002Fdetection-ligo-and-the-first-events":26036,"\u002Frelativity\u002Fcosmological-bridge\u002Fthe-cosmological-principle-and-flrw-metric":26040,"\u002Frelativity\u002Fcosmological-bridge\u002Ffriedmann-equations-and-cosmic-dynamics":26045,"\u002Frelativity":26049,"\u002Fphysical-computing":26052,"\u002Fquantum-mechanics\u002Fold-quantum-theory\u002Fblackbody-radiation-and-the-planck-quantum":26055,"\u002Fquantum-mechanics\u002Fold-quantum-theory\u002Fthe-photoelectric-effect-and-the-photon":26060,"\u002Fquantum-mechanics\u002Fold-quantum-theory\u002Fx-rays-and-the-compton-effect":26064,"\u002Fquantum-mechanics\u002Fold-quantum-theory\u002Fthe-old-quantum-theory-bohr-and-sommerfeld":26068,"\u002Fquantum-mechanics\u002Fmatter-waves\u002Fde-broglie-waves-and-electron-diffraction":26072,"\u002Fquantum-mechanics\u002Fmatter-waves\u002Fwave-packets-and-the-probability-interpretation":26077,"\u002Fquantum-mechanics\u002Fmatter-waves\u002Fthe-uncertainty-principle":26081,"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fthe-schrodinger-equation-in-one-dimension":26085,"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fthe-free-particle-and-wave-packet-dynamics":26090,"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fparticle-in-infinite-and-finite-square-wells":26094,"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Foperators-expectation-values-and-the-harmonic-oscillator":26098,"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fthe-dirac-delta-potential":26102,"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fbarrier-penetration-and-quantum-tunneling":26106,"\u002Fquantum-mechanics\u002Fformalism\u002Fhilbert-space-and-dirac-notation":26110,"\u002Fquantum-mechanics\u002Fformalism\u002Fobservables-hermitian-operators-and-eigenvalues":26115,"\u002Fquantum-mechanics\u002Fformalism\u002Fthe-postulates-and-quantum-measurement":26119,"\u002Fquantum-mechanics\u002Fformalism\u002Fposition-momentum-and-continuous-spectra":26123,"\u002Fquantum-mechanics\u002Fformalism\u002Fcommutators-and-the-generalized-uncertainty-principle":26127,"\u002Fquantum-mechanics\u002Fformalism\u002Ftime-evolution-schrodinger-and-heisenberg-pictures":26131,"\u002Fquantum-mechanics\u002Foscillator-and-symmetry\u002Fladder-operators-and-the-number-states":26135,"\u002Fquantum-mechanics\u002Foscillator-and-symmetry\u002Fcoherent-and-squeezed-states":26140,"\u002Fquantum-mechanics\u002Foscillator-and-symmetry\u002Fsymmetries-generators-and-conservation-laws":26144,"\u002Fquantum-mechanics\u002Foscillator-and-symmetry\u002Fparity-time-reversal-and-discrete-symmetries":26148,"\u002Fquantum-mechanics\u002Fangular-momentum\u002Forbital-angular-momentum-and-spherical-harmonics":26152,"\u002Fquantum-mechanics\u002Fangular-momentum\u002Fthe-angular-momentum-algebra":26156,"\u002Fquantum-mechanics\u002Fangular-momentum\u002Faddition-of-angular-momenta-and-clebsch-gordan":26160,"\u002Fquantum-mechanics\u002Fcentral-potentials\u002Fthe-schrodinger-equation-in-three-dimensions":26164,"\u002Fquantum-mechanics\u002Fcentral-potentials\u002Fthe-hydrogen-atom":26169,"\u002Fquantum-mechanics\u002Fcentral-potentials\u002Fthe-isotropic-oscillator-and-hidden-symmetry":26173,"\u002Fquantum-mechanics\u002Fspin\u002Fspin-half-pauli-matrices-and-stern-gerlach":26177,"\u002Fquantum-mechanics\u002Fspin\u002Fspin-in-a-magnetic-field-precession-and-resonance":26182,"\u002Fquantum-mechanics\u002Fspin\u002Ftwo-level-systems-and-the-bloch-sphere":26186,"\u002Fquantum-mechanics\u002Fidentical-particles\u002Fidentical-particles-and-exchange-symmetry":26190,"\u002Fquantum-mechanics\u002Fidentical-particles\u002Fthe-pauli-principle-atoms-and-the-periodic-table":26195,"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Ftime-independent-perturbation-theory":26199,"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Ffine-structure-and-the-real-hydrogen-atom":26204,"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Fthe-zeeman-and-stark-effects":26208,"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Fthe-variational-method":26212,"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Fthe-wkb-approximation":26216,"\u002Fquantum-mechanics":26220,"\u002Freal-analysis\u002Ffoundations\u002Fsets-logic-functions":26223,"\u002Freal-analysis\u002Ffoundations\u002Fordered-fields-completeness":26228,"\u002Freal-analysis\u002Ffoundations\u002Fabsolute-value-bounds":26232,"\u002Freal-analysis\u002Ffoundations\u002Fintervals-uncountability":26236,"\u002Freal-analysis\u002Fsequences-series\u002Fsequences-limits":26240,"\u002Freal-analysis\u002Fsequences-series\u002Flimit-laws-monotone":26245,"\u002Freal-analysis\u002Fsequences-series\u002Flimsup-bolzano-weierstrass":26249,"\u002Freal-analysis\u002Fsequences-series\u002Fcauchy-completeness":26253,"\u002Freal-analysis\u002Fsequences-series\u002Fseries-convergence":26257,"\u002Freal-analysis\u002Fsequences-series\u002Fabsolute-conditional-rearrangement":26261,"\u002Freal-analysis\u002Fmetric-spaces\u002Fmetric-spaces-norms":26265,"\u002Freal-analysis\u002Fmetric-spaces\u002Fopen-closed-sets":26270,"\u002Freal-analysis\u002Fmetric-spaces\u002Fconvergence-completeness":26274,"\u002Freal-analysis\u002Fmetric-spaces\u002Fcompactness":26278,"\u002Freal-analysis\u002Fmetric-spaces\u002Fconnectedness":26282,"\u002Freal-analysis\u002Fcontinuity\u002Flimits-of-functions":26286,"\u002Freal-analysis\u002Fcontinuity\u002Fcontinuous-functions":26290,"\u002Freal-analysis\u002Fcontinuity\u002Fevt-ivt":26294,"\u002Freal-analysis\u002Fcontinuity\u002Funiform-continuity":26298,"\u002Freal-analysis\u002Fcontinuity\u002Fcontinuity-metric-spaces":26302,"\u002Freal-analysis\u002Fcontinuity\u002Flimits-infinity-monotone":26306,"\u002Freal-analysis\u002Fdifferentiation\u002Fthe-derivative":26310,"\u002Freal-analysis\u002Fdifferentiation\u002Fmean-value-theorem":26315,"\u002Freal-analysis\u002Fdifferentiation\u002Ftaylors-theorem":26319,"\u002Freal-analysis\u002Fdifferentiation\u002Finverse-function-1d":26323,"\u002Freal-analysis\u002Friemann-integration\u002Fdarboux-integral":26327,"\u002Freal-analysis\u002Friemann-integration\u002Fintegrability-classes":26332,"\u002Freal-analysis\u002Friemann-integration\u002Fproperties-of-the-integral":26336,"\u002Freal-analysis\u002Friemann-integration\u002Ffundamental-theorem":26340,"\u002Freal-analysis\u002Friemann-integration\u002Flog-exp-improper":26343,"\u002Freal-analysis\u002Ffunction-sequences\u002Fpointwise-uniform-convergence":26347,"\u002Freal-analysis\u002Ffunction-sequences\u002Finterchange-of-limits":26352,"\u002Freal-analysis\u002Ffunction-sequences\u002Fpower-series-weierstrass":26356,"\u002Freal-analysis\u002Ffunction-sequences\u002Fpicard-ode":26360,"\u002Freal-analysis\u002Fseveral-variables\u002Fdifferentiability-rn":26364,"\u002Freal-analysis\u002Fseveral-variables\u002Fgradient-chain-rule":26369,"\u002Freal-analysis\u002Fseveral-variables\u002Fhigher-derivatives-taylor-extrema":26373,"\u002Freal-analysis\u002Fseveral-variables\u002Finverse-implicit-theorems":26377,"\u002Freal-analysis\u002Fseveral-variables\u002Fmultiple-integrals":26381,"\u002Freal-analysis":26385,"\u002Fabstract-algebra\u002Ffoundations\u002Fsets-functions-relations":26388,"\u002Fabstract-algebra\u002Ffoundations\u002Fintegers-and-modular-arithmetic":26392,"\u002Fabstract-algebra\u002Fgroups-and-symmetry\u002Fgroup-axioms-and-first-examples":26396,"\u002Fabstract-algebra\u002Fgroups-and-symmetry\u002Fdihedral-and-symmetric-groups":26401,"\u002Fabstract-algebra\u002Fgroups-and-symmetry\u002Fmatrix-and-quaternion-groups":26405,"\u002Fabstract-algebra\u002Fgroups-and-symmetry\u002Fhomomorphisms-and-group-actions":26409,"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fsubgroups-and-substructures":26413,"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fcyclic-groups":26418,"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fgeneration-and-subgroup-lattices":26422,"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fcosets-lagrange-and-normal-subgroups":26426,"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fisomorphism-theorems":26430,"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fcomposition-series-and-the-alternating-group":26434,"\u002Fabstract-algebra\u002Fgroup-actions-and-sylow\u002Factions-and-cayleys-theorem":26438,"\u002Fabstract-algebra\u002Fgroup-actions-and-sylow\u002Fconjugation-and-the-class-equation":26443,"\u002Fabstract-algebra\u002Fgroup-actions-and-sylow\u002Fsylow-theorems":26447,"\u002Fabstract-algebra\u002Fgroup-actions-and-sylow\u002Fautomorphisms-and-simple-groups":26451,"\u002Fabstract-algebra\u002Fproducts-and-group-structure\u002Fdirect-products-and-finite-abelian-groups":26455,"\u002Fabstract-algebra\u002Fproducts-and-group-structure\u002Fsemidirect-products":26460,"\u002Fabstract-algebra\u002Fproducts-and-group-structure\u002Fnilpotent-and-solvable-groups":26464,"\u002Fabstract-algebra\u002Fproducts-and-group-structure\u002Fclassifying-small-groups":26468,"\u002Fabstract-algebra\u002Fring-theory\u002Frings-definitions-and-examples":26472,"\u002Fabstract-algebra\u002Fring-theory\u002Fideals-quotients-and-homomorphisms":26477,"\u002Fabstract-algebra\u002Fring-theory\u002Ffractions-and-the-chinese-remainder-theorem":26481,"\u002Fabstract-algebra\u002Ffactorization-and-polynomials\u002Feuclidean-domains-pids-ufds":26485,"\u002Fabstract-algebra\u002Ffactorization-and-polynomials\u002Fpolynomial-rings-over-fields":26490,"\u002Fabstract-algebra\u002Ffactorization-and-polynomials\u002Fgauss-lemma-and-unique-factorization":26494,"\u002Fabstract-algebra\u002Ffactorization-and-polynomials\u002Firreducibility-criteria-and-groebner":26498,"\u002Fabstract-algebra\u002Fmodule-theory\u002Fintroduction-to-modules":26502,"\u002Fabstract-algebra\u002Fmodule-theory\u002Ffree-modules-and-direct-sums":26507,"\u002Fabstract-algebra\u002Fmodule-theory\u002Ftensor-products-and-exact-sequences":26511,"\u002Fabstract-algebra\u002Fmodule-theory\u002Fvector-spaces-and-linear-maps":26515,"\u002Fabstract-algebra\u002Fmodules-over-pids\u002Fstructure-theorem-over-pids":26519,"\u002Fabstract-algebra\u002Fmodules-over-pids\u002Frational-canonical-form":26524,"\u002Fabstract-algebra\u002Fmodules-over-pids\u002Fjordan-canonical-form":26528,"\u002Fabstract-algebra\u002Ffield-theory\u002Ffield-extensions-and-algebraic-elements":26532,"\u002Fabstract-algebra\u002Ffield-theory\u002Fstraightedge-and-compass-constructions":26537,"\u002Fabstract-algebra\u002Ffield-theory\u002Fsplitting-fields-and-algebraic-closure":26541,"\u002Fabstract-algebra\u002Ffield-theory\u002Fseparable-and-cyclotomic-extensions":26545,"\u002Fabstract-algebra\u002Fgalois-theory\u002Fthe-galois-correspondence":26549,"\u002Fabstract-algebra\u002Fgalois-theory\u002Ffinite-fields":26554,"\u002Fabstract-algebra\u002Fgalois-theory\u002Fcyclotomic-and-abelian-extensions":26558,"\u002Fabstract-algebra\u002Fgalois-theory\u002Fgalois-groups-of-polynomials":26562,"\u002Fabstract-algebra\u002Fgalois-theory\u002Fsolvability-by-radicals-and-the-quintic":26566,"\u002Fabstract-algebra\u002Fcapstone\u002Fcommutative-algebra-and-algebraic-geometry":26570,"\u002Fabstract-algebra\u002Fcapstone\u002Frepresentation-and-character-theory":26575,"\u002Fabstract-algebra":26579,"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fatomic-spectra-rutherford":26582,"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fbohr-model-hydrogen":26587,"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fx-ray-spectra-franck-hertz":26591,"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fbohr-sommerfeld-old-quantum-theory":26595,"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fold-quantum-theory-limits-wkb":26599,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fschrodinger-3d-hydrogen":26603,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fhydrogen-wave-functions":26608,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fradial-equation-in-full":26612,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fsymmetry-degeneracy-runge-lenz":26616,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fexpectation-values-virial":26620,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fquantum-defects-alkali-spectra":26624,"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Frydberg-atoms":26628,"\u002Fatomic-physics\u002Ffine-structure-and-the-dirac-atom\u002Frelativistic-kinetic-correction":26632,"\u002Fatomic-physics\u002Ffine-structure-and-the-dirac-atom\u002Fspin-orbit-thomas-precession":26637,"\u002Fatomic-physics\u002Ffine-structure-and-the-dirac-atom\u002Fdarwin-term-fine-structure-formula":26641,"\u002Fatomic-physics\u002Ffine-structure-and-the-dirac-atom\u002Fdirac-equation-hydrogen":26645,"\u002Fatomic-physics\u002Fqed-corrections-and-hyperfine-structure\u002Flamb-shift-qed":26649,"\u002Fatomic-physics\u002Fqed-corrections-and-hyperfine-structure\u002Fhyperfine-structure-21cm":26654,"\u002Fatomic-physics\u002Fqed-corrections-and-hyperfine-structure\u002Fnuclear-effects-isotope-shift":26658,"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fperiodic-table-atomic-spectra":26662,"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fcentral-field-self-consistent":26667,"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fidentical-particles-hartree-fock":26671,"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fhelium-two-electron-atom":26675,"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fls-jj-coupling-term-symbols":26679,"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fhund-rules-ground-terms":26683,"\u002Fatomic-physics\u002Fatoms-in-external-fields\u002Fzeeman-effect":26687,"\u002Fatomic-physics\u002Fatoms-in-external-fields\u002Fpaschen-back-intermediate":26692,"\u002Fatomic-physics\u002Fatoms-in-external-fields\u002Fstark-effect-polarizability":26696,"\u002Fatomic-physics\u002Fradiative-transitions-and-line-shapes\u002Ftime-dependent-perturbation-golden-rule":26700,"\u002Fatomic-physics\u002Fradiative-transitions-and-line-shapes\u002Fdipole-approximation-einstein-coefficients":26705,"\u002Fatomic-physics\u002Fradiative-transitions-and-line-shapes\u002Fselection-rules-forbidden-transitions":26709,"\u002Fatomic-physics\u002Fradiative-transitions-and-line-shapes\u002Flifetimes-and-line-shapes":26713,"\u002Fatomic-physics\u002Flasers-and-spectroscopy\u002Flaser-principles":26717,"\u002Fatomic-physics\u002Flasers-and-spectroscopy\u002Fspectroscopy-techniques":26722,"\u002Fatomic-physics\u002Flasers-and-spectroscopy\u002Fline-catalog-nist-asd":26726,"\u002Fatomic-physics\u002Fmodern-atomic-physics\u002Flaser-cooling-doppler":26730,"\u002Fatomic-physics\u002Fmodern-atomic-physics\u002Fsub-doppler-trapping":26735,"\u002Fatomic-physics\u002Fmodern-atomic-physics\u002Fbose-einstein-condensation":26739,"\u002Fatomic-physics\u002Fmodern-atomic-physics\u002Foptical-clocks-precision":26743,"\u002Fatomic-physics":26747,"\u002Fdatabases":26750,"\u002Fcategory-theory\u002Ffoundations\u002Fwhat-is-a-category":26753,"\u002Fcategory-theory\u002Ffoundations\u002Fexamples-of-categories":26757,"\u002Fcategory-theory\u002Ffoundations\u002Fspecial-morphisms":26761,"\u002Fcategory-theory\u002Ffoundations\u002Ffunctors":26765,"\u002Fcategory-theory\u002Ffoundations\u002Fnatural-transformations":26769,"\u002Fcategory-theory\u002Ffoundations\u002Fsize-and-set-theory":26773,"\u002Fcategory-theory\u002Funiversal-properties\u002Funiversal-properties":26777,"\u002Fcategory-theory\u002Funiversal-properties\u002Fproducts-and-coproducts":26782,"\u002Fcategory-theory\u002Funiversal-properties\u002Fconstructions-on-categories":26786,"\u002Fcategory-theory\u002Frepresentables-yoneda\u002Frepresentable-functors":26790,"\u002Fcategory-theory\u002Frepresentables-yoneda\u002Fyoneda-lemma":26795,"\u002Fcategory-theory\u002Frepresentables-yoneda\u002Fyoneda-consequences":26799,"\u002Fcategory-theory\u002Flimits-colimits\u002Flimits":26803,"\u002Fcategory-theory\u002Flimits-colimits\u002Fproducts-equalizers-pullbacks":26808,"\u002Fcategory-theory\u002Flimits-colimits\u002Fcolimits":26812,"\u002Fcategory-theory\u002Flimits-colimits\u002Fcomputing-limits":26816,"\u002Fcategory-theory\u002Flimits-colimits\u002Flimits-and-functors":26820,"\u002Fcategory-theory\u002Fadjunctions\u002Fadjunctions":26824,"\u002Fcategory-theory\u002Fadjunctions\u002Funits-and-counits":26829,"\u002Fcategory-theory\u002Fadjunctions\u002Fadjunctions-via-universal-arrows":26833,"\u002Fcategory-theory\u002Fadjunctions\u002Ffree-forgetful-adjunctions":26837,"\u002Fcategory-theory\u002Fadjoints-limits\u002Flimits-via-adjoints":26841,"\u002Fcategory-theory\u002Fadjoints-limits\u002Fpresheaf-limits-colimits":26846,"\u002Fcategory-theory\u002Fadjoints-limits\u002Fadjoints-preserve-limits":26850,"\u002Fcategory-theory\u002Fadjoints-limits\u002Fadjoint-functor-theorem":26854,"\u002Fcategory-theory\u002Fmonads-algebras\u002Fmonads":26858,"\u002Fcategory-theory\u002Fmonads-algebras\u002Falgebras-eilenberg-moore":26863,"\u002Fcategory-theory\u002Fmonads-algebras\u002Fkleisli-and-programming":26867,"\u002Fcategory-theory\u002Fmonads-algebras\u002Falgebras-for-endofunctors":26871,"\u002Fcategory-theory\u002Fcartesian-closed-lambda\u002Fcartesian-closed-categories":26875,"\u002Fcategory-theory\u002Fcartesian-closed-lambda\u002Flambda-calculus-correspondence":26880,"\u002Fcategory-theory\u002Fcartesian-closed-lambda\u002Ffixed-points-and-recursion":26884,"\u002Fcategory-theory":26888,"\u002Fdeep-learning\u002Fmathematical-background\u002Flinear-algebra-for-deep-learning":26891,"\u002Fdeep-learning\u002Fmathematical-background\u002Fprobability-and-information-theory":26895,"\u002Fdeep-learning\u002Fmathematical-background\u002Fnumerical-computation":26899,"\u002Fdeep-learning\u002Fmathematical-background\u002Fcalculus":26903,"\u002Fdeep-learning\u002Ffoundations\u002Fwhat-is-deep-learning":26906,"\u002Fdeep-learning\u002Ffoundations\u002Fmachine-learning-refresher":26910,"\u002Fdeep-learning\u002Ffoundations\u002Flinear-models-and-the-perceptron":26914,"\u002Fdeep-learning\u002Fneural-networks\u002Fthe-multilayer-perceptron":26918,"\u002Fdeep-learning\u002Fneural-networks\u002Factivation-functions":26923,"\u002Fdeep-learning\u002Fneural-networks\u002Funiversal-approximation":26927,"\u002Fdeep-learning\u002Fneural-networks\u002Fbackpropagation":26931,"\u002Fdeep-learning\u002Fneural-networks\u002Floss-functions-and-output-units":26935,"\u002Fdeep-learning\u002Foptimization\u002Fgradient-descent-and-sgd":26939,"\u002Fdeep-learning\u002Foptimization\u002Fmomentum-and-adaptive-methods":26944,"\u002Fdeep-learning\u002Foptimization\u002Finitialization":26948,"\u002Fdeep-learning\u002Foptimization\u002Fthe-optimization-landscape":26952,"\u002Fdeep-learning\u002Foptimization\u002Fsecond-order-and-approximate-methods":26956,"\u002Fdeep-learning\u002Fregularization\u002Fregularization-overview":26960,"\u002Fdeep-learning\u002Fregularization\u002Fdropout-and-data-augmentation":26965,"\u002Fdeep-learning\u002Fregularization\u002Fearly-stopping-and-parameter-sharing":26969,"\u002Fdeep-learning\u002Fregularization\u002Fnormalization":26973,"\u002Fdeep-learning\u002Farchitectures\u002Fconvolutional-networks":26976,"\u002Fdeep-learning\u002Farchitectures\u002Fcnn-architectures":26980,"\u002Fdeep-learning\u002Farchitectures\u002Frecurrent-networks":26983,"\u002Fdeep-learning\u002Farchitectures\u002Flstm-and-gru":26986,"\u002Fdeep-learning\u002Farchitectures\u002Fattention-and-transformers":26989,"\u002Fdeep-learning\u002Farchitectures\u002Fthe-transformer-architecture":26990,"\u002Fdeep-learning\u002Farchitectures\u002Ftransformers-in-practice":26993,"\u002Fdeep-learning\u002Farchitectures\u002Fgraph-neural-networks":26997,"\u002Fdeep-learning\u002Farchitectures\u002Fstate-space-models":27001,"\u002Fdeep-learning\u002Ftheory\u002Fgeneralization-theory":27005,"\u002Fdeep-learning\u002Ftheory\u002Fadversarial-robustness":27010,"\u002Fdeep-learning\u002Ftheory\u002Fadversarial-defenses":27014,"\u002Fdeep-learning\u002Ftheory\u002Fbayesian-and-ensemble-methods":27018,"\u002Fdeep-learning\u002Ftheory\u002Fdeep-equilibrium-models":27022,"\u002Fdeep-learning\u002Fgenerative-models\u002Flinear-factor-models":27026,"\u002Fdeep-learning\u002Fgenerative-models\u002Fautoencoders":27031,"\u002Fdeep-learning\u002Fgenerative-models\u002Fvariational-autoencoders":27035,"\u002Fdeep-learning\u002Fgenerative-models\u002Fgenerative-adversarial-networks":27039,"\u002Fdeep-learning\u002Fgenerative-models\u002Fautoregressive-and-normalizing-flows":27043,"\u002Fdeep-learning\u002Fgenerative-models\u002Fenergy-based-and-boltzmann-machines":27047,"\u002Fdeep-learning\u002Fgenerative-models\u002Fdiffusion-and-score-based-models":27051,"\u002Fdeep-learning\u002Fprobabilistic-methods\u002Fstructured-probabilistic-models":27055,"\u002Fdeep-learning\u002Fprobabilistic-methods\u002Fmonte-carlo-and-mcmc":27060,"\u002Fdeep-learning\u002Fprobabilistic-methods\u002Fapproximate-inference":27064,"\u002Fdeep-learning\u002Fpractical\u002Fpractical-methodology":27068,"\u002Fdeep-learning\u002Fpractical\u002Fhyperparameters-and-debugging":27073,"\u002Fdeep-learning\u002Fpractical\u002Frepresentation-learning":27077,"\u002Fdeep-learning\u002Fpractical\u002Ftransfer-learning":27081,"\u002Fdeep-learning\u002Fpractical\u002Fapplications":27085,"\u002Fdeep-learning\u002Fpractical\u002Fmodel-compression-and-distillation":27089,"\u002Fdeep-learning\u002Fpractical\u002Fmeta-learning-and-few-shot":27093,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Flarge-language-models":27097,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fscaling-inference-and-alignment":27102,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fseq2seq-pretraining-and-bart":27106,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Ftext-to-text-transfer-and-conditional-generation":27110,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fspeech-and-audio-models":27114,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fself-supervised-speech-and-synthesis":27118,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fai-agents":27122,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fagent-memory-retrieval-and-orchestration":27126,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fmixture-of-experts":27130,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fmultimodal-models":27134,"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Ffusion-and-vision-language-models":27138,"\u002Fdeep-learning\u002Freinforcement-learning\u002Ffoundations-of-reinforcement-learning":27142,"\u002Fdeep-learning\u002Freinforcement-learning\u002Fmodel-free-prediction-and-control":27147,"\u002Fdeep-learning\u002Freinforcement-learning\u002Fdeep-q-networks":27151,"\u002Fdeep-learning\u002Freinforcement-learning\u002Fpolicy-gradients-and-actor-critic":27155,"\u002Fdeep-learning\u002Freinforcement-learning\u002Frl-from-human-feedback":27159,"\u002Fdeep-learning":27163,"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Fequilibrium-state-variables-zeroth-law":27165,"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Ffirst-law-heat-and-work":27169,"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Fsecond-law-entropy-and-the-carnot-bound":27173,"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Fthermodynamic-potentials-and-maxwell-relations":27177,"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Fstability-response-functions-and-the-third-law":27181,"\u002Fstatistical-mechanics\u002Ffoundations\u002Fclassical-statistics-and-equipartition":27185,"\u002Fstatistical-mechanics\u002Ffoundations\u002Fphase-space-and-liouvilles-theorem":27190,"\u002Fstatistical-mechanics\u002Ffoundations\u002Fensembles-and-the-equal-probability-postulate":27194,"\u002Fstatistical-mechanics\u002Ffoundations\u002Fstatistical-entropy-boltzmann-and-gibbs":27198,"\u002Fstatistical-mechanics\u002Fmicrocanonical\u002Fmicrocanonical-ensemble-and-entropy":27202,"\u002Fstatistical-mechanics\u002Fmicrocanonical\u002Fequilibrium-conditions-temperature-pressure-chemical-potential":27207,"\u002Fstatistical-mechanics\u002Fmicrocanonical\u002Fideal-gas-phase-space-and-the-sackur-tetrode-entropy":27211,"\u002Fstatistical-mechanics\u002Fmicrocanonical\u002Ftwo-state-systems-paramagnets-and-negative-temperature":27215,"\u002Fstatistical-mechanics\u002Fcanonical\u002Fcanonical-ensemble-and-the-boltzmann-distribution":27219,"\u002Fstatistical-mechanics\u002Fcanonical\u002Fpartition-function-and-the-helmholtz-free-energy":27224,"\u002Fstatistical-mechanics\u002Fcanonical\u002Fenergy-fluctuations-and-ensemble-equivalence":27228,"\u002Fstatistical-mechanics\u002Fcanonical\u002Fthe-einstein-solid-and-harmonic-systems":27232,"\u002Fstatistical-mechanics\u002Fcanonical\u002Fparamagnetism-and-the-schottky-anomaly":27236,"\u002Fstatistical-mechanics\u002Fclassical-gas\u002Fideal-gas-partition-function-and-the-gibbs-paradox":27240,"\u002Fstatistical-mechanics\u002Fclassical-gas\u002Fequipartition-and-the-virial-theorem":27245,"\u002Fstatistical-mechanics\u002Fclassical-gas\u002Fmolecular-gases-rotation-and-vibration":27249,"\u002Fstatistical-mechanics\u002Fgrand-canonical\u002Fgrand-canonical-ensemble-and-the-grand-partition-function":27253,"\u002Fstatistical-mechanics\u002Fgrand-canonical\u002Fchemical-potential-fugacity-and-number-fluctuations":27258,"\u002Fstatistical-mechanics\u002Fgrand-canonical\u002Fensemble-summary-and-the-thermodynamic-web":27262,"\u002Fstatistical-mechanics\u002Fquantum-statistics\u002Fquantum-statistics-bose-einstein-and-fermi-dirac":27266,"\u002Fstatistical-mechanics\u002Fquantum-statistics\u002Fderiving-the-quantum-distributions":27271,"\u002Fstatistical-mechanics\u002Fquantum-statistics\u002Fthe-classical-limit-and-quantum-concentration":27275,"\u002Fstatistical-mechanics\u002Fquantum-statistics\u002Fideal-quantum-gases-general-framework":27279,"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fbose-einstein-condensation-and-the-fermion-gas":27283,"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fthe-photon-gas-and-plancks-radiation-law":27288,"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fblackbody-thermodynamics-and-radiation-pressure":27292,"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fphonons-and-the-debye-model":27296,"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fbose-einstein-condensation-derived":27300,"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fthermodynamics-of-the-bose-gas-and-superfluidity":27304,"\u002Fstatistical-mechanics\u002Ffermi-gas\u002Fthe-ideal-fermi-gas-at-zero-temperature":27308,"\u002Fstatistical-mechanics\u002Ffermi-gas\u002Fsommerfeld-expansion-and-electrons-in-metals":27313,"\u002Fstatistical-mechanics\u002Ffermi-gas\u002Fwhite-dwarfs-and-the-chandrasekhar-limit":27317,"\u002Fstatistical-mechanics\u002Ffermi-gas\u002Fneutron-stars-and-nuclear-matter":27321,"\u002Fstatistical-mechanics\u002Finteractions\u002Fthe-cluster-expansion-and-virial-coefficients":27325,"\u002Fstatistical-mechanics\u002Finteractions\u002Fthe-van-der-waals-gas-and-liquid-gas-coexistence":27330,"\u002Fstatistical-mechanics\u002Finteractions\u002Fquantum-gases-with-interactions-and-exchange":27334,"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fphases-coexistence-and-classification":27338,"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fthe-ising-model-and-exact-solutions":27343,"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fmean-field-theory-and-the-weiss-model":27347,"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fcritical-exponents-and-landau-theory":27351,"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fthe-renormalization-group-idea":27355,"\u002Fstatistical-mechanics\u002Ffluctuations\u002Fthermodynamic-fluctuations-and-response":27359,"\u002Fstatistical-mechanics\u002Ffluctuations\u002Fbrownian-motion-and-the-langevin-equation":27364,"\u002Fstatistical-mechanics\u002Ffluctuations\u002Flinear-response-and-the-fluctuation-dissipation-theorem":27368,"\u002Fstatistical-mechanics":27372,"\u002Fcondensed-matter\u002Fmolecules-and-bonding\u002Fbonding-mechanisms":27375,"\u002Fcondensed-matter\u002Fmolecules-and-bonding\u002Fmolecular-orbitals-and-h2-plus":27380,"\u002Fcondensed-matter\u002Fmolecules-and-bonding\u002Fhydrogen-molecule-and-exchange":27384,"\u002Fcondensed-matter\u002Fmolecules-and-bonding\u002Fvan-der-waals-forces":27388,"\u002Fcondensed-matter\u002Fmolecular-spectra\u002Frotational-vibrational-spectra":27392,"\u002Fcondensed-matter\u002Fmolecular-spectra\u002Fanharmonicity-and-rovibrational-structure":27397,"\u002Fcondensed-matter\u002Fmolecular-spectra\u002Framan-and-electronic-bands":27401,"\u002Fcondensed-matter\u002Fmolecular-spectra\u002Flasers-and-masers":27405,"\u002Fcondensed-matter\u002Fcrystal-structure\u002Fstructure-of-solids":27409,"\u002Fcondensed-matter\u002Fcrystal-structure\u002Fbravais-lattices-and-crystal-systems":27414,"\u002Fcondensed-matter\u002Fcrystal-structure\u002Freciprocal-lattice-and-brillouin-zones":27418,"\u002Fcondensed-matter\u002Fcrystal-structure\u002Fdiffraction-and-structure-factors":27422,"\u002Fcondensed-matter\u002Flattice-dynamics\u002Fphonon-dispersion":27426,"\u002Fcondensed-matter\u002Flattice-dynamics\u002Fphonons-quantization-and-dos":27431,"\u002Fcondensed-matter\u002Flattice-dynamics\u002Fdebye-einstein-heat-capacity":27435,"\u002Fcondensed-matter\u002Flattice-dynamics\u002Fanharmonicity-and-thermal-transport":27439,"\u002Fcondensed-matter\u002Ffree-electron-fermi-gas\u002Ffree-electron-gas-and-conduction":27443,"\u002Fcondensed-matter\u002Ffree-electron-fermi-gas\u002Fsommerfeld-model-and-heat-capacity":27448,"\u002Fcondensed-matter\u002Ffree-electron-fermi-gas\u002Ftransport-and-the-hall-effect":27452,"\u002Fcondensed-matter\u002Ffree-electron-fermi-gas\u002Fscreening-and-plasmons":27456,"\u002Fcondensed-matter\u002Fband-theory\u002Fblochs-theorem-and-energy-bands":27460,"\u002Fcondensed-matter\u002Fband-theory\u002Fnearly-free-electron-model":27465,"\u002Fcondensed-matter\u002Fband-theory\u002Ftight-binding-method":27469,"\u002Fcondensed-matter\u002Fband-theory\u002Ffermi-surfaces-and-semiclassical-dynamics":27473,"\u002Fcondensed-matter\u002Fsemiconductors\u002Fsemiconductor-bands-and-junctions":27477,"\u002Fcondensed-matter\u002Fsemiconductors\u002Fintrinsic-and-extrinsic-semiconductors":27482,"\u002Fcondensed-matter\u002Fsemiconductors\u002Fcarrier-transport-and-recombination":27486,"\u002Fcondensed-matter\u002Fsemiconductors\u002Fthe-pn-junction":27490,"\u002Fcondensed-matter\u002Fsemiconductors\u002Ftransistors-and-optoelectronics":27494,"\u002Fcondensed-matter\u002Fdielectrics-and-ferroelectrics\u002Fdielectrics-and-polarization":27498,"\u002Fcondensed-matter\u002Fdielectrics-and-ferroelectrics\u002Fferroelectrics-and-piezoelectrics":27503,"\u002Fcondensed-matter\u002Fmagnetism\u002Fdiamagnetism-and-paramagnetism":27507,"\u002Fcondensed-matter\u002Fmagnetism\u002Fexchange-and-ferromagnetism":27512,"\u002Fcondensed-matter\u002Fmagnetism\u002Fantiferromagnetism-and-domains":27516,"\u002Fcondensed-matter\u002Fmagnetism\u002Fspin-waves-and-magnons":27520,"\u002Fcondensed-matter\u002Fsuperconductivity\u002Fsuperconductivity-phenomenology":27524,"\u002Fcondensed-matter\u002Fsuperconductivity\u002Flondon-theory-and-the-meissner-effect":27529,"\u002Fcondensed-matter\u002Fsuperconductivity\u002Fginzburg-landau-theory":27533,"\u002Fcondensed-matter\u002Fsuperconductivity\u002Fbcs-theory":27537,"\u002Fcondensed-matter\u002Fsuperconductivity\u002Fjosephson-and-high-tc":27541,"\u002Fcondensed-matter\u002Fnanostructures\u002Fquantum-wells-wires-and-dots":27545,"\u002Fcondensed-matter\u002Fnanostructures\u002Finteger-quantum-hall-effect":27550,"\u002Fcondensed-matter\u002Fnanostructures\u002Ffractional-quantum-hall-and-topology":27554,"\u002Fcondensed-matter\u002Fnanostructures\u002Fgraphene-and-dirac-materials":27558,"\u002Fcondensed-matter":27562,"\u002Flogic\u002Ffoundations\u002Flogic-as-a-mathematical-model":27565,"\u002Flogic\u002Fsentential-logic\u002Fformal-languages-and-well-formed-formulas":27569,"\u002Flogic\u002Fsentential-logic\u002Ftruth-assignments-and-tautologies":27574,"\u002Flogic\u002Fsentential-logic\u002Funique-readability-and-parsing":27578,"\u002Flogic\u002Fsentential-logic\u002Finduction-and-recursion":27582,"\u002Flogic\u002Fsentential-logic\u002Fexpressive-completeness-and-normal-forms":27586,"\u002Flogic\u002Fsentential-logic\u002Fboolean-circuits":27590,"\u002Flogic\u002Fsentential-logic\u002Fcompactness-and-effectiveness":27594,"\u002Flogic\u002Ffirst-order-languages\u002Ffirst-order-languages":27598,"\u002Flogic\u002Ffirst-order-languages\u002Fstructures-truth-and-satisfaction":27603,"\u002Flogic\u002Ffirst-order-languages\u002Fdefinability-and-elementary-equivalence":27607,"\u002Flogic\u002Ffirst-order-languages\u002Fterms-substitution-and-parsing":27611,"\u002Flogic\u002Fdeductive-calculus\u002Fa-deductive-calculus":27615,"\u002Flogic\u002Fdeductive-calculus\u002Fdeduction-theorem-and-derived-rules":27620,"\u002Flogic\u002Fdeductive-calculus\u002Fsoundness":27624,"\u002Flogic\u002Fdeductive-calculus\u002Fcompleteness-and-consistency":27628,"\u002Flogic\u002Fmodels-and-theories\u002Fcompactness-and-lowenheim-skolem":27632,"\u002Flogic\u002Fmodels-and-theories\u002Ftheories-elementary-classes-and-categoricity":27637,"\u002Flogic\u002Fmodels-and-theories\u002Finterpretations-between-theories":27641,"\u002Flogic\u002Fmodels-and-theories\u002Fnonstandard-analysis":27645,"\u002Flogic\u002Farithmetic-and-definability\u002Fdefinability-in-arithmetic":27649,"\u002Flogic\u002Farithmetic-and-definability\u002Fnatural-numbers-with-successor":27654,"\u002Flogic\u002Farithmetic-and-definability\u002Fpresburger-and-reducts":27658,"\u002Flogic\u002Farithmetic-and-definability\u002Fa-subtheory-and-representability":27662,"\u002Flogic\u002Fincompleteness\u002Farithmetization-of-syntax":27666,"\u002Flogic\u002Fincompleteness\u002Fincompleteness-and-undecidability":27671,"\u002Flogic\u002Fincompleteness\u002Fsecond-incompleteness-theorem":27675,"\u002Flogic\u002Fcomputability-and-representability\u002Frecursive-functions":27679,"\u002Flogic\u002Fcomputability-and-representability\u002Frepresenting-exponentiation":27684,"\u002Flogic\u002Fsecond-order-logic\u002Fsecond-order-languages":27688,"\u002Flogic\u002Fsecond-order-logic\u002Fskolem-functions-and-many-sorted-logic":27693,"\u002Flogic\u002Fsecond-order-logic\u002Fgeneral-structures":27697,"\u002Flogic":27701,"\u002Freinforcement-learning\u002Ffoundations\u002Fwhat-is-reinforcement-learning":27704,"\u002Freinforcement-learning\u002Ffoundations\u002Fa-brief-history-of-rl":27708,"\u002Freinforcement-learning\u002Ffoundations\u002Fmulti-armed-bandits":27712,"\u002Freinforcement-learning\u002Ffoundations\u002Fbandit-exploration-algorithms":27716,"\u002Freinforcement-learning\u002Ffoundations\u002Fmarkov-decision-processes":27720,"\u002Freinforcement-learning\u002Ffoundations\u002Fvalue-functions-and-optimality":27724,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fdynamic-programming":27728,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fdp-async-and-gpi":27732,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fmonte-carlo-methods":27736,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fmonte-carlo-off-policy":27740,"\u002Freinforcement-learning\u002Ftabular-methods\u002Ftemporal-difference-learning":27744,"\u002Freinforcement-learning\u002Ftabular-methods\u002Ftd-control-sarsa-and-q-learning":27748,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fn-step-bootstrapping":27752,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fn-step-off-policy-methods":27756,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fplanning-and-learning":27760,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fplanning-focusing-and-decision-time":27764,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fdecision-time-planning":27768,"\u002Freinforcement-learning\u002Ftabular-methods\u002Fmonte-carlo-tree-search":27772,"\u002Freinforcement-learning\u002Fapproximation\u002Fon-policy-prediction":27776,"\u002Freinforcement-learning\u002Fapproximation\u002Ffeature-construction-and-nonlinear":27781,"\u002Freinforcement-learning\u002Fapproximation\u002Fon-policy-control":27785,"\u002Freinforcement-learning\u002Fapproximation\u002Faverage-reward-control":27789,"\u002Freinforcement-learning\u002Fapproximation\u002Foff-policy-and-the-deadly-triad":27793,"\u002Freinforcement-learning\u002Fapproximation\u002Fbellman-error-and-gradient-td":27797,"\u002Freinforcement-learning\u002Fapproximation\u002Feligibility-traces":27801,"\u002Freinforcement-learning\u002Fapproximation\u002Ftrue-online-and-sarsa-lambda":27805,"\u002Freinforcement-learning\u002Fapproximation\u002Fpolicy-gradient-methods":27809,"\u002Freinforcement-learning\u002Fapproximation\u002Factor-critic-and-continuous-actions":27813,"\u002Freinforcement-learning\u002Fapproximation\u002Fleast-squares-and-memory-based-methods":27817,"\u002Freinforcement-learning\u002Fapproximation\u002Fmemory-and-kernel-methods":27821,"\u002Freinforcement-learning\u002Fapproximation\u002Foff-policy-eligibility-traces":27825,"\u002Freinforcement-learning\u002Fapproximation\u002Fstable-off-policy-traces":27829,"\u002Freinforcement-learning\u002Fdeep-rl\u002Fdeep-q-networks":27833,"\u002Freinforcement-learning\u002Fdeep-rl\u002Fdqn-improvements":27837,"\u002Freinforcement-learning\u002Fdeep-rl\u002Factor-critic-and-ppo":27841,"\u002Freinforcement-learning\u002Fdeep-rl\u002Fppo-and-continuous-control":27845,"\u002Freinforcement-learning\u002Fdeep-rl\u002Fcase-studies":27849,"\u002Freinforcement-learning\u002Fdeep-rl\u002Frl-beyond-games":27853,"\u002Freinforcement-learning\u002Fdeep-rl\u002Ffrontiers":27857,"\u002Freinforcement-learning\u002Fdeep-rl\u002Freward-design-and-open-problems":27861,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fdistributional-and-rainbow":27865,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fdistributional-and-rainbow-part-2":27870,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fcontinuous-control":27874,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fcontinuous-control-part-2":27878,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmodel-based-rl":27882,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmodel-based-rl-part-2":27886,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fexploration":27890,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fexploration-part-2":27894,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Foffline-rl":27898,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Foffline-rl-part-2":27902,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fimitation-and-inverse-rl":27906,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fimitation-and-inverse-rl-part-2":27910,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmulti-agent-rl":27914,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmulti-agent-rl-part-2":27918,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fhierarchical-rl":27922,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fhierarchical-rl-part-2":27926,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Frlhf-and-language-models":27930,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fpartial-observability-pomdps":27934,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fpartial-observability-pomdps-part-2":27938,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fsafe-and-constrained-rl":27942,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fsafe-and-constrained-rl-part-2":27946,"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmeta-rl-and-generalization":27950,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fpsychology-of-reinforcement":27954,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Finstrumental-conditioning-and-control":27959,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fdopamine-and-td-error":27963,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fdopamine-in-the-brain":27967,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fanimal-learning-and-cognition":27971,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fcognitive-maps-and-planning":27975,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fneuroscience-of-reinforcement":27979,"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fseveral-learning-systems":27983,"\u002Freinforcement-learning":27987,"\u002Fartificial-intelligence\u002Ffoundations\u002Fwhat-is-ai":27989,"\u002Fartificial-intelligence\u002Ffoundations\u002Ffoundations-of-ai":27993,"\u002Fartificial-intelligence\u002Ffoundations\u002Fintelligent-agents":27997,"\u002Fartificial-intelligence\u002Ffoundations\u002Fagent-architectures":28001,"\u002Fartificial-intelligence\u002Fsearch\u002Funinformed-search":28005,"\u002Fartificial-intelligence\u002Fsearch\u002Fsearch-strategies-compared":28010,"\u002Fartificial-intelligence\u002Fsearch\u002Finformed-search":28014,"\u002Fartificial-intelligence\u002Fsearch\u002Fheuristic-functions":28018,"\u002Fartificial-intelligence\u002Fsearch\u002Flocal-search":28022,"\u002Fartificial-intelligence\u002Fsearch\u002Fpopulation-and-continuous-search":28026,"\u002Fartificial-intelligence\u002Fsearch\u002Fadversarial-search":28030,"\u002Fartificial-intelligence\u002Fsearch\u002Fgames-of-chance-and-imperfect-information":28034,"\u002Fartificial-intelligence\u002Fsearch\u002Fconstraint-satisfaction":28038,"\u002Fartificial-intelligence\u002Fsearch\u002Fcsp-search-and-structure":28042,"\u002Fartificial-intelligence\u002Fsearch\u002Fsearch-under-uncertainty":28046,"\u002Fartificial-intelligence\u002Fsearch\u002Fbelief-state-and-online-search":28050,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fpropositional-logic":28054,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fpropositional-inference":28059,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Ffirst-order-logic":28063,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Ffirst-order-logic-in-use":28067,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Finference-and-resolution":28071,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Ffirst-order-resolution":28075,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fclassical-planning":28079,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fplanning-graphs-and-graphplan":28083,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fplanning-in-the-real-world":28087,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fplanning-under-uncertainty":28091,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fknowledge-representation":28095,"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Freasoning-systems-and-defaults":28099,"\u002Fartificial-intelligence\u002Funcertainty\u002Fprobability-and-bayes":28103,"\u002Fartificial-intelligence\u002Funcertainty\u002Fbayes-rule-and-naive-bayes":28108,"\u002Fartificial-intelligence\u002Funcertainty\u002Fbayesian-networks":28112,"\u002Fartificial-intelligence\u002Funcertainty\u002Finference-in-bayesian-networks":28116,"\u002Fartificial-intelligence\u002Funcertainty\u002Freasoning-over-time":28120,"\u002Fartificial-intelligence\u002Funcertainty\u002Ftracking-and-data-association":28124,"\u002Fartificial-intelligence\u002Funcertainty\u002Fmaking-decisions":28128,"\u002Fartificial-intelligence\u002Funcertainty\u002Fmarkov-decision-processes":28132,"\u002Fartificial-intelligence\u002Funcertainty\u002Fdecision-networks-and-game-theory":28135,"\u002Fartificial-intelligence\u002Funcertainty\u002Fgame-theory-and-mechanism-design":28139,"\u002Fartificial-intelligence\u002Flearning\u002Flearning-from-examples":28143,"\u002Fartificial-intelligence\u002Flearning\u002Ftheory-and-model-families":28148,"\u002Fartificial-intelligence\u002Flearning\u002Fprobabilistic-learning":28152,"\u002Fartificial-intelligence\u002Flearning\u002Fexpectation-maximization":28156,"\u002Fartificial-intelligence\u002Flearning\u002Freinforcement-learning":28160,"\u002Fartificial-intelligence\u002Flearning\u002Fgeneralization-and-policy-search":28163,"\u002Fartificial-intelligence\u002Flearning\u002Fknowledge-in-learning":28167,"\u002Fartificial-intelligence\u002Flearning\u002Fknowledge-based-learning-methods":28171,"\u002Fartificial-intelligence\u002Ffrontiers\u002Fvision-and-perception":28175,"\u002Fartificial-intelligence\u002Ffrontiers\u002Freconstructing-the-3d-world":28180,"\u002Fartificial-intelligence\u002Ffrontiers\u002Frobotics":28184,"\u002Fartificial-intelligence\u002Ffrontiers\u002Frobot-planning-and-control":28188,"\u002Fartificial-intelligence\u002Ffrontiers\u002Fnatural-language-in-ai":28192,"\u002Fartificial-intelligence\u002Ffrontiers\u002Fnlp-grammar-translation-and-speech":28196,"\u002Fartificial-intelligence\u002Ffrontiers\u002Fphilosophy-and-future":28200,"\u002Fartificial-intelligence\u002Ffrontiers\u002Fai-ethics-and-future":28204,"\u002Fartificial-intelligence":28208,"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fnuclear-constituents-nuclide-chart":28211,"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fnuclear-size-charge-distributions":28216,"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fnuclear-masses-binding-energy":28220,"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fsemi-empirical-mass-formula":28224,"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fnuclear-moments-multipoles":28228,"\u002Fnuclear-physics\u002Fnuclear-force-deuteron\u002Fnuclear-force-shell-overview":28232,"\u002Fnuclear-physics\u002Fnuclear-force-deuteron\u002Fthe-deuteron":28237,"\u002Fnuclear-physics\u002Fnuclear-force-deuteron\u002Fnucleon-nucleon-scattering":28241,"\u002Fnuclear-physics\u002Fnuclear-force-deuteron\u002Fmeson-theory-isospin":28245,"\u002Fnuclear-physics\u002Fnuclear-models\u002Ffermi-gas-model":28249,"\u002Fnuclear-physics\u002Fnuclear-models\u002Fliquid-drop-collective-coordinates":28254,"\u002Fnuclear-physics\u002Fnuclear-models\u002Fshell-model-single-particle":28258,"\u002Fnuclear-physics\u002Fnuclear-models\u002Fcollective-model-rotations-vibrations":28262,"\u002Fnuclear-physics\u002Fradioactive-decay\u002Fdecay-law-modes":28266,"\u002Fnuclear-physics\u002Fradioactive-decay\u002Fdecay-kinetics-equilibrium":28271,"\u002Fnuclear-physics\u002Falpha-decay\u002Falpha-decay-gamow-theory":28275,"\u002Fnuclear-physics\u002Falpha-decay\u002Falpha-fine-structure-hindrance":28280,"\u002Fnuclear-physics\u002Fbeta-decay\u002Fbeta-decay-energetics-neutrino":28284,"\u002Fnuclear-physics\u002Fbeta-decay\u002Ffermi-theory-beta-decay":28289,"\u002Fnuclear-physics\u002Fbeta-decay\u002Fweak-interaction-parity-violation":28293,"\u002Fnuclear-physics\u002Fbeta-decay\u002Fdouble-beta-decay-neutrino-mass":28297,"\u002Fnuclear-physics\u002Fgamma-decay\u002Fgamma-multipole-radiation":28301,"\u002Fnuclear-physics\u002Fgamma-decay\u002Finternal-conversion-isomers":28306,"\u002Fnuclear-physics\u002Fgamma-decay\u002Fangular-correlations-mossbauer":28310,"\u002Fnuclear-physics\u002Fnuclear-reactions\u002Freaction-kinematics-cross-sections":28314,"\u002Fnuclear-physics\u002Fnuclear-reactions\u002Fcompound-nucleus-resonances":28319,"\u002Fnuclear-physics\u002Fnuclear-reactions\u002Fdirect-reactions-optical-model":28323,"\u002Fnuclear-physics\u002Ffission\u002Ffission-barrier-dynamics":28327,"\u002Fnuclear-physics\u002Ffission\u002Fchain-reactions-reactor-physics":28332,"\u002Fnuclear-physics\u002Ffusion-nucleosynthesis\u002Ffusion-reactions-confinement":28336,"\u002Fnuclear-physics\u002Ffusion-nucleosynthesis\u002Fstellar-nucleosynthesis":28341,"\u002Fnuclear-physics\u002Ffusion-nucleosynthesis\u002Fbig-bang-nucleosynthesis":28345,"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fcharged-particle-stopping-power":28349,"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fphoton-neutron-interactions":28354,"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fradiation-detectors":28358,"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fdosimetry-radiation-biology":28362,"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fnuclear-applications-dating-medicine":28366,"\u002Fnuclear-physics":28370,"\u002Fnatural-language-processing\u002Ffoundations\u002Fwhat-is-nlp":28373,"\u002Fnatural-language-processing\u002Ffoundations\u002Fregex-and-text-normalization":28377,"\u002Fnatural-language-processing\u002Ffoundations\u002Fminimum-edit-distance":28381,"\u002Fnatural-language-processing\u002Ffoundations\u002Fn-gram-language-models":28385,"\u002Fnatural-language-processing\u002Ffoundations\u002Fsmoothing-and-backoff":28389,"\u002Fnatural-language-processing\u002Fclassification\u002Fnaive-bayes-and-sentiment":28393,"\u002Fnatural-language-processing\u002Fclassification\u002Fevaluating-classifiers":28398,"\u002Fnatural-language-processing\u002Fclassification\u002Flogistic-regression":28402,"\u002Fnatural-language-processing\u002Fclassification\u002Fsentiment-and-affect-lexicons":28406,"\u002Fnatural-language-processing\u002Fsemantics\u002Fvector-semantics-and-embeddings":28410,"\u002Fnatural-language-processing\u002Fsemantics\u002Fstatic-word-embeddings":28415,"\u002Fnatural-language-processing\u002Fsemantics\u002Fneural-language-models":28419,"\u002Fnatural-language-processing\u002Fsequences\u002Fsequence-labeling":28423,"\u002Fnatural-language-processing\u002Fsequences\u002Fcrfs-and-neural-taggers":28427,"\u002Fnatural-language-processing\u002Fsequences\u002Frnns-and-lstms":28431,"\u002Fnatural-language-processing\u002Ftransformers\u002Ftransformers-and-attention":28435,"\u002Fnatural-language-processing\u002Ftransformers\u002Fthe-transformer-architecture":28439,"\u002Fnatural-language-processing\u002Ftransformers\u002Flarge-language-models":28442,"\u002Fnatural-language-processing\u002Ftransformers\u002Fllm-pretraining-and-scaling":28445,"\u002Fnatural-language-processing\u002Ftransformers\u002Ffine-tuning-and-prompting":28449,"\u002Fnatural-language-processing\u002Ftransformers\u002Fprompting-and-alignment":28453,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fconstituency-parsing":28457,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fcky-scoring-and-evaluation":28462,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fdependency-parsing":28466,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fgraph-based-and-neural-dependency-parsing":28470,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fword-senses-and-wsd":28474,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fwsd-in-practice-and-induction":28478,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fsemantic-roles-and-information-extraction":28482,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Frelations-events-and-templates":28486,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fcoreference-and-discourse":28490,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fcoherence-and-discourse-structure":28494,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Flogical-semantics":28498,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fcompositional-semantics-and-description-logics":28502,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fsemantic-parsing":28506,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fneural-semantic-parsing":28510,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Finformation-extraction":28514,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Ftimes-events-and-templates":28518,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fdiscourse-coherence":28522,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fentity-based-and-global-coherence":28526,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fconstituency-grammars":28530,"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Ftreebanks-and-lexicalized-grammars":28534,"\u002Fnatural-language-processing\u002Fapplications\u002Fmachine-translation":28538,"\u002Fnatural-language-processing\u002Fapplications\u002Fmachine-translation-decoding-and-evaluation":28542,"\u002Fnatural-language-processing\u002Fapplications\u002Fquestion-answering":28546,"\u002Fnatural-language-processing\u002Fapplications\u002Fquestion-answering-knowledge-and-llms":28550,"\u002Fnatural-language-processing\u002Fapplications\u002Fdialogue-and-chatbots":28554,"\u002Fnatural-language-processing\u002Fapplications\u002Fdialogue-systems-and-assistants":28558,"\u002Fnatural-language-processing\u002Fapplications\u002Ftext-summarization":28562,"\u002Fnatural-language-processing\u002Fapplications\u002Fabstractive-summarization-and-evaluation":28566,"\u002Fnatural-language-processing\u002Fspeech\u002Fphonetics":28570,"\u002Fnatural-language-processing\u002Fspeech\u002Facoustic-phonetics":28575,"\u002Fnatural-language-processing\u002Fspeech\u002Fautomatic-speech-recognition":28579,"\u002Fnatural-language-processing\u002Fspeech\u002Fasr-evaluation-and-applications":28583,"\u002Fnatural-language-processing":28587,"\u002Fparticle-physics\u002Ffoundations\u002Fhistorical-overview-particle-zoo":28590,"\u002Fparticle-physics\u002Ffoundations\u002Fparticle-physics-basic-concepts":28594,"\u002Fparticle-physics\u002Ffoundations\u002Ffundamental-interactions-force-carriers":28598,"\u002Fparticle-physics\u002Funits-kinematics\u002Fnatural-units-and-scales":28602,"\u002Fparticle-physics\u002Funits-kinematics\u002Ffour-vectors-invariant-mass":28607,"\u002Fparticle-physics\u002Funits-kinematics\u002Fdecay-scattering-kinematics-mandelstam":28611,"\u002Fparticle-physics\u002Funits-kinematics\u002Fcross-sections-golden-rule":28615,"\u002Fparticle-physics\u002Fsymmetries\u002Fconservation-laws-symmetries":28619,"\u002Fparticle-physics\u002Fsymmetries\u002Fdiscrete-symmetries-cpt":28624,"\u002Fparticle-physics\u002Fsymmetries\u002Fparity-violation-weak":28628,"\u002Fparticle-physics\u002Fsymmetries\u002Fsu2-su3-flavor-symmetry":28632,"\u002Fparticle-physics\u002Fquark-model\u002Feightfold-way-su3":28636,"\u002Fparticle-physics\u002Fquark-model\u002Fmeson-spectroscopy":28641,"\u002Fparticle-physics\u002Fquark-model\u002Fbaryon-spectroscopy":28645,"\u002Fparticle-physics\u002Fquark-model\u002Fcolor-confinement-exotics":28649,"\u002Fparticle-physics\u002Frelativistic-wave-equations\u002Fklein-gordon-equation":28653,"\u002Fparticle-physics\u002Frelativistic-wave-equations\u002Fdirac-equation-spinors":28658,"\u002Fparticle-physics\u002Frelativistic-wave-equations\u002Fantiparticles-hole-theory":28662,"\u002Fparticle-physics\u002Fqed\u002Ffeynman-rules-qed":28666,"\u002Fparticle-physics\u002Fqed\u002Fqed-tree-processes":28671,"\u002Fparticle-physics\u002Fqed\u002Frenormalization-running-coupling":28675,"\u002Fparticle-physics\u002Fqed\u002Felectron-g-2":28679,"\u002Fparticle-physics\u002Fweak-interaction\u002Fva-structure-weak":28683,"\u002Fparticle-physics\u002Fweak-interaction\u002Fw-z-bosons-decays":28688,"\u002Fparticle-physics\u002Fweak-interaction\u002Fckm-matrix":28692,"\u002Fparticle-physics\u002Fweak-interaction\u002Fcp-violation-kaons-b-mesons":28696,"\u002Fparticle-physics\u002Fqcd\u002Fcolor-su3-gluons":28700,"\u002Fparticle-physics\u002Fqcd\u002Fasymptotic-freedom-confinement":28705,"\u002Fparticle-physics\u002Fqcd\u002Fdeep-inelastic-scattering-partons":28709,"\u002Fparticle-physics\u002Fqcd\u002Fjets-hadronization":28713,"\u002Fparticle-physics\u002Felectroweak-higgs\u002Felectroweak-su2-u1":28717,"\u002Fparticle-physics\u002Felectroweak-higgs\u002Fspontaneous-symmetry-breaking":28722,"\u002Fparticle-physics\u002Felectroweak-higgs\u002Fhiggs-mechanism":28726,"\u002Fparticle-physics\u002Felectroweak-higgs\u002Fhiggs-boson-discovery":28730,"\u002Fparticle-physics\u002Felectroweak-higgs\u002Fstandard-model":28734,"\u002Fparticle-physics\u002Fneutrinos\u002Fneutrino-oscillations":28738,"\u002Fparticle-physics\u002Fneutrinos\u002Fneutrino-mass-pmns":28743,"\u002Fparticle-physics\u002Fneutrinos\u002Fdirac-majorana-experiments":28747,"\u002Fparticle-physics\u002Fexperiment\u002Faccelerators-luminosity":28751,"\u002Fparticle-physics\u002Fexperiment\u002Fdetectors-subsystems":28756,"\u002Fparticle-physics\u002Fexperiment\u002Fhow-discoveries-are-made":28760,"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fbeyond-standard-model":28764,"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fgrand-unified-theories":28768,"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fsupersymmetry":28772,"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fhierarchy-problem-naturalness":28776,"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fdark-matter-candidates":28780,"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fmatter-antimatter-open-questions":28784,"\u002Fparticle-physics":28788,"\u002Fastrophysics-cosmology\u002Forientation\u002Fthe-sun-and-stars":28791,"\u002Fastrophysics-cosmology\u002Forientation\u002Fstellar-death-final-states":28796,"\u002Fastrophysics-cosmology\u002Forientation\u002Fgalaxies-and-cosmology":28800,"\u002Fastrophysics-cosmology\u002Fobservational-foundations\u002Fmagnitudes-fluxes-and-the-distance-modulus":28804,"\u002Fastrophysics-cosmology\u002Fobservational-foundations\u002Fstellar-spectra-and-spectral-classification":28809,"\u002Fastrophysics-cosmology\u002Fobservational-foundations\u002Ftelescopes-and-detectors-across-the-spectrum":28813,"\u002Fastrophysics-cosmology\u002Fobservational-foundations\u002Fthe-cosmic-distance-ladder":28817,"\u002Fastrophysics-cosmology\u002Fradiation-and-matter\u002Fblackbody-radiation-and-specific-intensity":28821,"\u002Fastrophysics-cosmology\u002Fradiation-and-matter\u002Fradiative-transfer-and-the-transfer-equation":28826,"\u002Fastrophysics-cosmology\u002Fradiation-and-matter\u002Fspectral-line-formation-and-broadening":28830,"\u002Fastrophysics-cosmology\u002Fradiation-and-matter\u002Fopacity-and-the-rosseland-mean":28834,"\u002Fastrophysics-cosmology\u002Fstellar-structure\u002Fhydrostatic-equilibrium-and-the-virial-theorem":28838,"\u002Fastrophysics-cosmology\u002Fstellar-structure\u002Fthe-equations-of-stellar-structure":28843,"\u002Fastrophysics-cosmology\u002Fstellar-structure\u002Fthe-equation-of-state-and-polytropes":28847,"\u002Fastrophysics-cosmology\u002Fstellar-structure\u002Fthe-standard-solar-model":28851,"\u002Fastrophysics-cosmology\u002Fnuclear-astrophysics\u002Fthermonuclear-reaction-rates-and-the-gamow-peak":28855,"\u002Fastrophysics-cosmology\u002Fnuclear-astrophysics\u002Fhydrogen-burning-pp-chains-and-cno":28860,"\u002Fastrophysics-cosmology\u002Fnuclear-astrophysics\u002Fhelium-burning-and-the-triple-alpha-process":28864,"\u002Fastrophysics-cosmology\u002Fnuclear-astrophysics\u002Fadvanced-burning-and-neutron-capture-nucleosynthesis":28868,"\u002Fastrophysics-cosmology\u002Fism-and-star-formation\u002Fphases-of-the-interstellar-medium":28872,"\u002Fastrophysics-cosmology\u002Fism-and-star-formation\u002Fmolecular-clouds-and-gravitational-collapse":28877,"\u002Fastrophysics-cosmology\u002Fism-and-star-formation\u002Fprotostars-and-the-pre-main-sequence":28881,"\u002Fastrophysics-cosmology\u002Fstellar-evolution\u002Fthe-main-sequence-and-its-structure":28885,"\u002Fastrophysics-cosmology\u002Fstellar-evolution\u002Fpost-main-sequence-low-mass-evolution":28890,"\u002Fastrophysics-cosmology\u002Fstellar-evolution\u002Fthe-evolution-of-massive-stars":28894,"\u002Fastrophysics-cosmology\u002Fstellar-evolution\u002Fstellar-pulsation-and-the-instability-strip":28898,"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fwhite-dwarfs-and-the-chandrasekhar-limit":28902,"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fcore-collapse-supernovae":28906,"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fthermonuclear-supernovae-type-ia":28910,"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fneutron-stars-and-pulsars":28914,"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fblack-holes-schwarzschild-and-kerr":28918,"\u002Fastrophysics-cosmology\u002Fbinaries-and-gravitational-waves\u002Fbinary-systems-and-mass-transfer":28922,"\u002Fastrophysics-cosmology\u002Fbinaries-and-gravitational-waves\u002Faccreting-compact-objects":28927,"\u002Fastrophysics-cosmology\u002Fbinaries-and-gravitational-waves\u002Fgravitational-waves-from-inspiraling-binaries":28931,"\u002Fastrophysics-cosmology\u002Fbinaries-and-gravitational-waves\u002Fmultimessenger-astronomy-and-gamma-ray-bursts":28935,"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Fthe-milky-way":28939,"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Fgalaxy-morphology-and-classification":28944,"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Fgalaxy-rotation-curves-and-dark-matter":28948,"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Factive-galactic-nuclei-and-supermassive-black-holes":28952,"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Fgalaxy-clusters-and-large-scale-structure":28956,"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fthe-expanding-universe-and-hubbles-law":28960,"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fthe-frw-metric-and-cosmological-redshift":28965,"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fthe-friedmann-equations-and-cosmic-dynamics":28969,"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fcosmological-models-and-distances":28972,"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fdark-energy-and-the-accelerating-universe":28976,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fthe-thermal-history-of-the-universe":28980,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fbig-bang-nucleosynthesis":28985,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Frecombination-and-the-cosmic-microwave-background":28989,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fcmb-anisotropies-and-cosmological-parameters":28993,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fcosmic-inflation":28997,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fstructure-formation-and-the-growth-of-perturbations":29001,"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fdark-matter-dark-energy-and-open-questions":29005,"\u002Fastrophysics-cosmology":29009,"\u002Fcolophon":29012,"\u002F":29015},{"path":24338,"title":24339,"module":24340,"summary":24341},"\u002Falgorithms\u002Ffoundations\u002Fwhat-is-an-algorithm","What Is an Algorithm?","Foundations","An algorithm is a finite, mechanical recipe that transforms inputs into outputs. We define what counts as an algorithm, how we write one down, and the three things we always ask of it: is it correct, is it fast, and can we prove it.\n",{"path":24343,"title":24344,"module":24340,"summary":24345},"\u002Falgorithms\u002Ffoundations\u002Fproof-techniques","Proof Techniques","An algorithm without a proof is a conjecture. This lesson collects the handful\nof arguments that certify the algorithms in this course — direct proof,\ncontrapositive, contradiction, ordinary and strong induction, construction, and\ndisproof by counterexample — each with a small worked\nexample and a picture. Loop invariants are a form of induction,\nrecursive correctness falls to strong induction, and the classic broken proofs\n(all horses are the same color) show where inductions go wrong.\n",{"path":24347,"title":24348,"module":24340,"summary":24349},"\u002Falgorithms\u002Ffoundations\u002Fasymptotic-analysis","Asymptotic Analysis","We measure an algorithm's running time as a function of its input size, then strip away machine-specific constants and lower-order terms to compare algorithms cleanly. This lesson defines the RAM model and the $O$, $\\Omega$, $\\Theta$, $o$, and $\\omega$ notations, proves the polynomial theorem, and shows how to rank growth rates with the limit test, L'Hôpital, base substitution, and the logarithm identities the arguments lean on.\n",{"path":24351,"title":24352,"module":24340,"summary":24353},"\u002Falgorithms\u002Ffoundations\u002Fgrowth-rates-and-loop-analysis","Growth Rates and Loop Analysis","With the asymptotic notations in hand, we rank the functions that actually arise in running times — from constant to factorial — proving the orderings between rungs, then read the running time of a loop nest straight off the page. Sequential blocks add, nested loops multiply, index scaling gives logarithms; a worked trace and a tour of cache-aware and galactic algorithms close the lesson.\n",{"path":24355,"title":24356,"module":24340,"summary":24357},"\u002Falgorithms\u002Ffoundations\u002Frecurrences","Recurrences and the Master Theorem","Recursive and divide-and-conquer algorithms describe their own running time with a recurrence: $T(n)$ in terms of $T$ on smaller inputs. We solve recurrences three ways — drawing the recursion tree, guessing-and-verifying by induction, and applying the Master Theorem — using merge sort as the running example, then handle unequal splits with Akra–Bazzi.\n",{"path":24359,"title":24360,"module":24340,"summary":24361},"\u002Falgorithms\u002Ffoundations\u002Famortized-analysis","Amortized Analysis","Some operations are occasionally expensive but cheap on average across any\nsequence. Amortized analysis bounds the average cost per operation over a\nworst-case sequence — not an expectation — so a rare costly step is paid for by\nthe many cheap ones around it. This lesson develops the aggregate, accounting,\nand potential methods on dynamic-array doubling, the binary counter, and a\nstack with multipop.\n",{"path":24363,"title":24364,"module":24365,"summary":24366},"\u002Falgorithms\u002Fdivide-and-conquer\u002Fmergesort","Divide and Conquer & Mergesort","Divide & Conquer","Divide and conquer breaks a problem into smaller copies of itself, solves\nthem recursively, and stitches the answers together. We meet the paradigm\nthrough mergesort — its merge step, its loop-invariant proof, and the\nrecursion tree that pins its cost at $\\Theta(n\\log n)$ — then count inversions\nwith the same machinery and distill the whole pattern into the master theorem.\n",{"path":24368,"title":24369,"module":24365,"summary":24370},"\u002Falgorithms\u002Fdivide-and-conquer\u002Fquicksort","Quicksort","Quicksort sorts in place by partitioning around a pivot and recursing on\neach side. We give Lomuto and Hoare partitioning with a correctness\ninvariant, see why a bad pivot costs $\\Theta(n^2)$ while a balanced one gives\n$\\Theta(n\\log n)$, and prove that randomizing the pivot makes the expected\ncost $\\Theta(n\\log n)$ on every input.\n",{"path":24372,"title":24373,"module":24365,"summary":24374},"\u002Falgorithms\u002Fdivide-and-conquer\u002Fselection","Linear-Time Selection","Finding the $k$-th smallest element looks like it should require sorting, but\nit does not. Quickselect adapts quicksort's partition to recurse on just one\nside, achieving expected $O(n)$. The median-of-medians algorithm guarantees a\ngood pivot with the groups-of-five trick, pushing the worst case down to a\nprovable $O(n)$.\n",{"path":24376,"title":24377,"module":24365,"summary":24378},"\u002Falgorithms\u002Fdivide-and-conquer\u002Ffast-multiplication","Fast Multiplication","Grade-school multiplication is $\\Theta(n^2)$, yet divide and conquer beats it.\nKaratsuba multiplies $n$-bit integers with three half-size products instead of\nfour, giving $\\Theta(n^{\\log_2 3})$, and Strassen multiplies matrices with\nseven block products instead of eight, giving $\\Theta(n^{\\log_2 7})$. Both\nspend cheap additions to save an expensive multiplication, and the master\ntheorem quantifies the savings.\n",{"path":24380,"title":24381,"module":24382,"summary":24383},"\u002Falgorithms\u002Fsorting\u002Fheaps-and-heapsort","Heaps and Heapsort","Sorting & Order Statistics","A binary heap is a tree we store flat in an array, with index arithmetic\nstanding in for pointers. We build the max-heap property bottom-up in $O(n)$\ntime, sort in place in $\\Theta(n\\log n)$ by repeatedly extracting the maximum,\nand reuse the same structure to implement a priority queue.\n",{"path":24385,"title":24386,"module":24382,"summary":24387},"\u002Falgorithms\u002Fsorting\u002Fsorting-lower-bounds","Lower Bounds for Comparison Sorting","Every sort we have seen runs in $\\Omega(n\\log n)$, and that is no accident.\nModeling a sort as a decision tree of comparisons, we show any such tree must\nhave $n!$ leaves, forcing height $\\ge \\log_2(n!) = \\Omega(n\\log n)$ — a bound\nno comparison sort beats in the worst case, on average, or with randomness.\n",{"path":24389,"title":24390,"module":24382,"summary":24391},"\u002Falgorithms\u002Fsorting\u002Flinear-time-sorting","Sorting in Linear Time","The $\\Omega(n\\log n)$ barrier only binds algorithms that compare. By instead\nusing keys as array indices we slip past it: counting sort runs in\n$\\Theta(n+k)$ and is stable, radix sort layers it digit by digit, and bucket\nsort averages $\\Theta(n)$ on uniform data. We see exactly when each applies.\n",{"path":24393,"title":24394,"module":24382,"summary":24395},"\u002Falgorithms\u002Fsorting\u002Fexternal-sorting","External Sorting","When the data dwarfs main memory, the cost that matters is no longer\ncomparisons but block transfers to and from disk. External merge sort sorts\nmemory-sized runs, then folds them together with a heap-driven $k$-way merge in\n$\\Theta(\\log_k(N\u002FM))$ passes. Larger fan-out cuts passes; replacement selection\nbuilds longer runs to cut them further.\n",{"path":24397,"title":24398,"module":24399,"summary":24400},"\u002Falgorithms\u002Fdata-structures\u002Felementary-structures","Elementary Data Structures","Data Structures","Every container is built one of two ways: **contiguous** in an array, or\n**linked** through pointers. We trade cache-friendly random access against\n$O(1)$ splicing, derive the **amortized $O(1)$** append of a doubling dynamic\narray, and assemble the two ordered access disciplines — the LIFO **stack** and\nthe FIFO **queue** (with its generalization, the **deque**) — on top of both.\n",{"path":24402,"title":24403,"module":24399,"summary":24404},"\u002Falgorithms\u002Fdata-structures\u002Fhash-tables","Hash Tables","A hash table implements the dictionary — insert, search, delete — in expected\n$O(1)$ time by scattering keys across an array with a hash function. We build\nup from direct addressing, handle collisions by chaining and by open\naddressing, analyze the load factor $\\alpha$, and see how universal hashing\nachieves its expected-time guarantee against every input.\n",{"path":24406,"title":24407,"module":24399,"summary":24408},"\u002Falgorithms\u002Fdata-structures\u002Fbinary-search-trees","Binary Search Trees","A binary search tree keeps keys ordered so that every operation follows a\nsingle root-to-leaf path. We state the BST property, trace search, insert,\nsuccessor, and all three delete cases on concrete trees, prove the inorder\nwalk sorts, and note the drawback — every operation costs $O(h)$, and a\ncarelessly built tree degrades to height $h = \\Theta(n)$, motivating balance.\n",{"path":24410,"title":24411,"module":24399,"summary":24412},"\u002Falgorithms\u002Fdata-structures\u002Favl-trees","AVL Trees","An AVL tree is the first balanced BST: at every node the two subtrees' heights\ndiffer by at most $1$. A Fibonacci-style minimal-node argument forces height\n$h \\le 1.44\\log_2 n = O(\\log n)$, so search, insert, and delete are all\n$O(\\log n)$. Insertion rebalances with at most one of four rotation cases\n(LL, RR, LR, RL); deletion may rotate all the way to the root.\n",{"path":24414,"title":24415,"module":24399,"summary":24416},"\u002Falgorithms\u002Fdata-structures\u002Fbalanced-trees","Balanced Search Trees","An ordinary BST can degrade to height $\\Theta(n)$; balanced search trees\nguarantee $h = O(\\log n)$ by maintaining invariants and repairing them after\nevery update. We meet rotations, the local restructuring primitive, then\nred-black trees, whose color invariants force logarithmic height, and finally\nB-trees, which trade tall-and-thin for short-and-wide to win on disk.\n",{"path":24418,"title":24419,"module":24399,"summary":24420},"\u002Falgorithms\u002Fdata-structures\u002Funion-find","Disjoint Sets (Union-Find)","The disjoint-set data structure tracks a partition of elements into groups,\nanswering \"are these two in the same group?\" and merging groups on demand. A\nforest of parent pointers, sped up by union by rank and path compression,\ndrives every operation to near-constant $O(\\alpha(n))$ amortized time — the\nstructure behind connectivity queries and Kruskal's minimum spanning tree.\n",{"path":24422,"title":24423,"module":24399,"summary":24424},"\u002Falgorithms\u002Fdata-structures\u002Ffenwick-and-segment-trees","Fenwick & Segment Trees","A prefix-sum array answers a range sum in $O(1)$ but pays $O(n)$ per update;\na plain array updates in $O(1)$ but pays $O(n)$ per range sum. Fenwick and\nsegment trees give us _both_ in $O(\\log n)$. The Fenwick (binary indexed) tree\nis a tiny array keyed by the low bit; the segment tree is a general balanced\ntree over canonical ranges that handles any associative aggregate and, with\nlazy propagation, range updates too.\n",{"path":24426,"title":24427,"module":24399,"summary":24428},"\u002Falgorithms\u002Fdata-structures\u002Fspatial-data-structures","Spatial Data Structures","A balanced BST orders keys on a line, but points in the plane have no single\nnatural order. Quadtrees subdivide space recursively into quadrants; k-d trees\nsplit on alternating coordinates at the median. Both make range and\nnearest-neighbour queries fast by carving the plane into boxes a query can\nprune away. Range trees nest a y-tree in an x-tree for fast orthogonal range\nreporting; interval trees index intervals to answer stabbing queries.\n",{"path":24430,"title":24431,"module":24399,"summary":24432},"\u002Falgorithms\u002Fdata-structures\u002Fskip-lists-and-probabilistic-structures","Skip Lists & Probabilistic Structures","Balanced trees achieve $O(\\log n)$ with rotations and invariants; randomization\ngives the same bound far more simply. A skip list is a layered linked list whose\nexpress lanes are chosen by coin flips, giving expected $O(\\log n)$ search and\ninsert with no rebalancing. A Bloom filter trades exactness for space: a bit\narray and a few hashes answer set membership with no false negatives and a\ntunable false-positive rate, but cannot delete.\n",{"path":24434,"title":24435,"module":24399,"summary":24436},"\u002Falgorithms\u002Fdata-structures\u002Fb-trees","B-Trees","When data lives on disk, the cost that dominates is block transfers, not\ncomparisons — and a binary tree of a billion keys is thirty reads deep. A\nB-tree of minimum degree $t$ is short and wide: $t-1$ to $2t-1$ keys per node,\nall leaves at one depth, height $O(\\log_t n)$. Insertion splits a full node on\nthe way down and pushes its median up; deletion borrows or merges to keep nodes\nfull enough. High fan-out is what minimizes disk I\u002FO.\n",{"path":24438,"title":24439,"module":24399,"summary":24440},"\u002Falgorithms\u002Fdata-structures\u002Fdata-stream-algorithms","Data-Stream Algorithms","Most of this course assumes data sits in fast memory, addressable at will.\nExternal sorting relaxed that to a re-readable disk. The streaming model goes\nfurther: items arrive one at a time, are seen once, and must be discarded, with\nonly sublinear, often polylogarithmic, memory. In exchange, the answers are\napproximate and probabilistic. We set up the model, then meet reservoir\nsampling for a uniform sample of an unknown-length stream and Morris counting\nfor an approximate tally in doubly-logarithmic space.\n",{"path":24442,"title":24443,"module":24399,"summary":24444},"\u002Falgorithms\u002Fdata-structures\u002Fstreaming-sketches","Streaming Sketches","Sampling and counting kept a random subset or a single approximate tally.\nSketches go further: fixed, tiny summaries that answer questions about a\nstream's frequencies. We meet the Count–Min sketch for point frequency\nestimation, Misra–Gries for heavy hitters, and HyperLogLog for distinct\ncounts, each trading a controlled error for space that never grows with the\nstream.\n",{"path":24446,"title":24447,"module":24448,"summary":24449},"\u002Falgorithms\u002Fsequences\u002Ftwo-pointers-and-windows","Two Pointers & Sliding Windows","Sequences & Strings","A family of array idioms that collapse an obvious $O(n^2)$ scan into a single\n$O(n)$ pass by maintaining an invariant as indices move. We meet two pointers\n(converging on a sorted array, and a fast\u002Fslow pair for in-place rewriting)\nand the sliding window (fixed and variable size, amortized $O(n)$). The\ncompanion lesson on prefix sums picks up where the window's positivity\nassumption fails.\n",{"path":24451,"title":24452,"module":24448,"summary":24453},"\u002Falgorithms\u002Fsequences\u002Fprefix-sums","Prefix Sums & Difference Arrays","Prefix sums precompute the running total once so that any range-sum query is a\nsingle subtraction, $P[r{+}1]-P[l]$, in $O(1)$. A hash map of prefix\nfrequencies then counts subarrays summing to $k$ in $O(n)$ — even with negative\nentries, where the sliding window fails. The difference-array dual turns $m$\nrange-adds into $O(m+n)$, and the whole idea lifts to 2-D rectangle sums by\ninclusion–exclusion.\n",{"path":24455,"title":24456,"module":24448,"summary":24457},"\u002Falgorithms\u002Fsequences\u002Fmonotonic-stacks","Monotonic Stacks & Queues","A **monotonic stack** keeps its contents sorted by popping every element that\nwould break the order before each push — turning a family of \"previous\u002Fnext\ngreater (or smaller) element\" questions into a single $O(n)$ scan. We trace\nthe next-greater-element routine push by push and prove its amortized bound,\nfuse two such scans to measure the **largest rectangle in a histogram** in\nlinear time, extend the idea to a **monotonic deque** that streams the\n**sliding-window maximum** in $O(n)$, and use asymmetric tie-breaking to\ncount **subarray minimums** without double-counting duplicates.\n",{"path":24459,"title":24460,"module":24448,"summary":24461},"\u002Falgorithms\u002Fsequences\u002Fbinary-search-on-the-answer","Binary Search on the Answer","Binary search locates the boundary of a **monotone predicate** $p(x)$ in\n$O(\\log(\\text{range}))$ probes; sorted arrays are only one instance. We first\nestablish the half-open `while (lo \u003C hi)` template for $\\textsc{lower\\_bound}$\nand $\\textsc{upper\\_bound}$, then generalize to \"binary search on the answer\":\nwhenever feasibility is monotone in a numeric parameter, we binary search the\nparameter itself, calling a feasibility check at each step.\n",{"path":24463,"title":24464,"module":24448,"summary":24465},"\u002Falgorithms\u002Fsequences\u002Fstring-matching","String Matching: Naive & Rabin–Karp","Given a text $T$ of length $n$ and a pattern $P$ of length $m$, find every\noccurrence of $P$ in $T$. The naive scan costs $O(nm)$ and re-reads text it has\nalready seen. Rabin–Karp fixes the first inefficiency with a **rolling hash**:\neach length-$m$ window is summarized by one number, updated in $O(1)$ per slide,\nverified on a hash match to kill collisions, for expected $O(n+m)$. A companion\nlesson removes the re-reading entirely with KMP and the Z-function.\n",{"path":24467,"title":24468,"module":24448,"summary":24469},"\u002Falgorithms\u002Fsequences\u002Fkmp-and-z-function","String Matching: KMP & the Z-Function","Two linear-time matchers that beat Rabin–Karp's expected bound with a\nworst-case guarantee and no randomness. KMP precomputes a **failure function**\n$\\pi$ so a mismatch slides the pattern by $q-\\pi[q-1]$ and the text pointer\nnever backs up, for $O(n+m)$. The **Z-function** computes the longest\nprefix-match at every position via the Z-box, giving the same bound from a\ndifferent angle; the two encodings of a string's self-overlap convert freely.\n",{"path":24471,"title":24472,"module":24448,"summary":24473},"\u002Falgorithms\u002Fsequences\u002Ftries","Tries & Prefix Trees","A **trie** stores a set of strings in a tree keyed by _characters_, so that\ninsert, search, delete, and prefix-test all run in $O(L)$ time — the length\nof the key, _independent of how many keys are stored_. Shared prefixes are\nstored once, which makes tries the natural structure for autocomplete,\nwildcard dictionaries, board word-search, and — over the alphabet $\\{0,1\\}$\n— the maximum-XOR-pair problem. Radix (Patricia) trees compress the chains.\n",{"path":24475,"title":24476,"module":24448,"summary":24477},"\u002Falgorithms\u002Fsequences\u002Fsuffix-arrays-and-aho-corasick","Suffix Arrays, LCP & Aho–Corasick","A **suffix array** sorts all $n$ suffixes of a string, indexing every substring\nat once; built in $O(n\\log n)$, it locates a pattern by binary search in\n$O(m\\log n)$. Its companion **LCP array** (Kasai's $O(n)$ algorithm) counts\ndistinct substrings and finds the longest repeated substring. **Aho–Corasick**\ngeneralises KMP to a whole dictionary: a trie of patterns plus failure links\nscans the text once in $O(\\text{text} + \\text{matches})$ to report every\noccurrence of every pattern. Manacher's algorithm finds all palindromic\nsubstrings in $O(n)$.\n",{"path":24479,"title":24480,"module":24481,"summary":24482},"\u002Falgorithms\u002Fgraphs\u002Frepresentations-and-traversal","Graph Representations and Traversal","Graphs","A graph captures _relationships_ — who connects to whom. We fix the\nvocabulary, weigh the two standard representations (adjacency list versus\nmatrix), then meet the single search skeleton behind everything that follows:\nWhatever-First-Search, and its breadth-first reading, which finds shortest\npaths by number of edges in $O(V + E)$.\n",{"path":24484,"title":24485,"module":24481,"summary":24486},"\u002Falgorithms\u002Fgraphs\u002Fdepth-first-search","Depth-First Search","Swap BFS's queue for a stack and the search plunges instead of fanning out.\nDepth-first search stamps every vertex with discovery and finish times that\nnest like parentheses, classifies each edge as tree, back, forward, or cross,\nand — through the back edge — decides in one pass whether a graph has a cycle.\nThese timestamps underpin topological sort, strong\nconnectivity, and the rest of this module.\n",{"path":24488,"title":24489,"module":24481,"summary":24490},"\u002Falgorithms\u002Fgraphs\u002Ftopological-sort-and-scc","Topological Sort and Strong Connectivity","Directed acyclic graphs model dependencies: tasks that must precede other\ntasks. A _topological order_ lays such a graph out in a line so every edge\npoints forward, and depth-first finish times yield one almost for free.\nWe then ask the harder question for graphs _with_ cycles: which vertices can\nreach each other? The answer is the strongly connected components, found by a\ntwo-pass DFS.\n",{"path":24492,"title":24493,"module":24481,"summary":24494},"\u002Falgorithms\u002Fgraphs\u002Fminimum-spanning-trees","Minimum Spanning Trees","Given a weighted network, how do we connect everything as cheaply as possible?\nThe answer is a minimum spanning tree, and one lemma — the cut property —\njustifies _every_ correct MST algorithm. We prove the cut and cycle\nproperties by exchange arguments, use them to settle uniqueness, and meet the\noldest MST algorithm, Borůvka's, whose parallel component-merging rounds fall\nstraight out of the cut rule.\n",{"path":24496,"title":24497,"module":24481,"summary":24498},"\u002Falgorithms\u002Fgraphs\u002Fkruskal-and-prim","Kruskal and Prim","The two minimum-spanning-tree algorithms you will actually implement.\nKruskal grows a forest edge by edge, cheapest first, using a union-find\nstructure to reject cycle-closing edges; Prim grows one tree outward from a\nroot with a priority queue, exactly Dijkstra rekeyed by attachment cost. Both\ntraced in full on a nine-town graph, with the edge cases, the bottleneck\nproperty, and where each one wins.\n",{"path":24500,"title":24501,"module":24481,"summary":24502},"\u002Falgorithms\u002Fgraphs\u002Fshortest-paths","Shortest Paths","Finding the cheapest route through a weighted network is one of the most-used\nalgorithms in computing, and a single operation — _relaxation_ — underlies\nevery method. We build the primitive, prove the triangle inequality and\noptimal substructure that make it work, then meet Dijkstra's algorithm: the\ngreedy solution for non-negative weights, traced vertex by vertex, with the\ncut argument that proves each extraction is final.\n",{"path":24504,"title":24505,"module":24481,"summary":24506},"\u002Falgorithms\u002Fgraphs\u002Fall-pairs-and-negative-weights","All-Pairs and Negative Weights","Dijkstra's greedy schedule breaks the moment an edge goes negative. We give it\nup for dynamic programming: Bellman-Ford derived as a DP over edge budgets,\nwith its negative-cycle detector, and Floyd-Warshall computing the distance\nbetween _every_ pair of vertices via a DP over which vertices a path may pass\nthrough. We close with Johnson's algorithm and the arbitrage problems that\nnegative cycles encode.\n",{"path":24508,"title":24509,"module":24481,"summary":24510},"\u002Falgorithms\u002Fgraphs\u002Fnetwork-flow","Network Flow","How much can flow through a network from source to sink? We build flow\nnetworks with capacity and conservation constraints, increase a flow by\npushing along augmenting paths in the residual graph, and see how reverse\nedges let the algorithm undo earlier routing. Ford-Fulkerson and its BFS refinement\nEdmonds-Karp find a maximum flow, traced end to end on a worked network.\n",{"path":24512,"title":24513,"module":24481,"summary":24514},"\u002Falgorithms\u002Fgraphs\u002Fmax-flow-min-cut","Max-Flow Min-Cut and Applications","Why is the flow found when no augmenting path remains actually optimal? The\nanswer is a duality theorem: the maximum flow equals the minimum cut. We prove\nit, read the minimum cut off the final residual graph, then derive bipartite\nmatching and a catalog of modeling reductions from the flow\nabstraction — before touching the modern algorithms that supersede\nEdmonds-Karp.\n",{"path":24516,"title":24517,"module":24481,"summary":24518},"\u002Falgorithms\u002Fgraphs\u002Fbridges-and-articulation-points","Bridges & Articulation Points","A **bridge** is an edge whose removal disconnects the graph; an **articulation\npoint** is a vertex whose removal does. Both are single points of failure in a\nnetwork. A single depth-first search computes discovery times and **low-links**,\nand two local criteria — $low[v] > disc[u]$ for bridges, $low[v] \\ge disc[u]$\nfor cut vertices — find them all in $O(V+E)$.\n",{"path":24520,"title":24521,"module":24481,"summary":24522},"\u002Falgorithms\u002Fgraphs\u002Flowest-common-ancestor","Lowest Common Ancestor & Binary Lifting","Given a rooted tree, the lowest common ancestor of $u$ and $v$ is the deepest\nnode that is an ancestor of both. A naive walk answers one query in $O(h)$;\n**binary lifting** precomputes the $2^k$-th ancestor of every node in\n$O(n\\log n)$, then answers $k$-th-ancestor and LCA queries in $O(\\log n)$ each.\nWe derive both jumps, apply them to tree distance, and compare against the\nEuler-tour + RMQ and Tarjan offline alternatives.\n",{"path":24524,"title":24525,"module":24481,"summary":24526},"\u002Falgorithms\u002Fgraphs\u002Ftwo-sat","2-SAT via Implication Graphs","A boolean formula whose every clause has exactly two literals can be solved in\n_linear_ time — even though its three-literal cousin is NP-complete. The idea\nis to read each clause as a pair of implications, build a directed graph on the\n$2n$ literals, and ask a question we already know how to answer: which literals\nshare a strongly connected component? The formula is satisfiable iff no variable\nlands in the same SCC as its own negation, and the SCCs' topological order\nyields a satisfying assignment for free.\n",{"path":24528,"title":24529,"module":24481,"summary":24530},"\u002Falgorithms\u002Fgraphs\u002Feulerian-tours","Eulerian Tours","An **Eulerian tour** uses every _edge_ of a graph exactly once. We give the\nexact parity and balance conditions under which one exists (even degree\nfor undirected graphs, in-degree equal to out-degree for directed) and Hierholzer's\n$O(E)$ algorithm that constructs one by splicing closed sub-tours. We contrast\nthis sharply with the **Hamiltonian** problem (visit every _vertex_ once),\nwhich is NP-complete: visiting edges is easy, visiting vertices is hard.\n",{"path":24532,"title":24533,"module":24481,"summary":24534},"\u002Falgorithms\u002Fgraphs\u002Fbipartite-matching","Bipartite Matching","Pairing applicants to jobs, students to slots, files to disks: all are\n**maximum bipartite matching**. We solve it combinatorially with **augmenting\npaths** (Kuhn's algorithm, $O(VE)$), speed it up to $O(E\\sqrt V)$ with\n**Hopcroft–Karp**, and uncover the structure behind it — **König's theorem**\n(max matching equals min vertex cover) and **Hall's marriage theorem** (a\nperfect matching exists iff every set has enough neighbors).\n",{"path":24536,"title":24537,"module":24538,"summary":24539},"\u002Falgorithms\u002Fgreedy\u002Fthe-greedy-method","The Greedy Method","Greedy Algorithms","A greedy algorithm builds a solution one locally-best choice at a time and\nnever looks back. We isolate the two properties that make this work — the\ngreedy-choice property and optimal substructure — prove the canonical\nactivity-selection algorithm correct with an exchange argument, watch greedy\nfail on the 0\u002F1 knapsack, and glimpse matroids as the theory\nthat says exactly when the greedy method is optimal.\n",{"path":24541,"title":24542,"module":24538,"summary":24543},"\u002Falgorithms\u002Fgreedy\u002Fscheduling-and-intervals","Scheduling & Interval Partitioning","Three classic scheduling problems all yield to greedy algorithms — and all\nthree turn on a single design decision: which key to sort by. Interval\nscheduling sorts by **finish** time to pack the most compatible jobs;\ninterval partitioning sorts by **start** time and proves the rooms needed\nequal the maximum overlap **depth**; minimizing maximum lateness sorts by\n**deadline** and is justified by an adjacent-swap exchange argument.\n",{"path":24545,"title":24546,"module":24538,"summary":24547},"\u002Falgorithms\u002Fgreedy\u002Fhuffman-codes","Huffman Codes","Huffman coding builds a\nprovably optimal prefix-free binary code by repeatedly merging the two least\nfrequent symbols. We develop prefix-free codes as binary trees, give the\nalgorithm with a priority queue, build a Huffman tree from example\nfrequencies, prove optimality with the same greedy-choice-plus-substructure\nargument, and pin the running time at $O(n\\log n)$.\n",{"path":24549,"title":24550,"module":24538,"summary":24551},"\u002Falgorithms\u002Fgreedy\u002Fmatroids","Matroids & Exchange Arguments","The capstone of the greedy module: _why_ and _when_ a greedy algorithm is\nprovably optimal. We recap the two correctness templates — **greedy-stays-ahead**\nand the **exchange argument** — then meet the **matroid** $M=(S,\\mathcal{I})$, an\nabstraction whose **exchange property** is the structure greedy needs.\nThe matroid–greedy theorem says sorting by weight and taking what stays\nindependent yields a maximum-weight basis _if and only if_ the structure is a\nmatroid. Kruskal's MST is the canonical instance; 0\u002F1 knapsack and TSP are the\ncanonical failures.\n",{"path":24553,"title":24554,"module":24538,"summary":24555},"\u002Falgorithms\u002Fgreedy\u002Fstable-matching","Stable Matching (Gale–Shapley)","Two sides each rank the other; we want a matching with no **blocking pair** — no\ntwo participants who both prefer each other to their assigned partners. The\n**Gale–Shapley deferred-acceptance** algorithm has proposers propose in\npreference order while receivers tentatively hold the best offer so far. We prove\nit terminates in $\\O(n^2)$ proposals, returns a **perfect** matching, and that\nthe matching is **stable**. A sharper asymmetry follows: deferred acceptance is\n**proposer-optimal** and **receiver-pessimal**, the structural fact behind the\nresidency match and school-choice systems.\n",{"path":24557,"title":24558,"module":24559,"summary":24560},"\u002Falgorithms\u002Fdynamic-programming\u002Fprinciples","Principles of Dynamic Programming","Dynamic Programming","Dynamic programming is recursion with memory: when a recursive solution\nre-solves the same subproblems again and again, we solve each one once and\nstore the answer. We identify the two structural conditions that make this\nwork — overlapping subproblems and optimal substructure — contrast top-down\nmemoization with bottom-up tabulation, and distil the whole method into a\nfive-step recipe.\n",{"path":24562,"title":24563,"module":24559,"summary":24564},"\u002Falgorithms\u002Fdynamic-programming\u002Fsequence-dp","Sequence Alignment & LCS","Two strings can be compared by how much of one appears inside the\nother. The longest common subsequence (LCS) and edit distance are the two\nclassic measures, and they are the _same_ dynamic program with different\ncosts. We derive the LCS recurrence by examining the last characters, fill a\nworked DP table, reconstruct the subsequence, and then show edit distance as\nthe identical $\\Theta(mn)$ pattern.\n",{"path":24566,"title":24567,"module":24559,"summary":24568},"\u002Falgorithms\u002Fdynamic-programming\u002Flongest-increasing-subsequence","Longest Increasing Subsequence","Given a sequence of numbers, how long is its longest strictly increasing\nsubsequence? A first dynamic program indexes subproblems by the element each\nsubsequence _ends at_, giving an $O(n^2)$ solution with parent-pointer\nreconstruction. A sharper idea, the patience-sorting _tails_ array searched by\nbinary search, drops the time to $O(n\\log n)$. We then fold in the\nvariants: non-decreasing, counting, Russian-doll envelopes, and bitonic.\n",{"path":24570,"title":24571,"module":24559,"summary":24572},"\u002Falgorithms\u002Fdynamic-programming\u002Fknapsack","Knapsack & Subset Problems","We start from $\\textsc{Subset-sum}$ — does some sublist hit a target $t$? — and its\ninclude\u002Fexclude recurrence over a boolean table $A(i, u)$, then bolt on values\nto get 0\u002F1 knapsack as the same machine with $\\lor$ promoted to $\\max$. We fill\nboth tables, recover the chosen items, and confront the surprise that the\n$\\Theta(nt)$ running time is only _pseudo-polynomial_ — exponential in the bit\nlength $b$, and unimprovable unless $\\mathrm{P}=\\mathrm{NP}$ since subset-sum is\n$\\textsc{NP-complete}$. The fractional variant reveals the sharp line between greedy\nand dynamic programming.\n",{"path":24574,"title":24575,"module":24559,"summary":24576},"\u002Falgorithms\u002Fdynamic-programming\u002Fcoin-change-and-unbounded","Coin Change & Unbounded Knapsack","The previous lesson let each item be taken at most once. Drop that cap — items\nmay be reused _any number of times_ — and the 0\u002F1 knapsack collapses from a\ntwo-dimensional table to a one-dimensional one, because there is no longer a\nprefix of \"already-used\" items to track. We meet **unbounded knapsack**, then\nits most famous instance, **coin change**: the minimum-coins recurrence\n$C[a] = 1 + \\min_c C[a-c]$, and the counting variant where the _order of the\nloops_ decides whether you count unordered combinations or ordered sequences —\nthe classic bug. Greed fails in general but works for canonical coin systems.\n",{"path":24578,"title":24579,"module":24559,"summary":24580},"\u002Falgorithms\u002Fdynamic-programming\u002Finterval-dp","Interval DP","Many problems ask for the best way to combine a contiguous range of items, and\nthe answer is a dynamic program over subintervals $[i,j]$ that chooses a split\npoint $k$. We derive the pattern from matrix-chain multiplication —\nparenthesising a product to minimize scalar multiplications in $O(n^3)$ — distil\nit into a reusable template filled by increasing interval length, and then meet\nits sharpest variant: the \"last operation\" trick behind Burst Balloons and\ncutting a stick, where fixing the _last_ move (not the first) makes the two\nsides independent.\n",{"path":24582,"title":24583,"module":24559,"summary":24584},"\u002Falgorithms\u002Fdynamic-programming\u002Ftree-dp","Dynamic Programming on Trees","When the subproblems of a dynamic program are _rooted subtrees_, a single\npost-order DFS solves the whole thing in $O(n)$: each node combines the\nalready-computed answers of its children. We meet the archetype — maximum-weight\nindependent set on a tree — then the \"path through a node\" pattern behind tree\ndiameter and maximum path sum, and finally **rerooting**, which computes a\nper-node answer for _every_ node as root in $O(n)$ with two passes.\n",{"path":24586,"title":24587,"module":24559,"summary":24588},"\u002Falgorithms\u002Fdynamic-programming\u002Fbitmask-dp","Bitmask DP","When a subproblem depends not on an index or a prefix but on _which subset_ of\na small ground set has been used, we can encode that subset as the bits of an\ninteger and index a DP table by it. With $n \\le \\sim 20$ the $2^n$ subsets fit\nin a table, turning $\\Theta(n!)$ brute force into $O(2^n \\cdot \\text{poly}(n))$.\nWe meet the bit tricks, the Held–Karp TSP archetype, assignment by mask,\nsubset-sum partitioning, and submask enumeration with its $3^n$ bound.\n",{"path":24590,"title":24591,"module":24559,"summary":24592},"\u002Falgorithms\u002Fdynamic-programming\u002Fdp-optimizations","DP Optimizations","A correct DP recurrence is only half the battle; its naive evaluation is often\na factor of $n$ slower than necessary. This capstone surveys five techniques,\nmonotonic-queue, the convex hull trick, divide-and-conquer optimization,\nKnuth's optimization, and SOS DP, that each exploit _structure in the\ntransition_ (a sliding window, linear costs, monotone optimal splits, the\nquadrangle inequality, or subset lattices) to shave an $O(n)$, $O(\\log n)$, or\nworse factor off the running time.\n",{"path":24594,"title":24595,"module":24559,"summary":24596},"\u002Falgorithms\u002Fdynamic-programming\u002Fdp-on-graphs","Dynamic Programming on Graphs","Many graph algorithms are dynamic programs: the subproblem is the\n_best value reachable under a restricted resource_ — intermediate vertices\nallowed, edges allowed, or a topological prefix — and edge _relaxation_ is the\nDP transition. We frame Floyd–Warshall as the archetype ($O(V^3)$ all-pairs\nshortest paths), Bellman–Ford as a DP over path length (the at-most-$K$-stops\nvariant), DAG-DP in topological order ($O(V+E)$), and Warshall's transitive\nclosure as the boolean analog.\n",{"path":24598,"title":24599,"module":24559,"summary":24600},"\u002Falgorithms\u002Fdynamic-programming\u002Fdigit-and-probability-dp","Digit & Probability DP","Two DP patterns with unusual state. _Digit DP_ counts the\nintegers in a range $[L, R]$ that satisfy a digit constraint by walking the\ndecimal places of the bound, carrying a _tight_ flag that marks when the prefix\nstill equals the bound's. _Probability\u002FExpectation DP_ replaces \"best value\" with\n\"expected value,\" using linearity of expectation to make each state an\naverage over its weighted transitions — the natural tool for expected step\ncounts and absorbing Markov chains.\n",{"path":24602,"title":24603,"module":24604,"summary":24605},"\u002Falgorithms\u002Fbacktracking\u002Fbacktracking-fundamentals","Backtracking: Subsets, Permutations & Combinations","Backtracking & Search","Backtracking builds a solution one choice at a time and abandons a partial\nsolution the moment it cannot be completed, exploring a state-space tree by\ndepth-first search. We meet the universal choose\u002Fexplore\u002Fun-choose template,\nderive the canonical enumerations — subsets ($2^n$), permutations ($n!$), and\ncombinations ($\\binom{n}{k}$) — handle duplicate elements by skipping equal\nsiblings, and see how pruning turns an exponential search into a tractable one.\n",{"path":24607,"title":24608,"module":24604,"summary":24609},"\u002Falgorithms\u002Fbacktracking\u002Fconstraint-search","Constraint Search: N-Queens & Sudoku","Many hard puzzles are **constraint satisfaction problems**: assign each\nvariable a value from its domain so that every constraint holds. Backtracking\nsolves them by assigning variables one at a time and rejecting a partial\nassignment the instant a constraint breaks. We make the rejection cheap — $O(1)$\nconflict checks for N-Queens via column and diagonal sets — and prune harder\nwith **forward checking**, **MRV** ordering, and **constraint propagation**,\nwhich is what lets an exponential search actually finish.\n",{"path":24611,"title":24612,"module":24604,"summary":24613},"\u002Falgorithms\u002Fbacktracking\u002Fbranch-and-bound","Branch & Bound and Meet in the Middle","Plain backtracking prunes a search tree by _feasibility_; for _optimization_\nproblems we can prune far more aggressively by _value_. **Branch and bound**\nkeeps the best complete solution found so far and discards any partial solution\nwhose optimistic bound cannot beat it. **Meet in the middle** splits the\ninstance in two, enumerates each half, and recombines by binary search — turning\n$2^n$ into $O(2^{n\u002F2}\\,n)$ and pushing exact search out to $n \\approx 40$.\n",{"path":24615,"title":24616,"module":24604,"summary":24617},"\u002Falgorithms\u002Fbacktracking\u002Fgraph-backtracking","Graph Backtracking: m-Coloring & Hamiltonian Paths","Two famous graph problems have no known efficient algorithm, yet yield cleanly\nto backtracking with the right pruning. **Graph $m$-coloring** assigns one of\n$m$ colors to each vertex so no edge is monochromatic; we color vertices in turn\nand reject a color the instant a neighbor already has it. **Hamiltonian\npath\u002Fcycle** asks for a walk visiting every vertex exactly once; we extend a path\ngreedily and backtrack on dead ends. Both are NP-complete, so the worst case is\nexponential — but feasibility pruning and good vertex ordering make real\ninstances tractable, and the contrast with the easy Eulerian condition shows why.\n",{"path":24619,"title":24620,"module":24621,"summary":24622},"\u002Falgorithms\u002Fmathematical-algorithms\u002Fnumber-theory-basics","Number Theory: GCD & Modular Arithmetic","Mathematical Algorithms","This lesson opens the mathematical-algorithms module with the bedrock of\ncomputational number theory. We prove Euclid's recurrence\n$\\gcd(a,b)=\\gcd(b,\\,a\\bmod b)$ and its $O(\\log\\min(a,b))$ running time, extend\nit to recover Bézout coefficients $x,y$ with $ax+by=\\gcd(a,b)$, and build\nmodular arithmetic on residue classes — including when a modular inverse\n$a^{-1}\\bmod m$ exists and how to compute it.\n",{"path":24624,"title":24625,"module":24621,"summary":24626},"\u002Falgorithms\u002Fmathematical-algorithms\u002Fmodular-exponentiation-and-primality","Modular Exponentiation & Primality","Computing $a^n \\bmod m$ naively costs $n$ multiplications; **repeated squaring**\ndoes it in $O(\\log n)$ by reading the bits of the exponent. We use this routine\nto state **Fermat's little theorem** (and the modular inverse it gives), then to\ntest primality — trial division, the probabilistic **Fermat** and **Miller–Rabin**\ntests, and the deterministic witness set that settles primality for every 64-bit\nnumber.\n",{"path":24628,"title":24629,"module":24621,"summary":24630},"\u002Falgorithms\u002Fmathematical-algorithms\u002Fsieve-and-factorization","Sieves & Factorization","The previous lesson tested one number for primality; here we ask for _all_\nprimes up to $n$ at once. The **sieve of Eratosthenes** cross-cuts composites\nin $O(n\\log\\log n)$, and a **linear sieve** does it in $O(n)$ while recording\neach number's **smallest prime factor**, which then factors any $x \\le n$ in\n$O(\\log x)$. From a factorization $x = \\prod p_i^{e_i}$ the multiplicative\nfunctions $\\tau$, $\\sigma$, and Euler's totient $\\varphi$ fall out immediately.\n",{"path":24632,"title":24633,"module":24621,"summary":24634},"\u002Falgorithms\u002Fmathematical-algorithms\u002Fcombinatorics","Combinatorics & Counting","Counting is the arithmetic of finite sets. We build up from permutations\n$n!$ and combinations $\\binom{n}{k}$, prove Pascal's rule by a bijection,\nand count multisets with stars and bars. The practical core is computing\n$\\binom{n}{k}\\bmod p$ in $O(1)$ from precomputed factorials and inverse\nfactorials. We close with inclusion–exclusion and the Chinese Remainder\nTheorem, both of which lean on the modular inverse from the previous lesson.\n",{"path":24636,"title":24637,"module":24621,"summary":24638},"\u002Falgorithms\u002Fmathematical-algorithms\u002Fmatrix-exponentiation","Matrix Exponentiation","A linear recurrence advances by a fixed linear rule, so one step is a\n**matrix–vector** product and $n$ steps are a **matrix power**. Packaging\nFibonacci, and any $k$-term recurrence, into a transition matrix lets us jump\nto the $n$-th term in $O(k^3 \\log n)$ by **exponentiation by squaring** — the\nsame doubling trick from modular exponentiation, now over matrices.\n",{"path":24640,"title":24641,"module":24621,"summary":24642},"\u002Falgorithms\u002Fmathematical-algorithms\u002Ffast-fourier-transform","Fast Fourier Transform","Multiplying two degree-$n$ polynomials by the schoolbook method costs\n$\\Theta(n^2)$. Evaluating them at the **$n$-th roots of unity** turns\nmultiplication into pointwise products, and the **Cooley–Tukey FFT** computes\nall those evaluations in $\\Theta(n\\log n)$ by splitting even and odd\ncoefficients. The inverse FFT interpolates back, giving $\\Theta(n\\log n)$\npolynomial and big-integer multiplication.\n",{"path":24644,"title":24645,"module":24621,"summary":24646},"\u002Falgorithms\u002Fmathematical-algorithms\u002Fgradient-descent","Numerical Optimization and Gradient Descent","Most of this course chases **discrete** optima over finite structures; here the\nsearch space is **continuous** and the objective $f$ is differentiable. The\n**gradient** points uphill, so stepping against it —\n$x_{t+1} = x_t - \\eta\\,\\nabla f(x_t)$ — walks downhill. **Convexity** makes every\nlocal minimum global; for convex $L$-smooth $f$ gradient descent converges at\n$O(1\u002Ft)$, and **geometrically** under strong convexity. **Newton's method** uses\nthe Hessian for local quadratic convergence, and **bisection** is the robust\nbracketing fallback for roots.\n",{"path":24648,"title":24649,"module":24650,"summary":24651},"\u002Falgorithms\u002Fcomputational-geometry\u002Fgeometric-primitives","Geometric Primitives & Orientation","Computational Geometry","Computational geometry is built on a single reliable primitive — the\n**orientation test**, a sign of a cross product that tells whether three points\nturn left, right, or lie collinear. From points-as-vectors and the dot and\ncross products we derive orientation, segment intersection, the shoelace area\nformula, and point-in-polygon tests, keeping all arithmetic **exact and\ninteger** so that no floating-point rounding can corrupt a sign.\n",{"path":24653,"title":24654,"module":24650,"summary":24655},"\u002Falgorithms\u002Fcomputational-geometry\u002Fconvex-hull","Convex Hull","The convex hull is the smallest convex polygon enclosing a point set — the\nrubber band snapped around the nails. We build it with Andrew's monotone chain,\nsorting by $(x,y)$ and sweeping a lower and upper hull while popping any\nnon-left turn via the orientation primitive, in $O(n\\log n)$. A reduction from\nsorting shows that bound is optimal, and the hull yields diameter, smallest\nenclosing rectangle, and more through rotating calipers.\n",{"path":24657,"title":24658,"module":24650,"summary":24659},"\u002Falgorithms\u002Fcomputational-geometry\u002Fsweep-line","Sweep-Line Algorithms","The plane-sweep paradigm turns a static $2$-D geometry problem into a dynamic\n$1$-D ordered-set problem: a vertical line sweeps left to right, stopping at an\n$x$-sorted **event queue** while a balanced-BST **status structure** tracks the\nobjects it currently crosses, ordered by $y$. We derive Bentley–Ottmann segment\nintersection in $O((n+k)\\log n)$, recover closest-pair in $O(n\\log n)$, and\nreduce skyline, rectangle-area, and overlap problems to $\\pm1$ event sweeps.\n",{"path":24661,"title":24662,"module":24650,"summary":24663},"\u002Falgorithms\u002Fcomputational-geometry\u002Fpolygons-and-proximity","Polygons & Proximity","Four classics that live on top of the orientation primitive and the convex\nhull. **Closest pair** falls to divide-and-conquer in $\\Theta(n\\log n)$, where a\npacking argument caps the cross-boundary combine at seven neighbours per point.\n**Point-in-polygon** is the ray-casting parity test or the winding-number count\nthat also handles self-intersecting boundaries, both with their edge caveats. The **shoelace formula**\ngives signed area as a sum of cross products, and **rotating calipers** walk the\nhull to read off diameter and width in $O(n)$.\n",{"path":24665,"title":24666,"module":24667,"summary":24668},"\u002Falgorithms\u002Fintractability\u002Fp-np-reductions","P, NP, and Reductions","Intractability","Most problems we have met so far have fast algorithms. A vast and important\nfamily seemingly does not. This lesson builds the vocabulary for that\ndivide: decision problems, the class $\\mathsf{P}$ of problems we can solve\nquickly, the class $\\mathsf{NP}$ of problems whose solutions we can _check_\nquickly, and polynomial-time reductions, the tool that lets us compare the\ndifficulty of two problems without solving either.\n",{"path":24670,"title":24671,"module":24667,"summary":24672},"\u002Falgorithms\u002Fintractability\u002Fnp-completeness","NP-Completeness","Some problems in $\\mathsf{NP}$ are universally hardest: every other problem\nin $\\mathsf{NP}$ reduces to them. This lesson defines $\\mathsf{NP}$-hard and\n$\\mathsf{NP}$-complete, states the Cook–Levin theorem that anchors the\ntheory on **SAT**, walks the web of reductions that grows from it, and gives\nthe four-step recipe for proving a brand-new problem $\\mathsf{NP}$-complete.\n",{"path":24674,"title":24675,"module":24667,"summary":24676},"\u002Falgorithms\u002Fintractability\u002Fcoping-with-hardness","Coping with NP-Hardness","An $\\mathsf{NP}$-hardness proof rules out an exact polynomial-time algorithm,\nnot the need for answers. This lesson surveys four practical responses to\nhardness: approximation algorithms with a provable ratio (worked through a\n2-approximation for vertex cover), heuristics and local search, exact\nexponential methods like branch and bound, and exploiting special structure\nin the instances you actually face.\n",{"path":24678,"title":24679,"module":24667,"summary":24680},"\u002Falgorithms\u002Fintractability\u002Fapproximation-algorithms","Approximation Algorithms","When a problem is $\\mathsf{NP}$-hard we can still ask for a solution\nprovably close to optimal. This lesson makes the approximation ratio\n$\\rho$ precise, separates absolute from relative guarantees, and proves the\nratios of four classic algorithms: greedy set cover ($H_n \\approx \\ln n$),\nthe MST-doubling $2$-approximation for metric TSP, load balancing, and the\nknapsack FPTAS. It closes with the hierarchy PTAS \u002F FPTAS and the limits of\ninapproximability.\n",{"path":24682,"title":24683,"module":6,"summary":6},"\u002Falgorithms","Algorithms",{"path":24685,"title":24686,"module":24687,"summary":24688},"\u002Fcalculus\u002Flimits-and-continuity\u002Ffunctions-and-models","Functions and Mathematical Models","Limits and Continuity","A function assigns exactly one output to each input and can be presented four ways: verbally, numerically, graphically, or by a formula. The elementary families — linear, polynomial, power, rational, trigonometric, exponential — model most elementary phenomena, and transformation, combination, and composition build every other function from them.\n",{"path":24690,"title":24691,"module":24687,"summary":24692},"\u002Fcalculus\u002Flimits-and-continuity\u002Fthe-limit-of-a-function","The Limit of a Function","The tangent and velocity problems both ask for a value a ratio approaches but never reaches — the limit. Its intuitive two-sided form splits into one-sided limits that must agree; a limit fails to exist when they disagree or when the function grows without bound, the latter producing a vertical asymptote.\n",{"path":24694,"title":24695,"module":24687,"summary":24696},"\u002Fcalculus\u002Flimits-and-continuity\u002Flimit-laws-and-the-precise-definition","Limit Laws and the ε–δ Definition","The Limit Laws reduce a limit to arithmetic on simpler limits, and direct substitution settles polynomials and rational functions outright. The 0\u002F0 forms that resist substitution yield to algebra or the Squeeze Theorem, and the ε–δ definition makes \"arbitrarily close\" precise as a pair of quantified inequalities.\n",{"path":24698,"title":24699,"module":24687,"summary":24700},"\u002Fcalculus\u002Flimits-and-continuity\u002Fcontinuity","Continuity","A function is continuous at a point when its limit there equals its value, so the graph has no break. Continuity fails in three geometric ways; it is closed under arithmetic and composition, so the elementary families and their combinations are continuous; and on a closed interval it forces the Intermediate Value Theorem, which locates roots.\n",{"path":24702,"title":24703,"module":24704,"summary":24705},"\u002Fcalculus\u002Fderivatives\u002Fthe-derivative-and-rates-of-change","The Derivative and Rates of Change","Derivatives","A single limit with three readings: the slope of the tangent line, the instantaneous velocity of a moving object, and the rate of change of one quantity with respect to another. Built from the difference quotient, extended from a value at one point to a function of x, and undefined exactly where a corner, jump, or vertical tangent appears.\n",{"path":24707,"title":24708,"module":24704,"summary":24709},"\u002Fcalculus\u002Fderivatives\u002Fdifferentiation-rules-and-the-chain-rule","Differentiation Rules and the Chain Rule","Computing every derivative from the limit definition is tedious. A short list of rules — power, constant multiple, sum, product, quotient — differentiates any polynomial or rational function by inspection. The trigonometric derivatives follow from one limit, and the chain rule extends everything to composite functions by multiplying rates along the composition.\n",{"path":24711,"title":24712,"module":24704,"summary":24713},"\u002Fcalculus\u002Fderivatives\u002Fimplicit-differentiation-and-related-rates","Implicit Differentiation and Related Rates","Not every curve is the graph of y = f(x). Implicit differentiation finds a slope from an equation in x and y directly, treating y as an unknown function and differentiating both sides. The same chain-rule idea drives related rates, where one measured rate of change forces another through a geometric constraint, and interprets the derivative as a rate across the sciences.\n",{"path":24715,"title":24716,"module":24704,"summary":24717},"\u002Fcalculus\u002Fderivatives\u002Flinear-approximations-and-differentials","Linear Approximations and Differentials","A differentiable curve looks like its tangent line under enough magnification, so the tangent is a usable stand-in for the function near the point of contact. The linear approximation and its linearization, written in the language of differentials dy and dx, estimate both function values and the measurement error propagated into a computed quantity.\n",{"path":24719,"title":24720,"module":24721,"summary":24722},"\u002Fcalculus\u002Fapplications-of-derivatives\u002Fextrema-and-the-mean-value-theorem","Extrema and the Mean Value Theorem","Applications of Derivatives","Absolute and local extrema, the Extreme Value Theorem that guarantees them, and Fermat's Theorem pinning candidates to critical numbers. The Closed Interval Method turns the search for extrema into a finite checklist. Rolle's Theorem and the Mean Value Theorem then connect a function's values to its derivative, giving the tool that most of differential calculus rests on.\n",{"path":24724,"title":24725,"module":24721,"summary":24726},"\u002Fcalculus\u002Fapplications-of-derivatives\u002Fhow-derivatives-shape-a-graph","How Derivatives Shape a Graph","The sign of the first derivative fixes where a function rises and falls, and a sign change identifies each local extremum through the First Derivative Test. The second derivative sets concavity and inflection points and gives a faster Second Derivative Test. Limits at infinity describe end behavior and the horizontal asymptotes a curve settles toward.\n",{"path":24728,"title":24729,"module":24721,"summary":24730},"\u002Fcalculus\u002Fapplications-of-derivatives\u002Fcurve-sketching-and-optimization","Curve Sketching and Optimization","A checklist that synthesizes domain, symmetry, asymptotes, monotonicity, extrema, and concavity into a hand sketch of any function, plus the slant asymptote for rational functions whose degree exceeds the denominator's. The same extremum machinery, applied to a word problem, becomes the optimization template: model one quantity, reduce it to a function of a single variable, and find its absolute extremum.\n",{"path":24732,"title":24733,"module":24721,"summary":24734},"\u002Fcalculus\u002Fapplications-of-derivatives\u002Fnewtons-method-and-antiderivatives","Newton's Method and Antiderivatives","Newton's method solves $f(x) = 0$ by repeatedly replacing the curve with its tangent line and jumping to the tangent's root, converging fast when it works and diverging when the derivative is small. Antiderivatives reverse differentiation: every antiderivative of a function differs from another by a constant, so the general antiderivative is a family of parallel curves, pinned to one by an initial condition.\n",{"path":24736,"title":24737,"module":24738,"summary":24739},"\u002Fcalculus\u002Fintegrals\u002Farea-and-the-definite-integral","Area and the Definite Integral","Integrals","The area under a curve is defined as a limit of sums of rectangle areas. The same limit — a Riemann sum taken as the mesh shrinks to zero — defines the definite integral, a single number measuring signed area, total distance, and every accumulated quantity built the same way. Its properties, comparison bounds, and reading as net area follow directly from the limit.\n",{"path":24741,"title":24742,"module":24738,"summary":24743},"\u002Fcalculus\u002Fintegrals\u002Fthe-fundamental-theorem-of-calculus","The Fundamental Theorem of Calculus","Differentiation and integration are inverse operations. Part 1 says the derivative of an area-accumulation function is the integrand; Part 2 says a definite integral equals the change in any antiderivative across the interval. Together they replace limits of Riemann sums with antiderivative lookups, define the indefinite integral, and give the Net Change Theorem for rates.\n",{"path":24745,"title":24746,"module":24738,"summary":24747},"\u002Fcalculus\u002Fintegrals\u002Fthe-substitution-rule","The Substitution Rule","Substitution runs the Chain Rule backward: spotting an inner function whose derivative also appears in the integrand lets the variable change to $u$ and collapse a composite integral to a simple one. The rule applies to indefinite and definite integrals, with two ways to handle the limits, and it yields the symmetry shortcuts that double even integrands and vanish odd ones.\n",{"path":24749,"title":24750,"module":24751,"summary":24752},"\u002Fcalculus\u002Fapplications-of-integration\u002Fareas-and-volumes","Areas Between Curves and Volumes","Applications of Integration","A definite integral computes any quantity that a limit of Riemann sums approximates. Applied to geometry it gives the area between two curves and the volume of a solid: by cross-sections, by disks and washers when the region is revolved, and by cylindrical shells when inverting the boundary is awkward.\n",{"path":24754,"title":24755,"module":24751,"summary":24756},"\u002Fcalculus\u002Fapplications-of-integration\u002Fwork-average-value-and-arc-length","Work, Average Value, Arc Length, and Surface Area","The work done by a force that varies with position, the average value of a function and the Mean Value Theorem it satisfies, the length of a curve, and the area of a surface swept out by revolving that curve. Each is a limit of Riemann sums, hence a definite integral.\n",{"path":24758,"title":24759,"module":24751,"summary":24760},"\u002Fcalculus\u002Fapplications-of-integration\u002Fphysics-economics-and-probability","Applications to Physics, Economics, and Probability","Definite integrals in physics, economics, and statistics: the force a fluid exerts on a submerged plate, the balance point of a plane region, the money consumers save at a market price, and the probability that a continuous random variable lands in an interval, together with its mean.\n",{"path":24762,"title":24763,"module":24764,"summary":24765},"\u002Fcalculus\u002Fexponential-logarithmic-and-inverse-functions\u002Finverse-functions-logarithms-and-exponentials","Inverse Functions, Logarithms, and Exponentials","Exponential, Logarithmic, and Inverse Functions","A one-to-one function has an inverse that reverses it, with a graph mirrored across y = x and a derivative given by the reciprocal-slope rule. The exponential e^x is its own derivative and the natural logarithm has derivative 1\u002Fx; logarithmic differentiation turns products, quotients, and variable powers into sums.\n",{"path":24767,"title":24768,"module":24764,"summary":24769},"\u002Fcalculus\u002Fexponential-logarithmic-and-inverse-functions\u002Fgrowth-decay-inverse-trig-and-hyperbolic-functions","Growth, Decay, Inverse Trigonometric, and Hyperbolic Functions","Any quantity whose rate of change is proportional to its size grows or decays exponentially, the single equation y' = ky behind populations, radioactive decay, cooling, and continuously compounded interest. The inverse trigonometric functions have algebraic derivatives, and the hyperbolic functions, built from e^x and e^{-x}, describe the hanging cable.\n",{"path":24771,"title":24772,"module":24764,"summary":24773},"\u002Fcalculus\u002Fexponential-logarithmic-and-inverse-functions\u002Flhospitals-rule","Indeterminate Forms and l'Hospital's Rule","When a limit produces 0\u002F0 or infinity over infinity, the value is undetermined by the forms alone. l'Hospital's Rule resolves both by replacing the ratio of functions with the ratio of their derivatives. Products, differences, and powers reduce to a quotient the rule can handle, and repeated use ranks the growth of logarithms, powers, and exponentials.\n",{"path":24775,"title":24776,"module":24777,"summary":24778},"\u002Fcalculus\u002Ftechniques-of-integration\u002Fintegration-by-parts","Integration by Parts","Techniques of Integration","The product rule for derivatives reverses into integration by parts, trading the integral of $u\\,\\d v$ for the integral of $v\\,\\d u$ whenever the second is easier. The LIATE ordering fixes which factor to differentiate. Standard cases: a polynomial against a transcendental factor, repeated parts, cyclic integrals that solve for themselves, and reduction formulas that peel an exponent down by recursion.\n",{"path":24780,"title":24781,"module":24777,"summary":24782},"\u002Fcalculus\u002Ftechniques-of-integration\u002Ftrigonometric-integrals-and-substitution","Trigonometric Integrals and Substitution","Two related techniques. Trigonometric integrals evaluate powers and products of sine, cosine, tangent, and secant by splitting off one factor and converting the rest with a Pythagorean identity, or by dropping even powers with half-angle formulas. Trigonometric substitution runs the idea in reverse: replace x by a sine, tangent, or secant to clear a radical, integrate, then read the answer back off a reference triangle.\n",{"path":24784,"title":24785,"module":24777,"summary":24786},"\u002Fcalculus\u002Ftechniques-of-integration\u002Fpartial-fractions-and-integration-strategy","Partial Fractions and Integration Strategy","Any rational function integrates in closed form: factor the denominator, split the fraction into simple pieces by partial fractions, and integrate each piece as a logarithm or an arctangent. Four denominator cases exhaust the possibilities. A four-step strategy then sorts an arbitrary integrand by its shape to the technique that fits it, and a short catalog records elementary functions whose antiderivatives are not elementary.\n",{"path":24788,"title":24789,"module":24777,"summary":24790},"\u002Fcalculus\u002Ftechniques-of-integration\u002Fapproximate-and-improper-integrals","Approximate and Improper Integrals","Two definite integrals the Fundamental Theorem cannot reach. With no antiderivative available, the Midpoint, Trapezoidal, and Simpson rules approximate the integral from sample values, each carrying a provable error bound. With an infinite interval or an integrand that blows up, the improper integral is defined as a limit that either converges or diverges; the Comparison Test settles which without evaluating it.\n",{"path":24792,"title":24793,"module":24794,"summary":24795},"\u002Fcalculus\u002Fparametric-and-polar\u002Fparametric-curves-and-their-calculus","Parametric Curves and Their Calculus","Parametric Equations and Polar Coordinates","A parametric curve gives x and y separately as functions of a third variable, recording not only a path but the direction and timing with which it is traced. Eliminating the parameter recovers a Cartesian equation; the slope, area, arc-length, and surface-area formulas run directly on the parameter, with the cycloid and astroid as worked examples.\n",{"path":24797,"title":24798,"module":24794,"summary":24799},"\u002Fcalculus\u002Fparametric-and-polar\u002Fpolar-coordinates","Polar Coordinates","Polar coordinates locate a point by a distance from the pole and an angle from the polar axis, giving circles, spirals, and flower-shaped curves short equations. Conversion between the two systems is right-triangle trigonometry, and treating a polar curve as a parametric curve in the angle yields the tangent, area, and arc-length formulas.\n",{"path":24801,"title":24802,"module":24794,"summary":24803},"\u002Fcalculus\u002Fparametric-and-polar\u002Fconic-sections","Conic Sections","Parabolas, ellipses, and hyperbolas are the plane curves cut from a double cone. Each has a focus-based geometric definition and a standard Cartesian equation. A single number, the eccentricity, ties the three together, and placing a focus at the pole gives all of them one polar equation that describes planetary orbits.\n",{"path":24805,"title":24806,"module":24807,"summary":24808},"\u002Fcalculus\u002Fsequences-and-series\u002Fsequences","Sequences","Infinite Sequences and Series","A sequence is a function on the positive integers, and its limit is defined almost exactly like a limit at infinity. The Limit Laws and Squeeze Theorem carry over from functions, monotonic and bounded sequences give a convergence criterion, and the Monotonic Sequence Theorem guarantees a limit exists without naming it.\n",{"path":24810,"title":24811,"module":24807,"summary":24812},"\u002Fcalculus\u002Fsequences-and-series\u002Fseries-and-the-integral-test","Series and the Integral Test","Adding infinitely many terms is made precise as the limit of partial sums. The two series with closed-form partial sums are geometric and telescoping; the harmonic series diverges even as its terms shrink to zero. The Integral Test compares a positive series to an improper integral, settling the p-series and supplying a remainder bound for estimating sums.\n",{"path":24814,"title":24815,"module":24807,"summary":24816},"\u002Fcalculus\u002Fsequences-and-series\u002Fthe-convergence-tests","The Convergence Tests","The comparison, alternating-series, ratio, and root tests decide convergence without a closed-form partial sum. Absolute convergence is stronger than conditional convergence and is preserved under rearrangement; an alternating series errs by less than its first omitted term. A test is chosen from the shape of the general term.\n",{"path":24818,"title":24819,"module":24807,"summary":24820},"\u002Fcalculus\u002Fsequences-and-series\u002Fpower-series","Power Series","A power series is a polynomial of infinite degree whose convergence set is an interval centered at $a$, with a radius the Ratio Test finds and endpoints that must be tested by hand. Inside that interval the series represents a function that can be differentiated and integrated term by term, generating new representations from the geometric series.\n",{"path":24822,"title":24823,"module":24807,"summary":24824},"\u002Fcalculus\u002Fsequences-and-series\u002Ftaylor-and-maclaurin-series","Taylor and Maclaurin Series","If a function equals a power series, its coefficients are forced: the nth is the nth derivative at the center over n factorial. We derive that formula, use Taylor's Inequality to prove the standard series for the exponential, sine, and cosine, record the binomial series and a reference table, and bound the error when a Taylor polynomial replaces a function.\n",{"path":24826,"title":24827,"module":24828,"summary":24829},"\u002Fcalculus\u002Fvectors-and-space-curves\u002Fvectors-and-the-dot-product","Three-Dimensional Coordinates, Vectors, and the Dot Product","Vectors and the Geometry of Space","Space needs three coordinates, so we set up the rectangular system, the distance formula, and the equation of a sphere. Vectors then package magnitude and direction into a single algebraic object with its own arithmetic. The dot product turns two vectors into a number that measures the angle between them, gives a clean test for orthogonality, and produces the projection of one vector onto another.\n",{"path":24831,"title":24832,"module":24828,"summary":24833},"\u002Fcalculus\u002Fvectors-and-space-curves\u002Fthe-cross-product-lines-and-planes","The Cross Product, Lines, and Planes","The cross product multiplies two vectors into a third perpendicular to both, with length equal to the area of the parallelogram they span. That one construction supplies the direction of a line, the normal of a plane, and, through the scalar triple product, the volume of a parallelepiped. Lines carry a point and a direction vector; planes carry a point and a normal, which fixes the angle between planes and the distance from a point to a plane.\n",{"path":24835,"title":24836,"module":24828,"summary":24837},"\u002Fcalculus\u002Fvectors-and-space-curves\u002Fcylinders-and-quadric-surfaces","Cylinders and Quadric Surfaces","A surface whose equation omits one variable is a cylinder: the graph of a plane curve swept along the missing axis. A second-degree equation in three variables is a quadric, and translation and rotation reduce every one to a short standard list. Traces — the curves cut by planes parallel to the coordinate planes — sort the six quadrics into ellipsoid, the two paraboloids, the cone, and the two hyperboloids.\n",{"path":24839,"title":24840,"module":24828,"summary":24841},"\u002Fcalculus\u002Fvectors-and-space-curves\u002Fvector-functions-and-space-curves","Vector Functions and Space Curves","A vector function assigns a vector to each value of a parameter, and as the parameter runs its tip traces a space curve. Taking limits, derivatives, and integrals component by component carries all of single-variable calculus into three dimensions. The derivative of a vector function is the tangent vector to its curve, and normalizing it gives the unit tangent that points the way along the path.\n",{"path":24843,"title":24844,"module":24828,"summary":24845},"\u002Fcalculus\u002Fvectors-and-space-curves\u002Farc-length-curvature-and-motion","Arc Length, Curvature, and Motion in Space","Integrating the speed of a vector function gives the length of its curve and a natural parameter, arc length, that depends only on the curve's shape. Curvature measures how fast the unit tangent turns, and together with the normal and binormal it builds the moving TNB frame. Reading the same vector function as a trajectory, its first two derivatives are velocity and acceleration, and acceleration splits cleanly into tangential and normal parts.\n",{"path":24847,"title":24848,"module":24849,"summary":24850},"\u002Fcalculus\u002Fpartial-derivatives\u002Ffunctions-of-several-variables","Functions of Several Variables, Limits, and Continuity","Partial Derivatives","A function of several variables assigns one number to each point of a region in the plane or in space. Domain, graph, level curve, and level surface describe it; limits and continuity extend to two variables, where a limit must agree along every path of approach, not just from the left and the right.\n",{"path":24852,"title":24849,"module":24849,"summary":24853},"\u002Fcalculus\u002Fpartial-derivatives\u002Fpartial-derivatives","A partial derivative holds every variable but one fixed and differentiates in the ordinary sense. Geometrically it is the slope of a trace curve cut from the surface by a coordinate plane. The freeze-and-differentiate rule computes the two first partials; the four second partials follow, and the two mixed ones agree under Clairaut's Theorem when they are continuous.\n",{"path":24855,"title":24856,"module":24849,"summary":24857},"\u002Fcalculus\u002Fpartial-derivatives\u002Ftangent-planes-and-the-chain-rule","Tangent Planes, Linear Approximation, and the Chain Rule","Near a point, a smooth surface looks like its tangent plane, and the plane's equation is built from the two partial derivatives. That linearization defines the total differential and the meaning of differentiability in two variables. The chain rule then propagates derivatives through composed functions, tracked by a tree diagram, and yields clean formulas for implicit differentiation.\n",{"path":24859,"title":24860,"module":24849,"summary":24861},"\u002Fcalculus\u002Fpartial-derivatives\u002Fdirectional-derivatives-and-the-gradient","Directional Derivatives and the Gradient","The partial derivatives measure slope along the two axes; the directional derivative measures slope along any chosen direction, and equals the gradient dotted with a unit vector. The gradient points in the direction of steepest increase, its length is the greatest rate, and it stands perpendicular to level curves and surfaces, which fixes the tangent plane to a level surface.\n",{"path":24863,"title":24864,"module":24849,"summary":24865},"\u002Fcalculus\u002Fpartial-derivatives\u002Foptimization-and-lagrange-multipliers","Optimization and Lagrange Multipliers","Extrema of a two-variable function sit at critical points where the gradient vanishes; the Second Derivatives Test sorts them into peaks, valleys, and saddles by the sign of a discriminant. Absolute extrema on a closed region also need the boundary. When the domain is itself a constraint curve, Lagrange multipliers set the two gradients parallel and solve the constrained problem.\n",{"path":24867,"title":24868,"module":24869,"summary":24870},"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fdouble-integrals","Double Integrals","Multiple Integrals and Vector Calculus","The double integral extends the definite integral to functions of two variables: a limit of Riemann sums that measures signed volume under a surface. Fubini's Theorem turns it into two ordinary integrations done one after the other, general regions of type I and type II fix the inner limits, polar coordinates absorb circular symmetry through the factor r, and the same machine computes mass, center of mass, and moments of a lamina.\n",{"path":24872,"title":24873,"module":24869,"summary":24874},"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Ftriple-integrals-and-coordinate-systems","Triple Integrals and Coordinate Systems","The triple integral integrates a function of three variables over a solid, as a limit of Riemann sums evaluated by three nested single integrations. Cylindrical coordinates add the factor r to handle axial symmetry, spherical coordinates add rho-squared sine-phi for radial symmetry, and the general change of variables shows both volume elements are Jacobian determinants of the coordinate map. Surface area for a graph completes the measurement toolkit.\n",{"path":24876,"title":24877,"module":24869,"summary":24878},"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fvector-fields-and-line-integrals","Vector Fields and Line Integrals","A vector field assigns a vector to every point of space; the line integral of a field along a curve accumulates its tangential component, measuring work. Conservative fields are gradients of a potential, and for them the Fundamental Theorem for Line Integrals makes the integral depend only on the endpoints. Path independence, closed-loop integrals of zero, and the component test for a potential are three faces of the same property.\n",{"path":24880,"title":24881,"module":24869,"summary":24882},"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fgreens-theorem-curl-and-divergence","Green's Theorem, Curl, and Divergence","Green's Theorem equates the line integral of a field around a positively oriented closed curve with a double integral over the enclosed region, turning a boundary computation into an area computation and vice versa. Curl measures local circulation and divergence measures local outflow; the two vector forms of Green's Theorem express the boundary integral as the integrated curl or divergence, the planar case of Stokes' and the Divergence Theorem.\n",{"path":24884,"title":24885,"module":24869,"summary":24886},"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fsurface-integrals","Parametric Surfaces and Surface Integrals","A parametric surface is the image of a two-variable vector function; its area element is the magnitude of the cross product of the two tangent vectors. The surface integral of a scalar function sums it over that area, and the flux integral of a vector field sums the field's normal component, measuring flow through the surface. Orientation by a choice of unit normal makes flux well-defined, the integral Stokes' and the Divergence Theorem operate on.\n",{"path":24888,"title":24889,"module":24869,"summary":24890},"\u002Fcalculus\u002Fmultiple-integrals-and-vector-calculus\u002Fstokes-and-the-divergence-theorem","Stokes' Theorem and the Divergence Theorem","Stokes' Theorem lifts Green's Theorem into space: the line integral of a field around the boundary of a surface equals the flux of its curl through the surface. The Divergence Theorem relates the outward flux across a closed surface to the triple integral of divergence over the solid it encloses. Together with the Fundamental Theorem of Calculus and its line-integral and Green counterparts, they are one theorem: the integral of a derivative over a region equals the integral of the field over its oriented boundary.\n",{"path":24892,"title":24893,"module":6,"summary":6},"\u002Fcalculus","Calculus",{"path":24895,"title":24896,"module":24340,"summary":24897},"\u002Fmechanics\u002Ffoundations\u002Fmeasurement-and-dimensions","Measurement and Dimensions","Every physical quantity is a number attached to a unit, and that pairing is what lets you check an equation before computing anything, since terms that add together must carry the same dimensions. We build the SI base units and the notion of dimension, then use dimensional analysis to test a proposed relation and form scaling groups — a method that fixes a formula's shape but never its numerical constants. The lesson also sets how precisely a result may be stated, through significant figures, propagated uncertainty, and order-of-magnitude checks that catch errors a raw calculator answer hides.\n",{"path":24899,"title":24900,"module":24340,"summary":24901},"\u002Fmechanics\u002Ffoundations\u002Fvector-algebra","Vector Algebra","Force, velocity, and displacement all carry a direction, so mechanics needs an arithmetic that respects it; adding magnitudes alone gives the wrong answer the moment two arrows point different ways. We set up vectors and their components in a chosen basis, then build the two products that carry most of the physics — the dot product, which extracts the part of one vector along another and yields work and power, and the cross product, which measures oriented area and yields torque and angular momentum. Rotating the axes changes the components while leaving the vector itself untouched, and the same component method resolves a force along whatever directions a constraint picks out.\n",{"path":24903,"title":24904,"module":24905,"summary":24906},"\u002Fmechanics\u002Fkinematics\u002Fone-dimensional-motion","One-Dimensional Motion","Kinematics","Motion along a line already forces the two questions the whole of kinematics repeats: how fast is the object moving now, and where will it be next? Velocity and acceleration answer the first as derivatives of position; integrating them back — the signed area under a graph — answers the second. We derive the constant-acceleration equations, mark exactly where the \"constant\" assumption is load-bearing, and see why sign, not magnitude, is what carries direction.\n",{"path":24908,"title":24909,"module":24905,"summary":24910},"\u002Fmechanics\u002Fkinematics\u002Fmotion-graphs","Motion Graphs","Draw a motion as a graph and its two most useful facts turn geometric: the slope of the position curve is the velocity, and the area under the velocity curve is the displacement. We read motion in both directions — differentiating a graph for the next rate, integrating it back to recover position — and handle the curved, piecewise, and noisy graphs that real measurements produce. Along the way we see why a velocity estimated from two positions belongs to the midpoint of their interval, not its end.\n",{"path":24912,"title":24913,"module":24905,"summary":24914},"\u002Fmechanics\u002Fkinematics\u002Fprojectile-motion","Projectile Motion","Throw an object and it seems to trace one curved path, but the motion is really two independent one-dimensional motions running at once: constant velocity across the ground and free fall in the vertical. Splitting it that way turns every projectile question — how long it stays up, how far it lands, how high it climbs, whether it clears an obstacle — into a pair of equations you already know. We derive the parabolic trajectory, work both the forward and the inverse problems, and show why the familiar $45^\\circ$ range-maximizing angle holds only when launch and landing heights match.\n",{"path":24916,"title":24917,"module":24905,"summary":24918},"\u002Fmechanics\u002Fkinematics\u002Frelative-motion","Relative Motion","A velocity is only ever measured relative to some observer, so a boat's speed through the water, over the ground, and as seen from another boat are three different vectors. Choosing the right frame — and subtracting one motion from another — collapses river crossings, crosswind headings, pursuit, and closest-approach problems into a single vector equation. We build the relative-velocity and relative-position relations for uniformly moving frames, show why acceleration is the one quantity all such observers agree on, and note where rotating frames break the simple subtraction.\n",{"path":24920,"title":24921,"module":24905,"summary":24922},"\u002Fmechanics\u002Fkinematics\u002Fcircular-motion","Circular Motion","An object going around a circle at a steady speed is still accelerating, because its velocity is forever changing direction — the fact that governs everything from a car on a curve to a satellite in orbit. We tie the angular description (angle, angular velocity, angular acceleration) to the linear one through $v=r\\omega$, split the acceleration into an inward part that turns the velocity and a tangential part that changes its speed, and extend the inward $v^2\u002Fr$ result to any curved path through its local radius of curvature. Constant angular acceleration then mirrors straight-line motion equation for equation.\n",{"path":24924,"title":24925,"module":24926,"summary":24927},"\u002Fmechanics\u002Fdynamics\u002Fnewtons-laws","Newton's Laws","Dynamics","What makes a body change its motion, and in which frames does the answer take its simplest form? Newton's three laws settle both: inertial frames are the ones where a force-free body coasts, force is whatever changes momentum, and every interaction pushes back on its source. We write the second law as $\\sum\\vec F=\\d\\vec p\u002F\\d t$, reduce it to $m\\vec a$ at constant mass, and separate what a scale actually reads — the support force — from the weight $m\\vec g$ it is so often mistaken for.\n",{"path":24929,"title":24930,"module":24926,"summary":24931},"\u002Fmechanics\u002Fdynamics\u002Ffree-body-diagrams","Free-Body Diagrams","Once several forces act on a body at once, the reliable way to predict its motion is to isolate that one body and draw every external push and pull on it — nothing more, nothing less. The free-body diagram is that discipline. We fix a system boundary, resolve $\\sum\\vec F=m\\vec a$ into components along axes chosen to fit the geometry, and solve for the unknowns a problem hands us — normal forces, tensions, friction, and the acceleration a constraint permits — seeing why internal forces drop out only when the boundary encloses both bodies that share them.\n",{"path":24933,"title":24934,"module":24926,"summary":24935},"\u002Fmechanics\u002Fdynamics\u002Ffriction-and-curved-motion","Friction and Curved Motion","Real surfaces grip before they slip, fluids push back harder the faster you move through them, and anything rounding a bend must be pulled toward the inside of the curve by something. This lesson supplies the force laws for those three cases. We bound static friction by $|f_s|\\leq\\mu_sN$ and switch to kinetic friction $\\mu_kN$ once sliding starts, model drag as a speed-dependent resistance that levels off at a terminal speed, and show that circular motion demands an inward net force $mv^2\u002Fr$ furnished by real interactions — friction, a banked normal force, tension — never by an invented outward one.\n",{"path":24937,"title":24938,"module":24926,"summary":24939},"\u002Fmechanics\u002Fdynamics\u002Fnumerical-dynamics","Numerical Dynamics","Most force laws — quadratic drag, coupled oscillators, anything nonlinear — admit no closed-form trajectory, so we advance the motion one small time step at a time and let arithmetic do what algebra cannot. This lesson turns $\\d\\vec y\u002F\\d t=f(t,\\vec y)$ into a marching rule. We derive the Euler, Euler--Cromer, midpoint, and Verlet updates, weigh their accuracy and stability, watch a drifting energy expose a bad scheme, and use step-halving and conserved quantities to separate the error of the method from the error of the model.\n",{"path":24941,"title":24942,"module":24926,"summary":24943},"\u002Fmechanics\u002Fdynamics\u002Fcenter-of-mass-systems","Center-of-Mass Systems","A firework bursts into a dozen fragments, yet one point keeps gliding along the original parabola as though nothing had happened. That point is the centre of mass, and following it collapses a many-body tangle into a single equation of motion. We define $\\vec R=\\frac1M\\sum_i m_i\\vec r_i$ and its continuous form, show that internal forces cancel so that only external ones move it, $M\\vec A_{\\rm cm}=\\sum\\vec F_{\\rm ext}$, and put the result to work on recoil, collisions viewed from the centre-of-mass frame, and rocket propulsion, where mass leaving the boundary carries momentum with it.\n",{"path":24945,"title":24946,"module":24947,"summary":24948},"\u002Fmechanics\u002Fenergy\u002Fwork-and-kinetic-energy","Work and Kinetic Energy","Energy","A constant push along a straight path is trivial to score, but real forces vary and bend along curved trajectories, and only the component along the motion transfers any energy. Work captures exactly that transfer as the line integral $W=\\int\\vec F\\cdot\\d\\vec r$, and the work-kinetic-energy theorem turns it into a statement about speed: the net work on a particle equals the change in its $\\tfrac12 mv^2$. We build work up from the dot product to the signed area under a force curve, derive the theorem from Newton's second law, and read power as its instantaneous rate $P=\\vec F\\cdot\\vec v$.\n",{"path":24950,"title":24951,"module":24947,"summary":24952},"\u002Fmechanics\u002Fenergy\u002Fpotential-energy","Potential Energy","When a force does the same work no matter which path a particle takes, that work can be stored as a function of position alone, and solving for the motion becomes bookkeeping instead of integration. We single out the forces that qualify — the conservative ones, for which $\\oint\\vec F\\cdot\\d\\vec r=0$ — define their potential energy through $\\vec F=-\\nabla U$, and use conservation of mechanical energy to read speeds, turning points, and equilibria straight off a potential curve. Friction breaks the shortcut, so we also track where mechanical energy leaks away as heat.\n",{"path":24954,"title":24955,"module":24947,"summary":24956},"\u002Fmechanics\u002Fenergy\u002Fmultiparticle-work","Multiparticle Work","A single particle has one velocity and one kinetic energy; a system of many can spin, deform, explode, and warm up while its centre of mass glides along as if nothing happened. Splitting the motion into a centre-of-mass part and an internal part separates the energy that momentum already fixes from the energy left free for relative motion, $K=\\tfrac12MV_{\\rm cm}^2+K'$. We derive the centre-of-mass work theorem, see why an explosion or a released spring can raise total kinetic energy with no external work at all, and use the reduced-mass and centre-of-mass frames to make collisions and internal transfers clean.\n",{"path":24958,"title":24959,"module":24947,"summary":24960},"\u002Fmechanics\u002Fenergy\u002Fmass-energy-and-binding","Mass-Energy and Binding","Relativity puts rest itself on the energy ledger: a mass $m$ carries energy $mc^2$ even when it sits still, so weighing a system's separated pieces and weighing the assembled whole give different answers, and the gap is binding energy. We convert freely between mass units and MeV, compute the energy that holds a nucleus together, and read the binding-energy-per-nucleon curve that explains why fusing light nuclei and splitting heavy ones both release energy. Reaction $Q$ values, thresholds, and recoil then follow from the same mass-difference accounting, once the frame and mass convention are fixed.\n",{"path":24962,"title":24963,"module":24947,"summary":24964},"\u002Fmechanics\u002Fenergy\u002Fphotons-and-quantization","Photons and Quantization","Light delivers its energy in indivisible lumps: a photon of frequency $f$ carries exactly $hf$, and this one fact explains why a dim blue lamp ejects electrons that an intense red one cannot. We fix a photon's energy and momentum from its wavelength, follow the quanta through emission, absorption, and the photoelectric threshold $K_{\\rm max}=hf-\\phi$, and watch energy and momentum conservation together produce the Compton wavelength shift when a photon scatters from an electron. The recurring discipline is unit and frame care, where a stray factor of $10^9$ or a forgotten rest energy quietly ruins an answer.\n",{"path":24966,"title":24967,"module":24968,"summary":24969},"\u002Fmechanics\u002Fmomentum\u002Fmomentum-and-collisions","Momentum and Collisions","Momentum","When two objects collide, the forces between them are too brief and too tangled to integrate directly, yet the result is fixed by one conserved quantity. Linear momentum $\\vec p=m\\vec v$ turns Newton's second law into the impulse-momentum theorem $\\vec J=\\Delta\\vec p$, and for an isolated system into a conservation law that holds through any internal collision, however dissipative. We use it to separate elastic from inelastic collisions, follow the centre of mass, and read impulse as the signed area under a force-time curve — always tracking which external impulses the chosen system and interval let us drop.\n",{"path":24971,"title":24972,"module":24968,"summary":24973},"\u002Fmechanics\u002Fmomentum\u002Fcenter-of-mass-collisions","Center-of-Mass Collisions","A two-body collision that looks asymmetric in the laboratory becomes almost trivial in the frame that rides along with the centre of mass, where the total momentum is zero and the two momenta stay equal and opposite. We build that frame, reduce the pair to a single relative coordinate carrying the reduced mass $\\mu$, and show that an elastic collision there only rotates one momentum vector while its length holds fixed. Transforming back to the laboratory then handles elastic and inelastic collisions, scattering angles, and reaction thresholds with the same construction — and shows why relative speed, not laboratory kinetic energy, measures what a collision can convert.\n",{"path":24975,"title":24976,"module":24968,"summary":24977},"\u002Fmechanics\u002Fmomentum\u002Frocket-propulsion","Rocket Propulsion","A rocket speeds up by throwing mass backward, so its own mass drops as it flies and $\\vec F=m\\vec a$ no longer applies to a fixed body. Tracking the momentum the exhaust carries across the vehicle boundary gives thrust $T=Ru_e$ and, for a force-free burn, the rocket equation $\\Delta v=u_e\\ln(m_i\u002Fm_f)$ — a logarithm that makes large velocity changes expensive in propellant and forces staging. We then add the forces a real ascent cannot ignore, gravity, drag, and steering, and show how thrust and mass-flow records are cross-checked to infer the exhaust speed.\n",{"path":24979,"title":24980,"module":24981,"summary":24982},"\u002Fmechanics\u002Frotation\u002Frotational-inertia","Rotational Inertia","Rotation","Push a wheel and a merry-go-round with the same force and they speed up at wildly different rates: the same mass resists rotation differently depending on where it sits relative to the axis. That single fact is the moment of inertia, $I=\\int r_\\perp^2\\,\\d m$, and this lesson builds it from the ground up. We tie angular motion to linear through $s=r\\theta$, $v=r\\omega$, and $a_t=r\\alpha$, derive $I$ for rods, disks, and spheres, and use the parallel- and perpendicular-axis theorems to move between axes — always naming the axis, because the same body has as many moments of inertia as it has lines to spin about.\n",{"path":24984,"title":24985,"module":24981,"summary":24986},"\u002Fmechanics\u002Frotation\u002Frotational-dynamics","Rotational Dynamics","A force applied to a wheel does nothing unless it acts off the axis: what turns a rigid body is torque, force times lever arm. This lesson makes that precise and turns it into the rotational Newton's second law, $\\sum\\tau=I\\alpha$ about a fixed axis, the exact analogue of $\\sum F=ma$. From there we get rotational work $W=\\int\\tau\\,\\d\\theta$ and power $P=\\tau\\omega$, size a motor to a load, and solve pulleys and Atwood machines where the pulley's own inertia can no longer be ignored — always insisting that every torque be measured about the same axis.\n",{"path":24988,"title":24989,"module":24981,"summary":24990},"\u002Fmechanics\u002Frotation\u002Frolling-motion","Rolling Motion","A rolling wheel is doing two things at once — translating and spinning — but the no-slip condition $v_{cm}=R\\omega$ locks them together, and that single constraint is what makes rolling tractable. We use it to split the kinetic energy into $\\tfrac12Mv_{cm}^2+\\tfrac12I\\omega^2$, find how fast a cylinder reaches the bottom of an incline, and show why the contact point is instantaneously at rest. The static friction that enforces rolling does no work; we track its direction from the tendency to slip, and mark exactly where the model breaks once the required friction exceeds $\\mu_sN$.\n",{"path":24992,"title":24993,"module":24981,"summary":24994},"\u002Fmechanics\u002Frotation\u002Fangular-momentum","Angular Momentum","A skater pulls in her arms and spins faster, with no torque acting: that is angular momentum conservation, and it lets us answer questions that would be hopeless force by force. We build $\\vec L=\\vec r\\times\\vec p$, show it obeys $\\vec\\tau_{ext}=\\d\\vec L\u002F\\d t$, and use its conservation under zero external torque to link before and after in collisions, reconfigurations, and coupled rotors without ever resolving the internal forces. The catch is bookkeeping: the origin, the system boundary, and the frame must be fixed first, and a change in total $\\vec L$ always points to an external impulse someone forgot.\n",{"path":24996,"title":24997,"module":24981,"summary":24998},"\u002Fmechanics\u002Frotation\u002Frolling-resistance","Rolling Resistance","Ideal rolling should coast forever, yet every real wheel slows down. The reason is that a deformable tire and road do not press through a single point: the contact patch spreads, the normal-force resultant shifts ahead of the axle, and that offset is a resisting moment even with no gross sliding. We package it as an equivalent force $F_{rr}=C_{rr}N$, tie the coefficient to load, surface, speed, and temperature, and use coast-down, towing, and traction tests to separate this contact loss from aerodynamic drag, bearing friction, and the adhesion limit where rolling gives way to skidding.\n",{"path":25000,"title":25001,"module":24981,"summary":25002},"\u002Fmechanics\u002Frotation\u002Fgyroscopic-precession","Gyroscopic Precession","A spinning top leans over but does not fall — it swings its axis in a slow horizontal circle instead. The paradox dissolves once torque is read as the rate of change of a vector: gravity's torque is perpendicular to the spin angular momentum, so it turns $\\vec L$ rather than toppling it. We derive the steady precession rate $\\Omega\\simeq Mgr\u002F(I_s\\omega_s)$ in the fast-top limit, state the assumptions it leans on — dominant spin, slow tilt, negligible bearing torque — and read nutation, support motion, and a decaying spin as the ways real gyroscopes depart from it.\n",{"path":25004,"title":25005,"module":25006,"summary":25007},"\u002Fmechanics\u002Fgravity-and-matter\u002Fkeplerian-orbits","Keplerian Orbits","Gravitation and Matter","Why do the planets trace ellipses rather than any other curve? Newton's inverse-square law collapses the two-body problem onto a single conic section, and the answer falls out of two conserved quantities: a central force can exert no torque, so angular momentum is fixed, and gravity is conservative, so energy is fixed. We read an orbit's size and shape straight off those invariants, recover all three of Kepler's laws, and derive escape speed, the vis-viva relation, and the timing of a pass. We also mark where the ideal ellipse breaks down — drag, oblateness, and a third body slowly move a real orbit.\n",{"path":25009,"title":25010,"module":25006,"summary":25011},"\u002Fmechanics\u002Fgravity-and-matter\u002Fgravitational-fields","Gravitational Fields","Instead of tracking the force between every pair of masses, we attach a field to the source and ask a test mass to read it off locally. That move pays off because gravity is conservative: the field is the gradient of a single scalar potential, and potentials from many sources simply add. We build the field-potential picture, use spherical symmetry and the shell theorem to get the point-mass exterior field and the zero interior field of a shell, and read tides straight out of the field's gradient. Along the way we mark exactly when the constant-$g$ and point-mass shortcuts hold and when a shape correction is needed.\n",{"path":25013,"title":25014,"module":25006,"summary":25015},"\u002Fmechanics\u002Fgravity-and-matter\u002Fstatic-equilibrium","Static Equilibrium","What does it take for a loaded structure to stay put? A body at rest needs its forces to cancel and its turning effects to cancel — $\\sum\\vec F=0$ and $\\sum\\vec\\tau=0$ about any point — and almost all of statics is the craft of turning a physical setup into those equations. We build free-body diagrams, replace supports, cables, friction, couples, and distributed loads with their idealized reactions, and locate the centre of gravity that decides whether a body tips. We also count equations against unknowns to separate a determinate problem from one that needs the material's deformation to resolve, and read every negative or inconsistent reaction as a sign that a contact or a boundary was chosen wrong.\n",{"path":25017,"title":25018,"module":25006,"summary":25019},"\u002Fmechanics\u002Fgravity-and-matter\u002Ffluid-statics","Fluid Statics","A fluid at rest cannot support a shear, so the only stress it carries is a pressure that must grow with depth to hold up the fluid above it. That single balance, $\\d p\u002F\\d z=-\\rho g$, runs the whole subject: it sets manometer readings, the force on a dam, and — integrated over a submerged boundary — Archimedes' buoyant force $F_B=\\rho g V_{\\rm disp}$. We derive these, use them to decide when a body floats and whether it floats upright, and mark where acceleration, rotation, compressibility, or capillarity forces a richer pressure model.\n",{"path":25021,"title":25022,"module":25006,"summary":25023},"\u002Fmechanics\u002Fgravity-and-matter\u002Ffluid-flow","Fluid Flow","Two accounting rules carry most of steady flow: mass cannot pile up, so the same volume crosses every section each second, and mechanical energy is conserved along a streamline when the fluid is ideal. From those we get continuity, Bernoulli's relation between pressure, speed, and height, and the results that follow — Torricelli's efflux speed, the Venturi meter, the Pitot tube. We then let go of the ideal assumptions one at a time: viscosity adds wall shear and head loss, Reynolds number decides laminar versus turbulent, and Mach number marks where a gas stops behaving as incompressible.\n",{"path":25025,"title":25026,"module":25006,"summary":25027},"\u002Fmechanics\u002Fgravity-and-matter\u002Forbital-motion","Orbital Motion","A circular orbit is nothing but free fall with enough sideways speed to keep missing the ground, and setting gravity equal to the centripetal requirement fixes that speed and the period once and for all. From the same energy bookkeeping we read off escape speed, sort orbits into bound, parabolic, and hyperbolic by the sign of their specific energy, and see why a tangential burn is the efficient way to change an orbit. We build the Hohmann transfer and its launch window, work the numbers for a geostationary orbit and an escape burn, and mark where finite thrust, perturbations, and an uncertain initial state pull a real trajectory off the ideal.\n",{"path":25029,"title":25030,"module":25006,"summary":25031},"\u002Fmechanics\u002Fgravity-and-matter\u002Fstress-and-elasticity","Stress and Elasticity","Rigid bodies are a fiction; every real material stretches, shears, or squeezes under load, and the useful question is how much. We define stress as force per area and strain as fractional deformation, then find that for small deformations the two are simply proportional — Hooke's law — with Young's, shear, and bulk moduli as the constants for stretch, twist, and volume change. From these we compute extensions, torsional twist, and stored elastic energy, and read a tensile curve for the yield, ultimate, and fracture points where linear elasticity ends. We also mark the practical limits: stress concentrations, fatigue, and the multiaxial states a single uniaxial modulus cannot capture.\n",{"path":25033,"title":25034,"module":25035,"summary":25036},"\u002Fmechanics\u002Foscillations-waves\u002Fdamped-oscillators","Damped Oscillators","Oscillations and Waves","Every real oscillator eventually stops: friction, drag, and internal loss drain its energy, so free motion is a decay rather than a permanent swing. Adding a velocity-proportional resistance to the spring-mass equation produces one dimensionless number, $b\u002F(2\\sqrt{mk})$, that decides whether the mass rings down through many cycles, returns once without overshoot, or crawls back slowly. We solve the three regimes, tie the observed decay to the power balance $b\\dot x^2$, and turn a measured ring-down into the decay rate and quality factor of the apparatus — reading damping off the data instead of assuming it.\n",{"path":25038,"title":25039,"module":25035,"summary":25040},"\u002Fmechanics\u002Foscillations-waves\u002Ftravelling-waves","Travelling Waves","A wave carries a shape, not the material: each element of a rope or air column oscillates in place while the disturbance travels through it. Writing that shape as $f(x\\mp vt)$ turns \"the pattern moves\" into a statement about the cosine's argument, and a local force balance on one string segment fixes the speed at $v=\\sqrt{T\u002F\\mu}$ — restoring stiffness over inertia, with amplitude nowhere in it. We build the sinusoidal wave and its phase, derive the wave equation from Newton's second law, and follow the energy a travelling wave transports, then check speed and power against those predictions.\n",{"path":25042,"title":25043,"module":25035,"summary":25044},"\u002Fmechanics\u002Foscillations-waves\u002Fwave-superposition","Wave Superposition","When two waves cross the same point, what does a probe read? In a linear medium the answer is arithmetic: the displacements add, $y=y_1+y_2$, and the pulses pass through each other unchanged. That one rule produces interference — reinforcement where the signs agree, cancellation where they oppose — and it guards against a common mistake, since displacement can vanish at an instant while the energy sits in transverse motion instead. We work out the signed sum, the phase bookkeeping for equal-frequency components, and why a null in the record is not a null in the wave.\n",{"path":25046,"title":25047,"module":25035,"summary":25048},"\u002Fmechanics\u002Foscillations-waves\u002Fstanding-waves","Standing Waves","Clamp a string at both ends and only certain frequencies survive: the ends must be nodes, and that single geometric demand quantizes the wave into a discrete set of modes $f_n=nv\u002F(2L)$. The travelling wave becomes a fixed pattern of nodes and antinodes — standing, not moving — because equal waves running in opposite directions superpose. We build the standing wave from its counter-propagating pieces, read the harmonic sequence off the boundary conditions (half-wavelengths for a fixed-fixed string, odd quarter-wavelengths for a closed pipe), and test the ideal model against node scans and resonance peaks.\n",{"path":25050,"title":25051,"module":25035,"summary":25052},"\u002Fmechanics\u002Foscillations-waves\u002Fsound-waves","Sound Waves","Sound is a pressure wave so small that a loud tone displaces air molecules by less than the width of an atom, yet a microphone reads it easily — because pressure, not displacement, is what the ear and the instrument sense. The acoustic impedance $Z=\\rho c$ ties pressure, density, and particle velocity together, fixes the intensity a wave carries, and sets the reference for the decibel, a logarithm that tames a $10^{12}$ range in power. We derive the sound speed from the gas's stiffness, convert between pressure and intensity levels, and treat the measurement itself — calibration, geometry, background, averaging — as part of the physics.\n",{"path":25054,"title":25055,"module":25035,"summary":25056},"\u002Fmechanics\u002Foscillations-waves\u002Fdoppler-effect","Doppler Effect","A passing siren drops in pitch not because the source changes but because motion repacks the wavefronts: an approaching source crowds its crests, a receding one stretches them, and a moving listener samples them at a different rate. For mechanical waves every velocity is measured against the medium, and one signed ratio $f_r=f_s(v-u_r)\u002F(v-u_s)$ captures both effects at once. We separate source motion, which sets crest spacing, from receiver motion, which sets arrival rate, invert the shift to recover radial velocity, and mark where the model breaks — supersonic sources, moving air, and reflected paths that carry two shifts, not one.\n",{"path":25058,"title":25059,"module":25035,"summary":25060},"\u002Fmechanics\u002Foscillations-waves\u002Fwave-packets","Wave Packets","No real signal is a single frequency: a disturbance that starts and stops is built from a band of wave numbers, and the width of that band is what makes it local. We ask how such a packet moves — carrier crests at the phase velocity $v_\\mathrm p=\\omega\u002Fk$, the envelope at the group velocity $v_\\mathrm g=\\d\\omega\u002F\\d k$ — and why the two differ once a medium is dispersive. Curvature $\\d^2\\omega\u002F\\d k^2$ spreads and chirps the packet as it travels, and the Fourier reciprocity that ties bandwidth to duration explains why a finite record, aliasing, or a coarse probe can imitate that spreading unless the sampling limits are respected.\n",{"path":25062,"title":25063,"module":25035,"summary":25064},"\u002Fmechanics\u002Foscillations-waves\u002Fbeats-and-coupling","Beats and Coupling","Add two tones a few hertz apart and the sum swells and fades at their difference frequency — a beat — though neither source is changing. We work out that envelope, then ask the mechanical version of the same question: join two oscillators and a single resonance splits into normal modes, with energy sloshing between the coordinates at their frequency difference. The lesson identifies when a slow amplitude envelope signals genuine coupling rather than two independent sources, drift, or deliberate modulation, reading it from envelope timing, spectral sidebands, and the mode shapes.\n",{"path":25066,"title":25067,"module":25035,"summary":25068},"\u002Fmechanics\u002Foscillations-waves\u002Fsimple-harmonic-motion","Simple Harmonic Motion","Any system pushed back toward equilibrium by a force proportional to its displacement obeys one equation, $\\ddot x+\\omega_0^2x=0$, and so moves sinusoidally at $\\omega_0=\\sqrt{k\u002Fm}$ whatever the amplitude. We derive that motion, follow its energy $E=mv^2\u002F2+kx^2\u002F2$ trading between kinetic and potential form at constant total, and read the elliptical phase-space orbit Hooke's law implies. Period, amplitude, velocity, and acceleration then supply redundant checks: an amplitude-dependent period or a curved force residual is the signature that the linear model has failed, and mass-loading and offset tests separate a calibration error from a real frequency shift.\n",{"path":25070,"title":25071,"module":25035,"summary":25072},"\u002Fmechanics\u002Foscillations-waves\u002Fpendulum-motion","Pendulum Motion","A pendulum keeps time only because, for small swings, gravity supplies a restoring torque proportional to the angle — and $T=2\\pi\\sqrt{L\u002Fg}$ then follows without the mass appearing at all. We derive that result, mark exactly which assumptions carry it (small angle, negligible pivot loss, a rigid support), then relax them: finite amplitude lengthens the period through an elliptic integral, and an extended body replaces $L$ with the ratio of its moment of inertia to its center-of-mass distance. How the period drifts with amplitude or pivot position is what diagnoses the geometric, damping, and distributed-mass corrections.\n",{"path":25074,"title":25075,"module":25035,"summary":25076},"\u002Fmechanics\u002Foscillations-waves\u002Fdriven-oscillators","Driven Oscillators","Drive a damped oscillator at a frequency you control and it eventually forgets its own: $m\\ddot x+b\\dot x+kx=F_0\\cos\\omega t$ settles into a steady response whose amplitude and phase depend sharply on how close the drive sits to resonance. We solve for that response, show how damping alone fixes the resonance width, the peak power, and the settling time, and treat base excitation as the same problem with a different input. The steady-state formulas hold only for constant $m$, $b$, and $k$; level-dependent peaks or hysteresis between up- and down-sweeps are how nonlinearity or an extra mode announces itself.\n",{"path":25078,"title":25079,"module":25035,"summary":25080},"\u002Fmechanics\u002Foscillations-waves\u002Fwave-boundaries","Wave Boundaries","A pulse traveling along a string does something abrupt where the string's properties change: part reflects, part transmits, and which is which is set by the impedance mismatch alone. We impose continuity of displacement and transverse force at the join to get the reflection and transmission coefficients in terms of $Z=\\sqrt{T\\mu}$, fix their signs and the polarity flip, and balance the energy. The clean result assumes linear, nondispersive segments meeting at a localized join; pulse polarity, return timing, and energy ratios are the measurements that expose a real connector's mass, loss, or distributed transition.\n",{"path":25082,"title":25083,"module":25084,"summary":25085},"\u002Fmechanics\u002Fthermodynamics\u002Fkinetic-theory-of-ideal-gases","Kinetic Theory of Ideal Gases","Thermodynamics","A gas has no springs and no gears, yet it pushes on its container with a definite pressure and stores energy in a lawful way. Kinetic theory explains both from the motion of the molecules alone: pressure is the accumulated recoil of countless elastic impacts, and temperature is the average translational kinetic energy each molecule carries. We derive $pV=\\tfrac13Nm\\overline{v^2}$ from momentum transfer, read off $\\overline{K}_{\\rm tr}=\\tfrac32kT$, and use the Maxwell–Boltzmann distribution to separate the most probable, mean, and rms speeds — each the right average for a different question — while marking where the dilute, classical assumptions stop holding.\n",{"path":25087,"title":25088,"module":25084,"summary":25089},"\u002Fmechanics\u002Fthermodynamics\u002Ffirst-law-of-thermodynamics","First Law of Thermodynamics","Heat a gas and it may warm, expand, or both; compress it and the same energy can reappear as a temperature rise. The first law settles the bookkeeping: internal energy is a state property whose change equals the heat added plus the work done on the system, $\\Delta E_{\\rm int}=Q_{\\rm in}+W_{\\rm on}$. We fix a system boundary and one sign convention, compute boundary work as $\\int p\\,\\d V$ along a path, and use calorimetry to measure heat and heat capacities. The recurring point is that heat and work are path-dependent transfers while their sum is not, so an energy ledger closes only once every boundary crossing is named.\n",{"path":25091,"title":25092,"module":25084,"summary":25093},"\u002Fmechanics\u002Fthermodynamics\u002Fentropy-and-the-second-law","Entropy and the Second Law","The first law lets energy flow either way; it never says which way heat actually goes. The second law supplies the missing arrow. Entropy, defined through the reversible transfer $\\d S=\\delta Q_{\\rm rev}\u002FT$, can only increase in an isolated system, and that single inequality fixes the direction of heat flow and caps every engine, refrigerator, and heat pump at its Carnot value. We build entropy ledgers for reservoirs and working substances, separate the entropy carried by heat from the entropy generated by irreversibility, and read the sign of the total as a hard check on any proposed thermal machine.\n",{"path":25095,"title":25096,"module":25084,"summary":25097},"\u002Fmechanics\u002Fthermodynamics\u002Fthermal-processes","Thermal Processes","Heat rarely sits still: it stretches solids, pushes real gases off their ideal isotherms, and leaks across walls by conduction, convection, and radiation. Each behavior becomes a number a designer can use. Thermal expansion sets the gaps in a bridge and the stress in a clamped rod; the van der Waals equation and a phase diagram fix when $pV=nRT$ or a latent-heat term applies; Fourier's law, Newton cooling, and Stefan–Boltzmann radiation give the rate of heat flow. We assemble these into thermal-resistance networks and transient time constants, then mark where contact resistance, phase change, or a hidden thermal bridge breaks the simple model.\n",{"path":25099,"title":25100,"module":25084,"summary":25101},"\u002Fmechanics\u002Fthermodynamics\u002Fphase-changes","Phase Changes","Add heat to ice and its temperature climbs — until it reaches $0\\ ^\\circ\\mathrm C$, where the thermometer stalls while the ice melts. That plateau is the whole subject: at a phase boundary the energy rearranges molecules, $Q=mL$, instead of raising temperature, which resumes only once one phase is gone. We stage a heating path into sensible-heat legs ($Q=mc\\Delta T$) and latent plateaus, use the Clausius–Clapeyron relation to track how a boiling point moves with pressure, and solve calorimetry by testing each coexistence endpoint — so a melt fraction that lands outside $[0,1]$ flags a wrong final-state guess rather than a real state.\n",{"path":25103,"title":25104,"module":25084,"summary":25105},"\u002Fmechanics\u002Fthermodynamics\u002Fthermal-machines","Thermal Machines","An engine, a refrigerator, and a heat pump are one machine read three ways: each shuttles heat between a hot and a cold reservoir while trading work at the boundary, and only the flow you call useful separates them. A heat engine turns part of $Q_h$ into work, $W=Q_h-Q_c$; a refrigerator spends work to pull $Q_c$ from the cold side; a heat pump counts the warm-side delivery instead. We measure each with its own ratio — efficiency or coefficient of performance — bound them all by the Carnot limit that reservoir temperatures alone set, and track how finite temperature differences, throttling, and friction generate entropy and pull real machines below that bound.\n",{"path":25107,"title":25108,"module":6,"summary":6},"\u002Fmechanics","Mechanics & Dynamics",{"path":25110,"title":25111,"module":25112,"summary":25113},"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Fcharge-and-conductors","Charge and Conductors","Electric Fields","Rub two objects together and one pulls electrons from the other; nothing is created, only moved. We define what electric charge is — conserved, additive, and quantized in units of $e$ — and why a conductor's mobile carriers rearrange until its interior field vanishes and its surface sits at one potential. We follow charge through contact, induction, and grounding, treat the field-free cavity that turns a conductor into a shield, and mark where finite conductivity and leakage set the limits of the electrostatic picture.\n",{"path":25115,"title":25116,"module":25112,"summary":25117},"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Fcoulombs-law","Coulomb's Law","Two charges at rest push or pull along the line joining them, and the whole of electrostatics is assembled by adding up such pairs. We measure that force — its inverse-square falloff, its linear dependence on each charge, the sign that says attract or repel — and write it as a vector so direction survives superposition. We work the magnitude and component forms on real numbers, check them against limiting cases and dimensions, and fix the point-charge approximation to source sizes small against every separation.\n",{"path":25119,"title":25120,"module":25112,"summary":25121},"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Felectric-field-and-force","Electric Field and Force","Rather than ask how one charge reaches across empty space to another, we credit the source with a field that fills the space and let a second charge respond to whatever field sits at its own location. Electric field is force per unit positive test charge, $\\vec E=kq\\hat r\u002Fr^2$ for a point source, and source fields add before any receiving charge is placed. We compute those fields and the force $\\vec F=q\\vec E$ they exert, then follow a charge along its parabolic path through a uniform field and into nonuniform fields where the dynamics turn position-dependent.\n",{"path":25123,"title":25124,"module":25112,"summary":25125},"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Felectric-field-maps","Electric Field Maps","A field is a vector at every point of space, and the quickest way to grasp one is to draw it. We build the two standard pictures — continuous field lines tangent to $\\vec E$, and scaled vector arrows — and read direction, magnitude, and the location of nulls straight off them. We fix what a line drawing can and cannot say: density encodes magnitude only under a stated seeding rule, and integral curves never cross at a regular point. From there we work the topology near sources, sinks, and conductor surfaces, and state the step-size and interpolation checks a numerical map must pass.\n",{"path":25127,"title":25128,"module":25112,"summary":25129},"\u002Felectricity-and-magnetism\u002Felectric-fields\u002Felectric-dipoles","Electric Dipoles","Most neutral matter carries no net charge yet still responds to an electric field, because its positive and negative charge sit slightly apart. That separation is a dipole, moment $\\vec p=q\\vec d$ pointing from the negative to the positive charge, and it is the leading term in how any neutral distribution looks from far away. We derive the torque $\\vec p\\times\\vec E$ and energy $-\\vec p\\cdot\\vec E$ a uniform field imposes, the net force a field gradient adds, and the axial and equatorial $1\u002Fr^3$ fields the pair produces, then measure how far out the point-dipole approximation still holds.\n",{"path":25131,"title":25132,"module":25133,"summary":25134},"\u002Felectricity-and-magnetism\u002Fcontinuous-charge-distributions\u002Fcontinuous-charge-fields","Continuous Charge Fields","Continuous Charge Distributions","A charged rod, ring, or disk is not a point, yet its field is still nothing but Coulomb's law added up over the charge it carries. We replace the discrete sum by an integral, with $\\d q=\\lambda\\d\\ell$, $\\sigma\\d A$, or $\\rho\\d V$, so the real work becomes geometry: writing the vector from each source element to the field point, and letting symmetry cancel the components that must cancel before any integral is attempted. We carry the line, ring, and disk fields through in full, then check each result against its near field, its far field, and its dimensions.\n",{"path":25136,"title":25137,"module":25133,"summary":25138},"\u002Felectricity-and-magnetism\u002Fcontinuous-charge-distributions\u002Fgauss-law-and-conductors","Gauss's Law and Conductors","Adding up Coulomb's law over a whole distribution is laborious; Gauss's law trades that sum for a single statement, that the flux of $\\vec E$ out of any closed surface counts the charge inside, $\\oint\\vec E\\cdot\\d\\vec A=Q_{\\rm enc}\u002F\\varepsilon_0$. The law is always true, but it hands over the field only when the source is symmetric enough to pull $E$ outside the integral. We apply it to spheres, lines, and sheets, then turn it on conductors, where the zero interior field drives every excess charge to the surface and fixes the normal-field jump $\\sigma\u002F\\varepsilon_0$, the charge induced on a cavity wall, and electrostatic shielding.\n",{"path":25140,"title":25141,"module":25142,"summary":25143},"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Fpoint-charge-potential","Point-Charge Potential","Electric Potential","The electrostatic force is conservative, so the work it does between two points\ndepends only on the endpoints. That lets us trade the vector field for a single\nscalar attached to each point, the electric potential, the potential energy a unit\ncharge would have there. We build potential from the work integral, fix the usual\nreference at infinity, and add point sources as scalars, $V=k\\sum_i q_i\u002Fr_i$,\navoiding the vector bookkeeping the field demands. Signed charges, the reference\nchoice, equipotential motion, and far-field expansions each give an independent\ncheck on a result.\n",{"path":25145,"title":25146,"module":25142,"summary":25147},"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Fpotential-gradients-and-equipotentials","Potential Gradients and Equipotentials","Given the potential everywhere, how do we recover the field? The field is the\nnegative gradient, $\\vec E=-\\nabla V$: it points down the steepest local drop in\npotential, and its magnitude is set by how fast $V$ changes, not by the shape of a\ncontour. We read off components with directional derivatives, reconstruct fields\nfrom measured potential grids using centered differences, and use closed-loop\nintegrals and grid refinement to test whether a reconstructed field is physically\nconsistent.\n",{"path":25149,"title":25150,"module":25142,"summary":25151},"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Felectrostatic-energy-and-pressure","Electrostatic Energy and Pressure","Assembling a charge configuration takes work, and that work is stored, but where\nis it kept and how much is there? We total it two ways: as a sum over the charges,\n$U=\\tfrac12\\sum_i q_iV_i$, and as an integral over the field itself,\n$u_E=\\tfrac12\\varepsilon_0E^2$, energy the field carries in every region it fills.\nDifferentiating the stored energy at fixed charge or at fixed voltage recovers the\nmechanical force on a conductor, and at a charged surface the same field scale\nappears as an outward electrostatic pressure. We work the parallel-plate case in\nfull and mark where curvature and fringing make the pressure nonuniform.\n",{"path":25153,"title":25154,"module":25142,"summary":25155},"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Flaplace-boundary-problems","Laplace Boundary Problems","Often the charges are not given, only the conductors and the voltages held on\nthem, and the potential in the empty space between has to be found. There $V$ obeys\nLaplace's equation $\\nabla^2V=0$, and the boundary data alone determine a unique solution.\nWe solve it two ways: separation of variables into boundary-matched modes, whose\nhigher spatial frequencies die away with depth into the domain, and finite-difference\nrelaxation for boundaries no analytic mode fits. Residual norms, boundary error, and\nflux balance tell us when the computed potential and its field can be trusted.\n",{"path":25157,"title":25158,"module":25142,"summary":25159},"\u002Felectricity-and-magnetism\u002Felectric-potential\u002Fcontinuous-charge-potentials","Continuous Charge Potentials","When charge is spread over a line, a surface, or a volume, the sum over point\nsources becomes an integral, $V(\\vec r)=k\\int \\d q\u002F|\\vec r-\\vec r'|$. Because\npotential is a scalar, this integral sidesteps the component algebra the field\nwould force, until the field is actually wanted through $\\vec E=-\\nabla V$. We set\nup the right density element for each geometry, choose a workable reference, handle\nthe integrable singularities that arise when the observation point sits on the\ncharge, and check every result against symmetry, dimensions, and the far-field\nmultipole limit.\n",{"path":25161,"title":25162,"module":25163,"summary":25164},"\u002Felectricity-and-magnetism\u002Fcapacitance\u002Fcapacitance-fundamentals","Capacitance Fundamentals","Capacitance","How much charge must you separate onto two conductors to hold a given voltage between\nthem? That ratio, $C=Q\u002F\\Delta V$, is fixed by the conductor geometry and the medium,\nnot by how much charge is presently stored. We compute it from the field for the\nparallel-plate, isolated-sphere, concentric-sphere, and coaxial geometries, trace how\nsurface charge and boundary conditions set each result, and see where fringing,\nguarding, and stray coupling separate the ideal formula from what a bridge measures.\n",{"path":25166,"title":25167,"module":25163,"summary":25168},"\u002Felectricity-and-magnetism\u002Fcapacitance\u002Fcapacitor-networks","Capacitor Networks","Wire several capacitors together and the source sees one equivalent capacitance — but\nwhich? The answer comes not from how the symbols are drawn but from which conductors\nshare a node: parallel branches hold a common voltage and add, $C_{\\rm eq}=\\sum_iC_i$,\nwhile series branches share a common charge and add reciprocally. We derive both rules\nfrom charge conservation on the floating internal node, then extend the node-charge\nmethod to unequal, precharged, and stray-coupled branches and carry a worked reduction\nthrough to the charge and voltage on every element.\n",{"path":25170,"title":25171,"module":25163,"summary":25172},"\u002Felectricity-and-magnetism\u002Fcapacitance\u002Fcapacitor-energy-and-force","Capacitor Energy and Force","Charging a capacitor takes work, because every increment of charge is pushed through\nthe voltage the earlier charge already established. We total that work three\nequivalent ways, $U=Q^2\u002F(2C)=Q\\Delta V\u002F2=C(\\Delta V)^2\u002F2$, locate it in the field as\na density $u=\\tfrac12\\epsilon_0E^2$, then let the plates move. Differentiating the\nstored energy at fixed charge, or the coenergy at fixed voltage, gives the mechanical\nforce; the two boundaries differ only by the work the source supplies. We work the\nparallel-plate attraction and its electrostatic pressure in full, and follow the same\ngradient into pull-in, tilt, comb drives, and traceable force calibration.\n",{"path":25174,"title":25175,"module":25163,"summary":25176},"\u002Felectricity-and-magnetism\u002Fcapacitance\u002Fdielectric-polarization-and-breakdown","Dielectric Polarization and Breakdown","Slide a dielectric between the plates and the capacitance rises — but why, and how\nhard can you drive it before the insulator fails? Bound charge answers the first:\npolarization $\\vec P$ sets up surface and volume charge that partly cancels the\napplied field, so $\\vec D=\\varepsilon_0\\vec E+\\vec P$ separates what the circuit\ncontrols from what the material contributes. We follow the field across layered\ndielectrics and interfaces, tie permittivity and loss to their frequency dependence,\nand treat dielectric strength as a measured, geometry-dependent limit rather than one\nmaterial number.\n",{"path":25178,"title":25179,"module":25180,"summary":25181},"\u002Felectricity-and-magnetism\u002Fdirect-current-circuits\u002Fcurrent-and-resistance","Current and Resistance","Direct-Current Circuits","What does it mean, physically, for charge to flow, and what sets how hard a wire resists that flow? Current counts charge crossing a surface, $I=\\int\\vec J\\cdot\\d\\vec A$, and traces back to a slow drift of many carriers, $\\vec J=nq\\vec v_d$. We establish when the linear law $V=IR$ actually holds, how resistivity and geometry combine into bulk resistance, why real sources sag under load through their internal resistance, and how the three power forms $P=IV=I^2R=V^2\u002FR$ tie electrical work to heating and component ratings.\n",{"path":25183,"title":25184,"module":25180,"summary":25185},"\u002Felectricity-and-magnetism\u002Fdirect-current-circuits\u002Fkirchhoff-network-analysis","Kirchhoff Network Analysis","Once a circuit has more than one loop, no amount of series-parallel folding will reduce it — you need the two conservation laws written as equations. Kirchhoff's junction law is charge conservation at a node; his loop law is energy conservation around a closed path. We turn a labelled network into a linear system in node voltages or mesh currents, fix the sign conventions so a negative answer just means a reversed arrow, and use power balance as an independent check that the algebra describes the circuit that was actually built.\n",{"path":25187,"title":25188,"module":25180,"summary":25189},"\u002Felectricity-and-magnetism\u002Fdirect-current-circuits\u002Frc-transients","RC Transients","How does a circuit get from one steady state to the next when a capacitor refuses to change its voltage all at once? Because a jump would demand infinite current, an RC circuit slides between states exponentially, with a single time constant $\\tau=RC$ that sets the whole schedule: charging fills as $1-e^{-t\u002F\\tau}$, discharge empties as $e^{-t\u002F\\tau}$. We solve the first-order loop equation, read the response off three numbers — the switch-instant voltage, the final dc voltage, and the Thevenin resistance the capacitor sees — and mark where source and probe resistance shift $\\tau$ or where a second storage element hides a mode a one-$\\tau$ fit misses.\n",{"path":25191,"title":25192,"module":25193,"summary":25194},"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fmagnetic-trajectories","Magnetic Trajectories","Magnetic Field","A charged particle in a magnetic field never speeds up or slows down, yet its path curves relentlessly. We work out why: the magnetic force is always perpendicular to velocity, so it does no work and bends the transverse motion into a circle of radius $r=mv_\\perp\u002F(|q|B)$ while leaving the parallel motion untouched, producing a helix. We derive the cyclotron frequency, show why it is independent of speed until relativity intervenes, and turn the geometry around: a measured curvature reads back a particle's momentum, which is how tracking detectors weigh what they cannot see.\n",{"path":25196,"title":25197,"module":25193,"summary":25198},"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fhall-effect","Hall Effect","Current tells you charge is moving, but not whether the movers are positive or negative, nor how many there are. A magnetic field settles both questions. Push current through a strip in a transverse field and the carriers pile up on one edge until a transverse electric field just balances the magnetic deflection; the sign of the resulting Hall voltage names the carrier's charge and its size counts the carriers per volume. We derive the balance $q\\vec E+q\\vec v_d\\times\\vec B=0$, read off $V_H=IB\u002F(nqt)$, and see why field-and-current reversal is what separates the real Hall signal from the offsets that mimic it.\n",{"path":25200,"title":25201,"module":25193,"summary":25202},"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fmagnetic-force-on-conductors","Magnetic Force on Conductors","A magnet pushes on a current-carrying wire even though the wire is electrically neutral. The reason is that each moving carrier feels the Lorentz force, and those microscopic pushes add up to a force the wire's supports must hold. We sum them into $\\d\\vec F=I\\,\\d\\vec\\ell\\times\\vec B$, collapse it to $\\vec F=I\\vec L\\times\\vec B$ for a straight segment in a uniform field, and see exactly when that shortcut fails and the full path integral is needed. The same law runs backward as a measurement: a force-versus-current slope weighs a magnetic field against a known length.\n",{"path":25204,"title":25205,"module":25193,"summary":25206},"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fmagnetic-dipoles","Magnetic Dipoles","A compass needle turns to point north; a current loop in a field does the same thing, and for the same reason. Both are magnetic dipoles, and a uniform field cannot push a dipole anywhere, only twist it. We package a loop's response into one vector, the magnetic moment $\\vec\\mu=IA\\hat n$, from which torque $\\vec\\tau=\\vec\\mu\\times\\vec B$ and orientation energy $U=-\\vec\\mu\\cdot\\vec B$ both follow. Stable alignment sits at the energy minimum, a field gradient is what it takes to produce a net force $\\vec F=\\nabla(\\vec\\mu\\cdot\\vec B)$, and the same moment reappears whenever anything from an electron to a planet acts magnetic.\n",{"path":25208,"title":25209,"module":25193,"summary":25210},"\u002Felectricity-and-magnetism\u002Fmagnetic-field\u002Fmass-spectrometry","Mass Spectrometry","To weigh a single atom you cannot use a scale, so you use a magnetic field instead. A charged ion of unknown mass bends in a field by an amount that depends on its momentum and charge, so if every ion enters with the same velocity, its landing position reads off its mass-to-charge ratio directly. We build the instrument in two stages: crossed electric and magnetic fields that pass only ions with $v=E\u002FB$, and a magnetic sector that bends the survivors along $r=mv\u002F(|q|B)$. Then we ask what blurs a spectral line and how reference ions turn a position into a calibrated mass.\n",{"path":25212,"title":25213,"module":25214,"summary":25215},"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fmoving-charge-fields","Moving-Charge Fields","Magnetic Sources","Every magnetic field comes from charge in motion, and the simplest source is a single point charge drifting past. We work out the field it produces — normal to both the velocity and the line of sight, falling off as the inverse square — and read off why it vanishes straight ahead of the charge and peaks broadside. Summing many such charges is the bridge to steady currents, valid while speeds stay far below $c$ and the motion changes little during the time its field takes to propagate outward.\n",{"path":25217,"title":25218,"module":25214,"summary":25219},"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fbiot-savart-law","Biot–Savart Law","A steady current is a continuous stream of current elements, and the Biot–Savart law hands each one a magnetic contribution — a right-hand cross product that falls off as the inverse square of distance. Summing the contributions along a conductor is a vector line integral, which we carry out for the straight wire to get the endpoint-angle formula. The infinite-wire field $B=\\mu_0 I\u002F2\\pi s$ falls out as the limit where both ends recede, and we mark how fast a finite wire departs from it and when a thin-filament model is safe.\n",{"path":25221,"title":25222,"module":25214,"summary":25223},"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fcircular-current-loops","Circular Current Loops","A ring of current is the simplest source with a well-defined magnetic axis, and it is the building block of every coil and electromagnet. Symmetry kills the transverse Biot–Savart contributions along that axis and leaves a single clean integral; we evaluate it to get $B_z=\\mu_0 I R^2\u002F[2(R^2+z^2)^{3\u002F2}]$, read off the centre field $\\mu_0 I\u002F2R$, and watch it fall into the $1\u002Fz^3$ tail of a magnetic dipole far away. Stacking turns just adds their axial contributions, which is what makes a solenoid out of a pile of loops.\n",{"path":25225,"title":25226,"module":25214,"summary":25227},"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Famperes-law","Ampère’s Law","When a current arrangement is symmetric enough, the Biot–Savart integral is overkill: Ampère's law, $\\oint_C\\vec B\\cdot\\d\\vec\\ell=\\mu_0 I_{\\rm enc}$, gets the field from a single line of reasoning about how much current a loop encloses. We see why the law holds for any steady current, then use cylindrical, planar, and toroidal symmetry to turn the circulation into simple algebra — the field inside and outside a wire, an infinite sheet, a solenoid, and a toroid. We also mark the catch: without symmetry the law still holds but no longer hands you the field pointwise.\n",{"path":25229,"title":25230,"module":25214,"summary":25231},"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fgauss-law-for-magnetism","Gauss’s Law for Magnetism","Electric field lines start and end on charges; magnetic field lines do neither, because no one has ever found an isolated magnetic pole. That single experimental fact is Gauss's law for magnetism: the flux of $\\vec B$ through any closed surface is zero, $\\oint\\vec B\\cdot\\d\\vec A=0$, or in differential form $\\nabla\\cdot\\vec B=0$. We work through what it says — every field line that enters a closed surface must leave it, so field lines close on themselves — and, just as important, what it does not say, since flux through an open surface is generally nonzero.\n",{"path":25233,"title":25234,"module":25214,"summary":25235},"\u002Felectricity-and-magnetism\u002Fmagnetic-sources\u002Fmagnetic-materials","Magnetic Materials","Put matter in a magnetic field and its atoms respond, each acting as a tiny current loop; the aligned moments per unit volume are the magnetization $\\vec M$, whose bound currents add to the field. Separating what we control (the free current) from what the material supplies leads to $\\vec H$ and the relation $\\vec B=\\mu_0(\\vec H+\\vec M)$. We sort materials into diamagnets, paramagnets, and ferromagnets by how $\\vec M$ answers, follow a ferromagnet around its hysteresis loop, and see why the loop's area is the energy dissipated per cycle and why a sample's shape changes the field it actually feels.\n",{"path":25237,"title":25238,"module":25239,"summary":25240},"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Fmagnetic-flux","Magnetic Flux","Electromagnetic Induction","A magnetic field threading a loop collapses to one signed number, the flux, and every induced voltage in this module turns out to be a rate of change of that number — so defining the flux and its sign comes first. We define it as the surface integral of $\\vec B$ over an oriented surface, reduce it to $BA\\cos\\theta$ for a uniform field on a flat loop, and carry the flux linkage $N\\Phi_B$ of a coil. The chosen normal fixes the sign; reversing it flips the sign without touching the field. Nonuniform fields and curved surfaces force the integral, so we also build the numerical estimate and the checks that separate a reliable value from a nominal field-times-area product.\n",{"path":25242,"title":25243,"module":25239,"summary":25244},"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Ffaradays-law","Faraday's Law","Move a magnet toward a coil, or ramp the current in a nearby circuit, and a voltage appears with no battery in sight. Faraday's law names the cause: the emf around a loop equals minus the rate of change of the magnetic flux through it, so any change of field, area, orientation, or position that alters the flux drives an emf. We separate the emf, which lives around the boundary whether or not current can flow, from the current that follows only when the path is closed; fix the single sign convention that ties flux to loop orientation; and read the emf off rotating coils and off flux sampled at discrete times.\n",{"path":25246,"title":25247,"module":25239,"summary":25248},"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Flenzs-law","Lenz's Law","The minus sign in Faraday's law is not decoration: it decides which way the induced current flows, and it always chooses the direction that fights the change that produced it. Lenz's law reads that sign off energy conservation — a current that aided the change would be free energy — and turns it into a repeatable procedure. We fix a surface normal and a positive loop direction so the sign is calculable, then work through approaching magnets, expanding loops, coupled coils, and rotating generators, using mechanical work and Joule heating as an independent check on every direction we draw.\n",{"path":25250,"title":25251,"module":25239,"summary":25252},"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Fmotional-emf","Motional EMF","Push a wire through a magnetic field and its free charges feel a sideways magnetic force that piles them up at the ends — a battery made of motion. Motional emf is that effect: the work per unit charge a moving conductor supplies is the line integral of $\\vec v\\times\\vec B$ along it, which for a rod moving perpendicular to both its length and the field collapses to $B\\ell v$. We chase where the energy comes from — the hand or motor fighting the magnetic drag, never the magnetic force itself — solve the sliding-rail circuit from both flux and carrier forces, and carry the idea into rotating rods, homopolar disks, generators, and the back emf of a motor.\n",{"path":25254,"title":25255,"module":25239,"summary":25256},"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Feddy-currents","Eddy Currents","A wire carries current along one path; a solid block of metal offers a continuum of them, and any changing flux threading that block sets charge circulating in closed loops it chooses for itself. We ask what those eddy currents do — where they heat, where they drag, and how Lenz's law fixes their direction — and why the same circulation is a feature in an induction furnace and a loss to be suppressed in a transformer core. From a representative-loop estimate we get the scaling (heating grows with the square of frequency and flux rate) and the two design levers, lamination and resistivity, that break the paths a solid conductor would otherwise hand the current.\n",{"path":25258,"title":25259,"module":25239,"summary":25260},"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Fself-inductance","Self-Inductance","A coil resists changes to its own current. Drive current through it and the flux it produces threads its own turns; change that current and Faraday's law turns the coil against the source with a back emf $\\mathcal E_L=-L\\,\\d I\u002F\\d t$. We define self-inductance as the flux linkage per ampere fixed by winding and core geometry, derive the long-solenoid value $L=\\mu_0 N^2A\u002F\\ell$, and follow the consequence that dominates circuits: because a finite voltage can only sustain a finite $\\d I\u002F\\d t$, an inductor's current cannot jump — which is why opening a switch on a live coil throws a spark.\n",{"path":25262,"title":25263,"module":25239,"summary":25264},"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Fmagnetic-energy","Magnetic Energy","Building current in a coil means working against its back emf, and that work does not vanish — it sits in the magnetic field as recoverable energy $U_B=\\tfrac12LI^2$, spread through space at density $u_B=B^2\u002F(2\\mu_0)$. We derive both forms, show they agree for a solenoid, and read a force out of the same energy: an armature is pulled toward higher inductance, and $B^2\u002F(2\\mu_0)$ doubles as a magnetic pressure. The lesson closes on the accounting a real switching event demands, where recoverable energy, copper heating, core loss, and clamp dissipation must balance a single ledger.\n",{"path":25266,"title":25267,"module":25239,"summary":25268},"\u002Felectricity-and-magnetism\u002Felectromagnetic-induction\u002Frl-circuits","RL Circuits","Put a resistor and an inductor in series and the current cannot switch on or off at will: it climbs to $V_0\u002FR$ and falls away exponentially on a single time scale $\\tau=L\u002FR$ set by how much flux the coil hoards against how fast the resistor bleeds it. We solve the turn-on and turn-off, then confront the practical sting — because the coil's current refuses to stop instantly, breaking its path throws up a large voltage, which is why real inductive circuits carry freewheel diodes and clamps that trade voltage stress against how quickly the current dies.\n",{"path":25270,"title":25271,"module":25272,"summary":25273},"\u002Felectricity-and-magnetism\u002Falternating-current\u002Fac-fundamentals","AC Fundamentals","Alternating Current","A wall socket delivers a voltage that averages to zero over each cycle, yet it still heats a filament and runs a motor. The resolution is that dissipation follows the mean of the square, not the mean, so we define the root-mean-square value that makes an alternating source the equal of a DC one for resistive heating. We show a sinusoid's RMS is its peak divided by $\\sqrt2$, work out the average power an ideal resistor draws when its current stays in phase with the applied voltage, and separate the peak, average, and RMS descriptions that a single number cannot combine.\n",{"path":25275,"title":25276,"module":25272,"summary":25277},"\u002Felectricity-and-magnetism\u002Falternating-current\u002Freactance","Reactance","A resistor obeys Ohm's law instant by instant, but a capacitor responds to how fast its voltage changes and an inductor to how fast its current changes. Under a steady sinusoid that rate-dependence collapses to a fixed quarter-cycle phase shift and a frequency-dependent amplitude ratio, the reactance. We derive $X_C=1\u002F(\\omega C)$ and $X_L=\\omega L$, adopt phasors to turn the defining derivatives into multiplication by $j\\omega$ so a single complex impedance carries amplitude and phase together, and track the energy an ideal reactance stores and returns without dissipating it. Real windings and dielectrics add loss, leakage, and self-resonance that bound where the ideal formulas hold.\n",{"path":25279,"title":25280,"module":25272,"summary":25281},"\u002Felectricity-and-magnetism\u002Falternating-current\u002Frlc-resonance","RLC Resonance","Put a resistor, inductor, and capacitor in one loop and their reactances work against each other: inductive reactance grows with frequency while capacitive reactance shrinks, and at one frequency they cancel exactly. There the branch looks purely resistive, the current peaks, and the inductor and capacitor voltages can swing far above the source. We locate that resonance at $\\omega_0=1\u002F\\sqrt{LC}$, measure how sharp the peak is with the quality factor $Q=\\omega_0L\u002FR$, tie its half-power bandwidth $R\u002FL$ to the ringdown of the unforced circuit, and read the same poles off as bandpass and peaked filters at the R, L, or C terminals.\n",{"path":25283,"title":25284,"module":25272,"summary":25285},"\u002Felectricity-and-magnetism\u002Falternating-current\u002Fac-power","AC Power","Multiply an AC load's RMS voltage by its RMS current and you get an answer in volt-amperes that the wiring must carry, but not in general the watts the load consumes. The phase between voltage and current splits that product into a part that does net work and a part that merely sloshes energy back and forth. We derive the average power $P=V_{\\rm rms}I_{\\rm rms}\\cos\\phi$, package amplitude and phase into complex power $S=P+jQ$ so that real, reactive, and apparent power form one right triangle, and see why a harmonic-rich current forces the time-domain definition $P=\\langle vi\\rangle$ in place of a single phase angle.\n",{"path":25287,"title":25288,"module":25272,"summary":25289},"\u002Felectricity-and-magnetism\u002Falternating-current\u002Ftransformers","Transformers","Two coils sharing an iron core exchange no charge, yet a changing current in one drives a voltage in the other, and the ratio of their turns sets how voltage and current trade off between the windings. That lets a transformer step a voltage up or down, isolate two circuits, and make a load look larger or smaller to the source by the square of the turns ratio. We build the ideal ratio element from Faraday's law and the dot convention, derive the reflected-impedance rule, then add the winding resistance, leakage, magnetizing current, and core loss that turn the ideal ratios into real regulation, efficiency, and a bounded voltage-frequency range.\n",{"path":25291,"title":25292,"module":25293,"summary":25294},"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Fdisplacement-current","Displacement Current","Maxwell’s Equations and Electromagnetic Waves","Ampère's law asks for the current through a surface bounded by a loop, but a charging capacitor breaks it: slide the surface off the wire and into the gap and the enclosed conduction current drops to zero, while the magnetic field around the loop plainly does not. Maxwell's repair is to count a changing electric flux as itself a source of magnetic circulation. We derive the displacement-current term $\\varepsilon_0\\,\\d\\Phi_E\u002F\\d t$, show that charge continuity demands it, compute the magnetic field it produces inside a charging capacitor, and see how it closes the Ampère–Maxwell law so that electric and magnetic fields can sustain one another as a wave.\n",{"path":25296,"title":25297,"module":25293,"summary":25298},"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Felectromagnetic-waves","Electromagnetic Waves","Once a changing electric flux can drive a magnetic field, the two curl laws feed each other: a disturbance in one regenerates the other, and the pair walks off through empty space with no medium holding it up. We take the curl of Faraday's law, land on a wave equation whose speed is fixed entirely by $\\mu_0$ and $\\varepsilon_0$, and find that $c=1\u002F\\sqrt{\\mu_0\\varepsilon_0}$ falls out of purely electric and magnetic constants. The plane-wave solution then fixes the geometry — $\\vec E$, $\\vec B$, and the propagation direction mutually perpendicular, oscillating in phase, with amplitudes locked at $E=cB$ — a set of independent predictions any real measurement must meet at once.\n",{"path":25300,"title":25301,"module":25293,"summary":25302},"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Felectromagnetic-momentum","Electromagnetic Momentum","A light beam carries no mass, yet it pushes: shine it on a surface and the surface feels a force. We trace that force back to the fields, which store energy with density $\\varepsilon_0E^2$ and carry it along the Poynting vector $\\vec S=\\vec E\\times\\vec B\u002F\\mu_0$. Because that energy also carries momentum $U\u002Fc$, an absorbed beam presses with $I\u002Fc$ and a mirror with $2I\u002Fc$. We derive the Poynting theorem as local energy conservation, tie intensity to field amplitude, and work the momentum balance carefully enough that oblique incidence, partial reflection, and finite beams all drop out of one accounting.\n",{"path":25304,"title":25305,"module":25293,"summary":25306},"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Fdipole-radiation","Dipole Radiation","Only accelerating charge radiates, and the simplest accelerator is a charge sloshing back and forth: an oscillating electric dipole. We work out the field it throws off, keeping the part that survives to large distance — the $1\u002Fr$ radiation field whose intensity goes as $\\sin^2\\theta\u002Fr^2$, zero along the dipole axis and strongest broadside. From it follow the $\\omega^4$ scaling of total radiated power, radiation resistance as the feed's view of that escaping power, and, through reciprocity, the fact that a good transmitter receives well in the same directions. The near-zone terms that fall off faster carry no net power, and we mark carefully where each description is allowed to be used.\n",{"path":25308,"title":25309,"module":25293,"summary":25310},"\u002Felectricity-and-magnetism\u002Fmaxwell-electromagnetic-waves\u002Fpolarization","Polarization","A plane wave still leaves one thing free: which way its electric field points as it oscillates. That freedom is polarization, set entirely by the relative amplitude and phase of the two transverse field components — in phase gives a line, equal amplitudes a quarter cycle apart give a circle, everything else an ellipse. We work out how a linear analyzer reads a state through Malus's law $I=I_0\\cos^2\\theta$, why that scan alone cannot tell circular light from unpolarized, and how a quarter-wave plate plus a few analyzer settings recover the full Stokes vector and the degree of polarization.\n",{"path":25312,"title":25313,"module":25314,"summary":25315},"\u002Felectricity-and-magnetism\u002Foptics\u002Freflection-and-refraction","Reflection and Refraction","Geometrical Optics","Light meeting a boundary between two transparent media splits into a reflected ray and a bent transmitted one, and predicting where those rays go is the whole starting point of geometrical optics. Fixing one convention — every angle measured from the surface normal — we get reflection's equal angles and derive Snell's law $n_1\\sin\\theta_1=n_2\\sin\\theta_2$ from wavefront timing. That single relation, applied once or twice, yields the critical angle and total internal reflection, prism deviation, the lateral shift through a window, apparent depth, and a fiber's acceptance cone; a wavelength-dependent index then adds dispersion. We mark throughout where the ray picture is trustworthy: feature sizes large against the wavelength and clean interface geometry.\n",{"path":25317,"title":25318,"module":25314,"summary":25319},"\u002Felectricity-and-magnetism\u002Foptics\u002Fthin-lenses","Thin Lenses","A lens gathers the light spreading from one point back onto another, and a single paraxial relation $1\u002Fs+1\u002Fs'=1\u002Ff$ predicts where that image lands and how large it is. We collapse two refractions into one bending plane, read image position and orientation off the three principal rays, and trace focal length back to glass and curvature through the lensmaker equation. Sign conventions carry the physics here — they separate real from virtual images and upright from inverted — so we drill them before chaining lenses in sequence and in contact. The lesson ends on how focal length is actually measured on a bench, and where finite thickness, aperture, and dispersion break the thin-lens picture.\n",{"path":25321,"title":25322,"module":25314,"summary":25323},"\u002Felectricity-and-magnetism\u002Foptics\u002Fspherical-mirrors","Spherical Mirrors","Curve a mirror and it stops merely reflecting an image and starts forming one: the same $1\u002Fs+1\u002Fs'=1\u002Ff$ that governs lenses reappears, now with $f=R\u002F2$ and reflected rays and object sharing one side of the glass. We derive the mirror equation from the reflection geometry of a single paraxial ray, then let signed distances do the sorting — real inverted images on the near branch, virtual upright ones behind the surface — and check the concave, convex, and plane-mirror limits against each other. The second half turns to how focal length is actually measured on a bench, by finite conjugates, distant targets, return imaging, and sagitta, and to the aperture and off-axis aberrations the single paraxial focus cannot capture.\n",{"path":25325,"title":25326,"module":6,"summary":6},"\u002Felectricity-and-magnetism","Electricity & Magnetism",{"path":25328,"title":25329,"module":25330,"summary":25331},"\u002Flinear-algebra\u002Flinear-systems\u002Fsystems-and-echelon-forms","Systems of Linear Equations and Row Reduction","Linear Equations in Linear Algebra","A linear system is a finite set of linear equations in shared variables. Elementary row operations rewrite it without changing its solution set, and reducing the augmented matrix to echelon form decides both existence and uniqueness. Pivot positions say whether the solution set is empty, a single point, or infinite.\n",{"path":25333,"title":25334,"module":25330,"summary":25335},"\u002Flinear-algebra\u002Flinear-systems\u002Fvector-and-matrix-equations","Vector Equations and the Matrix Equation Ax = b","The same linear system reads three equivalent ways: a system of equations, a vector equation asking whether b is a linear combination of fixed vectors, and a matrix equation Ax = b. Ax is the linear combination of A's columns weighted by x, so consistency for a given b means b lies in the span of the columns, and consistency for every b means the columns span all of R^m.\n",{"path":25337,"title":25338,"module":25330,"summary":25339},"\u002Flinear-algebra\u002Flinear-systems\u002Fsolution-sets-and-applications","Solution Sets and Applied Linear Systems","A homogeneous system Ax = 0 has a solution set that is a span through the origin; a consistent Ax = b has that same span translated by any one particular solution. Parametric vector form writes both explicitly. The structure shows up in applied systems with many solutions: equilibrium prices, balanced chemical reactions, network flows, weight-loss diets, and migration models.\n",{"path":25341,"title":25342,"module":25330,"summary":25343},"\u002Flinear-algebra\u002Flinear-systems\u002Flinear-independence","Linear Independence","A set of vectors is linearly independent when the only linear combination equal to zero is the trivial one; otherwise a dependence relation writes one vector in terms of the others. For the columns of A the question becomes whether Ax = 0 has only the trivial solution — a pivot in every column. Counting pivots settles independence, and any set with more vectors than entries is automatically dependent.\n",{"path":25345,"title":25346,"module":25330,"summary":25347},"\u002Flinear-algebra\u002Flinear-systems\u002Flinear-transformations","Linear Transformations and Their Matrices","Reading A as an action rather than an array, x maps to Ax is a transformation from R^n to R^m. The ones that preserve addition and scalar multiplication are the linear transformations, and every one is x maps to Ax for a unique standard matrix whose columns are the images of the standard basis vectors. Onto and one-to-one translate into the span and independence of those columns.\n",{"path":25349,"title":25350,"module":25351,"summary":25352},"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fmatrix-operations","Matrix Operations","Matrix Algebra","Matrices add and scale entrywise, but their product is defined so that multiplication corresponds to composition of linear maps: the columns of AB are A applied to the columns of B. From that requirement follow the row-column rule, the algebra of products (associative and distributive but not commutative), powers, and the transpose.\n",{"path":25354,"title":25355,"module":25351,"summary":25356},"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fmatrix-inverse-and-invertibility","The Inverse and the Invertible Matrix Theorem","The inverse of a square matrix is the matrix analogue of a reciprocal, defined by AA⁻¹ = I. A closed form settles the 2×2 case; the Gauss–Jordan algorithm row reduces [A | I] to [I | A⁻¹] in general; and elementary matrices record single row operations. The Invertible Matrix Theorem collects a dozen equivalent conditions for invertibility into one statement.\n",{"path":25358,"title":25359,"module":25351,"summary":25360},"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fpartitioned-matrices-and-lu","Block Matrices and the LU Factorization","Partitioning a matrix into blocks lets sums, products, and inverses be computed block by block, as if the submatrices were scalars. Block structure also underlies the LU factorization A = LU, which splits solving Ax = b into two fast triangular solves and repays the cost whenever many systems share one coefficient matrix.\n",{"path":25362,"title":25363,"module":25351,"summary":25364},"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fsubspaces-dimension-rank","Subspaces of Rⁿ, Dimension, and Rank","A subspace is a set closed under addition and scalar multiplication. Every matrix carries two: the column space of all attainable outputs Ax, and the null space of all solutions of Ax = 0. A basis measures each with a minimal spanning set, dimension counts it, and the Rank Theorem ties pivots and free variables together as rank + nullity = n.\n",{"path":25366,"title":25367,"module":25351,"summary":25368},"\u002Flinear-algebra\u002Fmatrix-algebra\u002Fapplications-leontief-and-graphics","Applications: Leontief Economics and Computer Graphics","The Leontief input–output model balances an economy through (I − C)x = d and expands the inverse as a geometric series in the consumption matrix. Computer graphics moves figures with matrix products, using homogeneous coordinates so that translation and perspective projection become matrix multiplications too.\n",{"path":25370,"title":25371,"module":25372,"summary":25373},"\u002Flinear-algebra\u002Fdeterminants\u002Fdeterminants-and-cofactors","Introduction to Determinants","Determinants","The determinant of a square matrix is defined recursively by cofactor expansion: an n-by-n determinant is a signed sum of (n-1)-by-(n-1) determinants built from the first row. The expansion can equally run along any row or down any column, and a triangular matrix has determinant equal to the product of its diagonal.\n",{"path":25375,"title":25376,"module":25372,"summary":25377},"\u002Flinear-algebra\u002Fdeterminants\u002Fproperties-of-determinants","Properties of Determinants","Row operations act on the determinant in three predictable ways, and this turns row reduction into a fast algorithm: the determinant is the product of the pivots times a sign for the interchanges. The same properties yield the invertibility test det A is nonzero, the transpose identity, and the multiplicative law det(AB) equals det A times det B.\n",{"path":25379,"title":25380,"module":25372,"summary":25381},"\u002Flinear-algebra\u002Fdeterminants\u002Fcramer-volume-and-area","Cramer's Rule, Volume, and Linear Transformations","Cramer's rule writes each unknown of an invertible system as a ratio of determinants, and the same idea gives a closed formula for the inverse through the adjugate. Geometrically the absolute determinant is the area of the parallelogram or the volume of the parallelepiped spanned by the columns, so a linear map scales every region's measure by that factor.\n",{"path":25383,"title":25384,"module":25385,"summary":25386},"\u002Flinear-algebra\u002Fvector-spaces\u002Fvector-spaces-and-subspaces","Vector Spaces and Subspaces","Vector Spaces","A vector space is any set closed under addition and scalar multiplication that obeys ten algebraic axioms. The same axioms that govern arrows in the plane govern polynomials, functions, matrices, and infinite signals, so one theory covers them all. A subspace is a subset that is a vector space in its own right, tested by three conditions, and the span of any set of vectors is the smallest subspace containing them.\n",{"path":25388,"title":25389,"module":25385,"summary":25390},"\u002Flinear-algebra\u002Fvector-spaces\u002Fnull-and-column-spaces","Null Spaces, Column Spaces, and Linear Transformations","Two subspaces sit inside every matrix. The null space collects all solutions of $Ax = 0$ and lives in the domain; the column space collects every attainable $Ax$ and lives in the codomain. One is defined implicitly by a condition, the other explicitly by a spanning set, and the same pair appears for an abstract linear transformation as its kernel and range.\n",{"path":25392,"title":25393,"module":25385,"summary":25394},"\u002Flinear-algebra\u002Fvector-spaces\u002Fbases-and-independent-sets","Linearly Independent Sets and Bases","A basis is a spanning set with no redundancy: linearly independent and still large enough to reach every vector. The spanning-set theorem shows any spanning set can be trimmed to a basis by discarding dependent vectors, and the pivot columns of a matrix give a basis for its column space. Independence and spanning are defined for abstract spaces exactly as in $\\mathbb{R}^n$.\n",{"path":25396,"title":25397,"module":25385,"summary":25398},"\u002Flinear-algebra\u002Fvector-spaces\u002Fcoordinate-systems","Coordinate Systems","Fixing a basis assigns every vector a unique list of coordinates, turning an abstract space into $\\mathbb{R}^n$. The coordinate mapping is a one-to-one linear transformation onto $\\mathbb{R}^n$ — an isomorphism — so any $n$-dimensional space is indistinguishable from $\\mathbb{R}^n$ as far as vector-space computations go. In $\\mathbb{R}^n$ the change-of-coordinates matrix $P_B$ and its inverse convert between basis coordinates and standard coordinates.\n",{"path":25400,"title":25401,"module":25385,"summary":25402},"\u002Flinear-algebra\u002Fvector-spaces\u002Fdimension-and-rank","The Dimension of a Vector Space and Rank","Every basis of a space has the same number of vectors, and that number is the dimension. Rank is the dimension of the column space, equal to the dimension of the row space and to the number of pivots. The Rank Theorem, rank plus nullity equals the number of columns, ties the four fundamental subspaces of a matrix together and adds six lines to the Invertible Matrix Theorem.\n",{"path":25404,"title":25405,"module":25385,"summary":25406},"\u002Flinear-algebra\u002Fvector-spaces\u002Fchange-of-basis","Change of Basis","Two bases give the same vector two different coordinate vectors, and a single invertible matrix converts between them. Its columns are the coordinate vectors of the old basis expressed in the new one, and its inverse reverses the conversion. In $\\mathbb{R}^n$ the change-of-coordinates matrix between two bases is found by one row reduction.\n",{"path":25408,"title":25409,"module":25385,"summary":25410},"\u002Flinear-algebra\u002Fvector-spaces\u002Fdifference-equations-and-markov","Applications: Difference Equations and Markov Chains","The solutions of an nth-order linear difference equation form an $n$-dimensional vector space, so finding $n$ independent solutions gives them all. A Markov chain evolves a probability distribution by repeated multiplication by a stochastic matrix, and a regular chain converges to a unique steady-state vector fixed by that matrix. Both applications turn a dynamic process into a subspace or a fixed-point question.\n",{"path":25412,"title":25413,"module":25414,"summary":25415},"\u002Flinear-algebra\u002Feigenvalues\u002Feigenvectors-and-eigenvalues","Eigenvectors and Eigenvalues","Eigenvalues and Eigenvectors","An eigenvector of a square matrix is a nonzero vector the matrix only stretches; its eigenvalue is the stretch factor. The eigenspace of an eigenvalue is the null space of A minus lambda times the identity, the eigenvalues of a triangular matrix are its diagonal entries, and eigenvectors for distinct eigenvalues are linearly independent.\n",{"path":25417,"title":25418,"module":25414,"summary":25419},"\u002Flinear-algebra\u002Feigenvalues\u002Fthe-characteristic-equation","The Characteristic Equation","The eigenvalues of a matrix are the roots of its characteristic polynomial det(A minus lambda I). This degree-n polynomial carries an algebraic multiplicity at each repeated root, a nonzero determinant is equivalent to zero not being an eigenvalue, and similar matrices share a characteristic polynomial and hence the same eigenvalues.\n",{"path":25421,"title":25422,"module":25414,"summary":25423},"\u002Flinear-algebra\u002Feigenvalues\u002Fdiagonalization","Diagonalization","A matrix is diagonalizable when it factors as A equals P D P inverse with D diagonal, which happens exactly when it has n linearly independent eigenvectors. The factorization computes matrix powers cheaply, distinct eigenvalues guarantee it, and a repeated eigenvalue permits it only when its eigenspace dimension equals its multiplicity.\n",{"path":25425,"title":25426,"module":25414,"summary":25427},"\u002Flinear-algebra\u002Feigenvalues\u002Feigenvectors-and-linear-transformations","Eigenvectors and Linear Transformations","Every linear transformation between finite-dimensional spaces has a matrix relative to chosen bases, built from the coordinate vectors of the images of the basis vectors. For a map from a space to itself, an eigenvector basis makes that matrix diagonal, and that change of basis is diagonalization; the matrices similar to A are the representations of the map in every basis.\n",{"path":25429,"title":25430,"module":25414,"summary":25431},"\u002Flinear-algebra\u002Feigenvalues\u002Fcomplex-eigenvalues","Complex Eigenvalues","A real matrix with no real eigenvalues still has complex ones, occurring in conjugate pairs. A real 2-by-2 matrix with eigenvalue a plus b i is similar to a rotation-scaling matrix, whose rotation angle is the argument of the eigenvalue and whose scale factor is its modulus; the modulus decides whether the trajectories close up, spiral in, or spiral out.\n",{"path":25433,"title":25434,"module":25414,"summary":25435},"\u002Flinear-algebra\u002Feigenvalues\u002Fdynamical-systems","Discrete and Continuous Dynamical Systems","Eigenvalues govern the long-term behavior of a system that evolves by x becomes A x or by x prime equals A x. An eigenvector basis decouples both kinds of system into independent scalar equations; the eigenvalues then classify the origin as attractor, repeller, saddle, or spiral, and the dominant eigenpair fixes the growth rate and limiting direction.\n",{"path":25437,"title":25438,"module":25414,"summary":25439},"\u002Flinear-algebra\u002Feigenvalues\u002Fpower-method","Iterative Estimates for Eigenvalues","When only a numerical eigenvalue is needed, iteration is preferred over the characteristic polynomial. The power method repeatedly multiplies by A to converge on the dominant eigenvalue and its eigenvector; the Rayleigh quotient sharpens the estimate for symmetric matrices; and the inverse power method targets any eigenvalue near a known guess.\n",{"path":25441,"title":25442,"module":25443,"summary":25444},"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Finner-product-length-orthogonality","Inner Product, Length, and Orthogonality","Orthogonality and Least Squares","The dot product turns the algebra of vectors in R^n into geometry: length, distance, and perpendicularity. The inner product yields the norm, the Pythagorean theorem, and the orthogonal complement, and the null space of a matrix is the orthogonal complement of its row space.\n",{"path":25446,"title":25447,"module":25443,"summary":25448},"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Forthogonal-sets-and-projections","Orthogonal Sets and Orthogonal Projections","An orthogonal basis makes coordinates trivial: each weight is a single dot product, no linear system required. Orthogonal and orthonormal bases give a direct projection formula onto a line and onto a subspace, the orthogonal decomposition and best-approximation theorems, and the matrix form U U-transpose of a projection.\n",{"path":25450,"title":25451,"module":25443,"summary":25452},"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Fgram-schmidt-and-qr","The Gram-Schmidt Process and QR Factorization","Gram-Schmidt turns any basis into an orthogonal one by repeatedly subtracting off projections onto the span already built. Normalizing the result and recording the coefficients factors the matrix as A = QR, with Q orthonormal and R upper triangular, the factorization behind stable least-squares and eigenvalue algorithms.\n",{"path":25454,"title":25455,"module":25443,"summary":25456},"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Fleast-squares-problems","Least-Squares Problems","When Ax = b has no solution, the least-squares solution makes Ax as close to b as possible. The closest Ax is the projection of b onto the column space, and the vector that produces it solves the normal equations A-transpose A x = A-transpose b. Uniqueness, the residual error, and the stabler QR route follow.\n",{"path":25458,"title":25459,"module":25443,"summary":25460},"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Fleast-squares-applications","Applications to Linear Models","Curve fitting is a least-squares problem in statistical notation. The least-squares line, polynomial fits, and multiple regression all reduce to X beta = y with a design matrix X built from the data, solved by the same normal equations.\n",{"path":25462,"title":25463,"module":25443,"summary":25464},"\u002Flinear-algebra\u002Forthogonality-least-squares\u002Finner-product-spaces","Inner Product Spaces","Promoting the four properties of the dot product to axioms defines an inner product on any vector space, including spaces of functions. Length, distance, orthogonality, Gram-Schmidt, and best approximation all carry over, along with the Cauchy-Schwarz and triangle inequalities and the integral inner product behind Fourier approximation.\n",{"path":25466,"title":25467,"module":25468,"summary":25469},"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fdiagonalizing-symmetric-matrices","Diagonalization of Symmetric Matrices","Symmetric Matrices, Quadratic Forms, and the SVD","A symmetric matrix is one that equals its own transpose. Every such matrix can be diagonalized by an orthogonal change of basis, A = PDPᵀ, with real eigenvalues and perpendicular eigenvectors. This is the Spectral Theorem, and it rewrites A as a weighted sum of rank-one projections onto its eigenvectors.\n",{"path":25471,"title":25472,"module":25468,"summary":25473},"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fquadratic-forms","Quadratic Forms","A quadratic form xᵀAx is the second-degree analogue of a linear map, attached to a symmetric matrix A. Orthogonal diagonalization changes variables to the eigenbasis, removing all cross-terms and rotating the form into standard position. The signs of the eigenvalues then classify it as definite or indefinite.\n",{"path":25475,"title":25476,"module":25468,"summary":25477},"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fconstrained-optimization","Constrained Optimization","Maximizing a quadratic form xᵀAx over the unit sphere has an exact answer: the maximum is the largest eigenvalue of A, attained at its eigenvector, and the minimum is the smallest eigenvalue. Adding orthogonality constraints peels off the eigenvalues in order, characterizing the whole spectrum by optimization.\n",{"path":25479,"title":25480,"module":25468,"summary":25481},"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fsingular-value-decomposition","The Singular Value Decomposition","The singular value decomposition factors any m×n matrix as A = UΣVᵀ, with orthogonal U and V and a nonnegative diagonal Σ of singular values. The singular values are the square roots of the eigenvalues of AᵀA, and they describe the matrix geometrically as a rotation, an axiswise stretch, and another rotation, exposing rank, the four fundamental subspaces, and a best low-rank approximation.\n",{"path":25483,"title":25484,"module":25468,"summary":25485},"\u002Flinear-algebra\u002Fsymmetric-quadratic-svd\u002Fsvd-applications-pca-imaging","Applications: Image Processing and Statistics","Principal component analysis diagonalizes the covariance matrix of a data set, producing uncorrelated variables ordered by variance. The leading components capture most of the variation, which reduces dimension, compresses images through low-rank SVD approximation, and connects directly to the singular values of the data matrix.\n",{"path":25487,"title":25488,"module":25489,"summary":25490},"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fnumerical-thinking-and-matrix-computation","Numerical Thinking and Matrix Computation","Numerical Linear Algebra","Numerical analysis builds efficient discrete algorithms for continuous problems, and its cost is dominated as much by memory traffic as by arithmetic. Block matrix calculus, flop counts, and the BLAS efficiency ratio fix the cost model; triangular and unitary matrices are the two computational building blocks every factorization rests on.\n",{"path":25492,"title":25493,"module":25489,"summary":25494},"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Flu-and-cholesky","LU and Cholesky Factorization in Practice","Gaussian elimination, read as a factorization A = LU, turns a linear system into two triangular solves. A single near-zero pivot wrecks it, so partial pivoting reorders rows to pick the largest available pivot and makes the method work for every invertible matrix. For symmetric positive-definite systems, Cholesky halves the cost and needs no pivoting.\n",{"path":25496,"title":25497,"module":25489,"summary":25498},"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fconditioning-and-floating-point","Conditioning and Floating-Point Arithmetic","A problem's condition number measures how much its answer moves when its data is perturbed, independent of any algorithm. Subtraction is ill-conditioned under cancellation, and for a linear system the amplifier is the matrix condition number κ(A). Floating-point arithmetic supplies the perturbation: every real number is rounded to within a relative machine precision, so even perfect computation inherits an error of order κ times the unit roundoff.\n",{"path":25500,"title":25501,"module":25489,"summary":25502},"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fstability-and-error-analysis","Numerical Stability and Backward Error Analysis","An algorithm is backward stable when its computed answer is the exact answer to a slightly perturbed problem. Combined with the condition number this gives the governing rule of thumb: forward error is at most condition times stability. Three cancellation case studies make the point, then the residual-based backward error applies it to Ax = b and shows why partial pivoting keeps Gaussian elimination stable.\n",{"path":25504,"title":25505,"module":25489,"summary":25506},"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fqr-and-numerical-least-squares","QR, Householder, and Numerical Least Squares","The least-squares problem reduces to the normal equations, but forming AᵀA squares the condition number and can wreck accuracy. The stable route computes a QR factorization directly on A and solves Rx = Qᵀb. Householder reflectors build that QR one column at a time using length-preserving reflections, the unconditionally backward-stable building block behind every serious least-squares solver.\n",{"path":25508,"title":25509,"module":25489,"summary":25510},"\u002Flinear-algebra\u002Fnumerical-linear-algebra\u002Fnumerical-eigenvalues-and-svd","Numerical Eigenvalue Problems and the SVD","Eigenvalues cannot be found by a formula for large matrices, so they are found by iteration. Power and inverse iteration converge to one eigenvector at a rate set by the eigenvalue gap; the QR algorithm sweeps a matrix to Schur form and, with a good shift and a Hessenberg reduction, computes the whole spectrum in cubic time. Singular values follow from the same machinery applied without ever forming AᵀA.\n",{"path":25512,"title":25513,"module":25514,"summary":25515},"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Faffine-combinations","Affine Combinations","Geometry of Vector Spaces","An affine combination is a linear combination whose weights sum to one. The affine hull of a set is the smallest flat containing it: a point, a line, a plane, or a translated subspace. Homogeneous coordinates turn every affine combination into an ordinary linear combination one dimension up.\n",{"path":25517,"title":25518,"module":25514,"summary":25519},"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Faffine-independence-and-barycentric-coordinates","Affine Independence and Barycentric Coordinates","Affine independence is linear independence for the translated or lifted points, and it guarantees each point of an affine hull a unique weight vector. Those weights are barycentric coordinates: centers of mass, ratios of triangle areas, and the interpolation rule behind smooth shading in computer graphics.\n",{"path":25521,"title":25522,"module":25514,"summary":25523},"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Fconvex-combinations-and-convex-sets","Convex Combinations and Convex Sets","A convex combination is an affine combination with nonnegative weights, and the convex hull of a set is the smallest convex set containing it. Convex sets are closed under intersection, and Carathéodory's theorem bounds how many points a convex combination in $\\mathbb{R}^n$ ever needs: at most $n+1$.\n",{"path":25525,"title":25526,"module":25514,"summary":25527},"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Fhyperplanes-and-polytopes","Hyperplanes and Polytopes","A hyperplane is a level set of a linear functional, the set where an inner product equals a constant. Hyperplanes separate disjoint convex sets and support them at their boundaries. Polytopes are convex hulls of finite point sets; their vertices are the extreme points, and a linear functional attains its extremes there.\n",{"path":25529,"title":25530,"module":25514,"summary":25531},"\u002Flinear-algebra\u002Fgeometry-of-vector-spaces\u002Fcurves-and-surfaces","Curves and Surfaces","Bézier curves are affine combinations of control points with polynomial weights, so they lie in the convex hull of those points and bend toward them. The de Casteljau algorithm evaluates them by repeated interpolation, a matrix form factors them for computation, and matching endpoints and tangents joins segments into smooth curves and surfaces.\n",{"path":25533,"title":25534,"module":6,"summary":6},"\u002Flinear-algebra","Linear Algebra",{"path":25536,"title":25537,"module":6,"summary":6},"\u002Ftheory-of-computation","Theory of Computation",{"path":25539,"title":25540,"module":24340,"summary":25541},"\u002Fcomputer-architecture\u002Ffoundations\u002Fbits-bytes-and-words","Bits, Bytes, and Words","Everything a machine stores is a string of bits grouped into bytes. We set out binary and hexadecimal, the byte as the unit of addressing, the word as the machine's natural integer size, and byte ordering — why the same four bytes read as 0x01234567 on one machine and 0x67452301 on another.\n",{"path":25543,"title":25544,"module":24340,"summary":25545},"\u002Fcomputer-architecture\u002Ffoundations\u002Finteger-representation","Integer Representation","A fixed-width byte string is just a pattern; what makes it a number is the rule we read it by. We define unsigned encoding and two's complement — where the top bit carries a negative weight — derive the ranges UMax, TMin, and TMax, and show how the same bits reinterpret between signed and unsigned, how widening sign-extends, and what truncation throws away.\n",{"path":25547,"title":25548,"module":24340,"summary":25549},"\u002Fcomputer-architecture\u002Ffoundations\u002Finteger-arithmetic","Integer Arithmetic","Fixed-width integer arithmetic is arithmetic modulo a power of two: add past the top and the result wraps. We work out unsigned and two's-complement addition and the rules that detect their overflow, why negation is a complement-plus-one, how multiplication truncates to the low-order bits and how compilers turn constant multiplies into shifts and adds, why C declares signed overflow undefined, and the bias fix that keeps shift-based signed division rounding toward zero.\n",{"path":25551,"title":25552,"module":24340,"summary":25553},"\u002Fcomputer-architecture\u002Ffoundations\u002Ffloating-point","Floating Point","IEEE-754 trades the exactness of integers for enormous range by storing numbers as sign, exponent, and fraction — scientific notation in binary. We lay out the single and double formats, the bias that encodes the exponent, the three regimes (normalized, denormalized, special), a worked encode\u002Fdecode, the four rounding modes and round-to-even at the bit level, why addition is not associative, the pitfalls of float-int conversion, and why 0.1 has no exact binary representation.\n",{"path":25555,"title":25556,"module":24340,"summary":25557},"\u002Fcomputer-architecture\u002Ffoundations\u002Fboolean-algebra-and-bit-manipulation","Boolean Algebra and Bit Manipulation","Treat a word as a vector of independent bits and the bitwise operators become an algebra. We define AND, OR, NOT, and XOR as bit vectors, build the masking idioms that set, clear, toggle, and test individual bits, extract fields with zero- and sign-extension, count set bits three ways, derive the classic x & (x - 1) family of tricks, and distinguish bitwise operators from C's short-circuiting logical operators.\n",{"path":25559,"title":25560,"module":25561,"summary":25562},"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fthe-machines-view","The Machine's View","Machine-Level Programming","The instruction set architecture is the contract a compiler writes against: the program counter, sixteen integer registers with their sub-register widths, and the condition codes. We follow one C function down through gcc to assembly, learn to read an instruction as operation plus operands, and fix the vocabulary the rest of the module uses.\n",{"path":25564,"title":25565,"module":25561,"summary":25566},"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fdata-movement","Data Movement","Most instructions a program runs simply move data. We cover the mov family and its size suffixes, the three operand forms, the full memory addressing mode D(Rb,Ri,S) and its special cases, lea for address arithmetic, and how push and pop manipulate the stack pointer %rsp on a stack that grows toward lower addresses.\n",{"path":25568,"title":25569,"module":25561,"summary":25570},"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Farithmetic-and-logic","Arithmetic and Logic","The ALU instructions that compute on register and memory values: add, sub, and imul; the unary inc\u002Fdec\u002Fneg\u002Fnot; the shifts sal\u002Fshr\u002Fsar; the bitwise and\u002For\u002Fxor; and lea reused as a fast arithmetic trick. Each binary operation also sets the condition-code flags CF, ZF, SF, and OF, which cmp and test compute without keeping a result.\n",{"path":25572,"title":25573,"module":25561,"summary":25574},"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fcontrol-flow","Control Flow","How a flat instruction stream realizes branches and loops. The conditional jumps read the condition-code flags; set instructions turn flags into a 0\u002F1 byte. We translate if\u002Felse into the standard compare-and-branch pattern, while\u002Ffor loops into the guarded-do form, and dense switches into jump tables that index a target directly.\n",{"path":25576,"title":25577,"module":25561,"summary":25578},"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fprocedures","Procedures","How a function call works at the machine level: the run-time stack, call and ret passing control through a saved return address, the System V convention that routes the first six arguments through %rdi..%r9 and the result through %rax, the caller-saved versus callee-saved split, the stack frame, and a recursive factorial traced through its frames.\n",{"path":25580,"title":25581,"module":25561,"summary":25582},"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Farrays-structs-and-alignment","Arrays, Structs, and Alignment","How aggregate data lays out in memory. Arrays as base-plus-scaled-index, the row-major ordering of multidimensional arrays, pointer arithmetic in units of the pointed-to type, struct fields at fixed byte offsets, the overlapping storage of unions, and the alignment rules that force padding into a struct.\n",{"path":25584,"title":25585,"module":25561,"summary":25586},"\u002Fcomputer-architecture\u002Fmachine-level-x86-64\u002Fmemory-layout-and-buffer-overflows","Memory Layout and Buffer Overflows","The process address space — text, data, heap, and stack — and the classic vulnerability it enables. A stack buffer that is written past its end can overwrite the saved return address and redirect ret, so we sketch the mechanism defensively and then the three standard protections: stack canaries, a non-executable stack, and address-space layout randomization.\n",{"path":25588,"title":25589,"module":25590,"summary":25591},"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Fwhat-an-isa-is","What an ISA Is","Instruction Set Architecture","The instruction set architecture is the contract that lets a compiler and a chip be written by people who never meet: the stable interface software targets and hardware implements. We separate architecture from microarchitecture, read RISC and CISC as opposite answers to where complexity should live, price out what each choice costs in decode hardware, code density, and pipeline friendliness, and see how x86-64 endures by translating its instructions into RISC-like operations on the fly.\n",{"path":25593,"title":25594,"module":25590,"summary":25595},"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Finstruction-formats-and-operands","Instruction Formats and Operands","An instruction is an opcode plus a way to name its operands. We count operands — 3-address, 2-address, 1-address accumulator, and 0-address stack machines — by writing the same C = A + B four ways, weigh register operands against memory operands, then lay out the same add byte by byte in x86-64 (REX prefix, opcode, ModRM) and in Y86-64, and what fixed versus variable length costs at fetch time.\n",{"path":25597,"title":25598,"module":25590,"summary":25599},"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Faddressing-modes","Addressing Modes","Once an operand field exists, it needs a rule for turning its bits into the data it names. That rule is the addressing mode. We walk the standard set — immediate, register, direct, register-indirect, displacement, scaled-indexed, and PC-relative — fixing the effective-address computation for each, run every mode against one concrete machine state, and price out what Y86-64 loses by keeping only base plus displacement.\n",{"path":25601,"title":25602,"module":25590,"summary":25603},"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Fthe-y86-64-instruction-set","The Y86-64 Instruction Set","Y86-64 is a teaching ISA — a stripped-down x86-64 simple enough to implement by hand yet real enough to compile to. We fix its programmer-visible state (fifteen registers, three condition codes, the PC, memory, and a status code), give the instruction set with exact byte encodings, spell out how the condition codes decide every jXX and cmovXX, and run the encoding both directions: assembly to bytes and raw bytes back to meaning.\n",{"path":25605,"title":25606,"module":25590,"summary":25607},"\u002Fcomputer-architecture\u002Finstruction-set-architecture\u002Fy86-64-programming","Y86-64 Programming","With the encodings fixed, we write real Y86-64 assembly: the .pos, .align, and .quad directives, the calling convention borrowed from x86-64, a stack set up by hand, and complete programs — an array sum and a branch-free max. We watch the assembler turn the listing into the exact byte image the processor will execute, and trace the stack across the call.\n",{"path":25609,"title":25610,"module":25611,"summary":25612},"\u002Fcomputer-architecture\u002Fdigital-logic\u002Ftransistors-gates-and-boolean-functions","Transistors, Gates, and Boolean Functions","Digital Logic","A processor is built from millions of transistor switches. We start at the MOS transistor as a voltage-controlled switch, build the CMOS inverter and NAND transistor by transistor, meet the seven standard gates with their truth tables, show that NAND alone is functionally complete, price each gate in transistors and in time, and turn any truth table into a sum-of-products circuit.\n",{"path":25614,"title":25615,"module":25611,"summary":25616},"\u002Fcomputer-architecture\u002Fdigital-logic\u002Fcombinational-logic-and-hcl","Combinational Logic and HCL","A combinational circuit is a pure Boolean function of its current inputs — no memory, no clock. We draw the line between combinational and sequential logic, do the gate-delay accounting that finds a circuit's critical path and bounds the clock, meet don't-cares, then introduce CS:APP's Hardware Control Language: bit-level operators, word-level signals, equality nets, and the case expression that compiles to a multiplexer tree.\n",{"path":25618,"title":25619,"module":25611,"summary":25620},"\u002Fcomputer-architecture\u002Fdigital-logic\u002Fmultiplexers-decoders-and-the-alu","Multiplexers, Decoders, and the ALU","The combinational building blocks that make a datapath. We build the 2:1 and 4:1 multiplexer and tie it back to HCL's case expression, the n-to-2^n decoder, a one-bit full adder (sum is XOR, carry is majority), the ripple-carry adder that chains them, and finally the ALU — a function unit that selects among add, sub, and, and xor under a control input and exposes condition flags.\n",{"path":25622,"title":25623,"module":25611,"summary":25624},"\u002Fcomputer-architecture\u002Fdigital-logic\u002Fmemory-elements-latches-flip-flops-and-clocking","Memory Elements: Latches, Flip-Flops, and Clocking","A combinational circuit holds no state; feeding a circuit's output back to its input creates memory. We build the SR latch from cross-coupled gates, the level-sensitive D latch, and the master\u002Fslave edge-triggered D flip-flop, then introduce the clock and the synchronous design discipline, the setup\u002Fhold timing window, clock skew, metastability, and the register as n flip-flops sharing one clock.\n",{"path":25626,"title":25627,"module":25611,"summary":25628},"\u002Fcomputer-architecture\u002Fdigital-logic\u002Fregister-files-and-random-access-memory","Register Files and Random-Access Memory","Storage organized for access by address. We build the register file (a small bank of registers with addressed read ports and clocked write ports, the exact structure Y86-64's decode and write-back stages use), then descend to the SRAM and DRAM cells of main memory, why one is fast and dear and the other dense and slow, and how a row decoder picks a word out of a memory array.\n",{"path":25630,"title":25631,"module":25632,"summary":25633},"\u002Fcomputer-architecture\u002Fprocessor-design\u002Fthe-fetch-decode-execute-cycle","The Fetch-Decode-Execute Cycle","Processor Design","A processor is a machine that repeats one loop forever: read the next instruction from memory, figure out what it asks for, do it, and advance. We fix the stored-program idea, lay out the datapath at a high level — PC, instruction memory, register file, ALU, data memory — and the control unit that sequences them, break the work into the six stages the rest of the module builds in hardware, and work out exactly how fetch parses variable-length instructions and computes the next PC.\n",{"path":25635,"title":25636,"module":25632,"summary":25637},"\u002Fcomputer-architecture\u002Fprocessor-design\u002Fthe-seq-stages","The SEQ Stages","The six SEQ stages, made exact. For every Y86-64 instruction — halt, nop, the moves, OPq, the jumps, call and ret, pushq and popq — we write down what Fetch, Decode, Execute, Memory, Write-back, and PC update each compute, as per-instruction stage tables with every row justified. Once the tables are filled in, the processor is fully specified; the remaining lessons turn them into wires.\n",{"path":25639,"title":25640,"module":25632,"summary":25641},"\u002Fcomputer-architecture\u002Fprocessor-design\u002Fcontrol-logic-and-sequencing","Control Logic and Sequencing","The stage tables say what each instruction needs; the control logic computes it from icode. We write the HCL for the register-port selections (srcA, srcB, dstE, dstM), the ALU function and input selection, the memory read\u002Fwrite and address, the branch condition, and the next-PC mux — each a case expression on icode that compiles to a mux — and see how one blob of combinational logic serves every instruction at once. We close by contrasting hardwired control with the microprogrammed alternative.\n",{"path":25643,"title":25644,"module":25632,"summary":25645},"\u002Fcomputer-architecture\u002Fprocessor-design\u002Fassembling-seq","Assembling SEQ","We wire the whole thing together. The functional units from digital logic and the control signals from the last lesson assemble into the complete SEQ datapath, laid out the way CS:APP draws it — six stages stacked bottom to top, Fetch at the floor and PC update at the ceiling, signals flowing up the margins. Then the timing analysis: why everything must settle in one cycle, the no-reading-back principle that makes single-cycle execution consistent, and the critical path that sets the clock. We close by walking an OPq and a ret through the assembled machine.\n",{"path":25647,"title":25648,"module":25632,"summary":25649},"\u002Fcomputer-architecture\u002Fprocessor-design\u002Ftracing-a-program","Tracing a Program","To close the module, we take a complete Y86-64 program — a loop that sums 1 through 3 — and run it through SEQ one cycle at a time, recording the PC, the fetched instruction, every stage computation, and the registers, condition codes, and memory after each cycle. Then we examine single cycles in detail: every named signal of an OPq in concrete hex, and a second program whose call and ret we trace through the stack. The traces confirm that the assembled datapath and control logic behave as a processor.\n",{"path":25651,"title":25652,"module":25653,"summary":25654},"\u002Fcomputer-architecture\u002Fpipelining\u002Fpipelining-principles","Pipelining Principles","Pipelining","A processor that runs one instruction to completion before starting the next wastes most of its hardware most of the time. Pipelining splits the work into stages separated by registers so several instructions are in flight at once. We separate throughput from latency, work the 300 ps example through one, two, and three stages, and derive the three ceilings on the gain: uneven stages, register overhead, and the dependencies between instructions.\n",{"path":25656,"title":25657,"module":25653,"summary":25658},"\u002Fcomputer-architecture\u002Fpipelining\u002Ffrom-seq-to-pipe","From SEQ to PIPE","We turn the sequential Y86-64 processor into a pipelined one by inserting pipeline registers between its stages so each cycle holds one instruction per stage. Doing it correctly forces a rearrangement: the next-PC computation must move into Fetch as a prediction, because the later stages that used to compute it are now busy with other instructions. We walk SEQ to SEQ+ to PIPE, spell out exactly what each pipeline register carries, and fix the naming discipline (D_stat versus d_stat) that keeps five in-flight instructions straight.\n",{"path":25660,"title":25661,"module":25653,"summary":25662},"\u002Fcomputer-architecture\u002Fpipelining\u002Fdata-hazards-stalling-and-forwarding","Data Hazards: Stalling and Forwarding","Overlapping instructions collide when a later one needs a value an earlier one has not finished computing: a read-after-write data hazard. We map exactly which instruction distances are dangerous, fix hazards the slow way by stalling (three bubbles), then the fast way by forwarding from five distinct sources into Decode, in a priority order that sequential semantics forces. Forwarding handles almost everything; the load-use hazard still needs exactly one stall.\n",{"path":25664,"title":25665,"module":25653,"summary":25666},"\u002Fcomputer-architecture\u002Fpipelining\u002Fcontrol-hazards-and-branch-prediction","Control Hazards and Branch Prediction","A pipeline must fetch an instruction every cycle, but after a conditional jump or a ret the next address is not yet known: a control hazard. We measure the branch penalty, weigh predict-taken against its alternatives with real loop arithmetic, watch PIPE detect a misprediction in Execute and squash the two wrong-path instructions, and meet the ret hazard, which has nothing to predict and stalls three cycles. A 2-bit counter gives a taste of dynamic prediction.\n",{"path":25668,"title":25669,"module":25653,"summary":25670},"\u002Fcomputer-architecture\u002Fpipelining\u002Fthe-complete-pipe-processor","The Complete PIPE Processor","We assemble the full pipelined Y86-64: five stages, five pipeline registers, forwarding paths, and a small control unit that decides, each cycle, whether to stall or bubble each register. The subtle part is when hazards combine: one pairing hides a genuine bug. A fourth control case reads stat and keeps exceptions precise. Performance reduces to CPI = 1 + lp + mp + rp, worked out to 1.27 with realistic frequencies, and PIPE beats SEQ by several times despite every penalty.\n",{"path":25672,"title":25673,"module":25674,"summary":25675},"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Fstorage-technologies-and-the-latency-gap","Storage Technologies and the Latency Gap","The Memory Hierarchy","No single memory is both fast and large and cheap. We survey the technologies a machine can store bits in — SRAM, DRAM, flash, and rotating disk — open up a DRAM chip to find the row buffer, work a disk access down to the millisecond, and rank everything by speed, density, and cost per bit. Then we watch the processor outrun memory decade after decade. That widening gap is the whole reason a machine stacks fast small storage on top of slow large storage into a hierarchy.\n",{"path":25677,"title":25678,"module":25674,"summary":25679},"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Flocality","Locality","A hierarchy only pays off because programs do not touch memory at random. They reuse recently-used data (temporal locality) and touch nearby data soon after (spatial locality). We make both precise and then quantitative: miss rates for stride-1 and stride-k traversals against a concrete block size, and the loop-order pair on a 2-D array where the same sum misses 16 times one way and 64 times the other — why row-major versus column-major order can change a program's speed by an order of magnitude.\n",{"path":25681,"title":25682,"module":25674,"summary":25683},"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Fcache-memories-direct-mapped","Cache Memories and Direct Mapping","A cache is fast SRAM that holds copies of recently-used blocks of main memory. We fix its organization — S sets, E lines per set, B bytes per block — and the way it dissects an address into tag, set index, and block offset, worked bit by bit on a concrete 16-byte cache. Then we run the direct-mapped (E=1) access algorithm end to end on a seven-access trace: index to a set, compare the tag, hit or miss, evict. Cold and conflict misses fall out of the structure, and a two-array ping-pong shows conflict thrashing and its padding fix.\n",{"path":25685,"title":25686,"module":25674,"summary":25687},"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Fset-associative-and-write-policies","Set-Associative Caches and Write Policies","Give each set several lines and a block has a choice of homes — fewer conflict misses, at the cost of comparing E tags in parallel and choosing a victim to evict. We re-run the direct-mapped ping-pong trace on a 2-way cache and watch the conflicts vanish, weigh LRU against random replacement, then turn to writes: write-through versus write-back with a dirty bit on a hit, write-allocate versus no-write-allocate on a miss, and a worked traffic count showing when each pairing wins.\n",{"path":25689,"title":25690,"module":25674,"summary":25691},"\u002Fcomputer-architecture\u002Fmemory-hierarchy\u002Fcache-performance-and-cache-friendly-code","Cache Performance and Cache-Friendly Code","Turn the cache mechanism into a number. Hit time, miss rate, and miss penalty combine into the average memory access time; we compute AMAT for a two-level hierarchy with real numbers, weigh the design knobs against each other, and read the memory mountain. Then we write cache-friendly code — the matrix-multiply loop-order case study (ijk versus kij, misses counted per iteration) and loop blocking, where cache-sized tiles turn evicted reuse back into hits.\n",{"path":25693,"title":25694,"module":25695,"summary":25696},"\u002Fcomputer-architecture\u002Fvirtual-memory\u002Faddress-spaces-and-translation","Address Spaces and Translation","Virtual Memory","Every process runs as if it owns a private, contiguous span of memory — its virtual address space — while the hardware maps those addresses onto a single shared physical memory. We fix virtual memory's three jobs (a cache for disk, a memory manager, a protection boundary), the page as the unit of mapping, and the MMU replacing the virtual page number while the offset passes through untouched — then run one translation end to end at the bit level and trace the control flow of a page hit against a page fault.\n",{"path":25698,"title":25699,"module":25695,"summary":25700},"\u002Fcomputer-architecture\u002Fvirtual-memory\u002Fpage-tables-and-page-faults","Page Tables and Page Faults","The page table is an array of page-table entries indexed by virtual page number; each entry's valid bit says whether the page is in DRAM, on disk, or unallocated, and its permission, reference, and dirty bits drive protection and replacement. We walk translation as a table lookup, the page fault and demand paging, the clock algorithm the OS uses to approximate LRU, memory mapping and copy-on-write (why fork is cheap), the taxonomy of bad references, and thrashing.\n",{"path":25702,"title":25703,"module":25695,"summary":25704},"\u002Fcomputer-architecture\u002Fvirtual-memory\u002Fthe-tlb-and-multi-level-page-tables","The TLB and Multi-Level Page Tables","A page-table read on every access would double memory traffic; a flat table for a 48-bit space would occupy 512 GB per process. The TLB fixes the first: a small set-associative cache of PTEs inside the MMU whose tag and index come from the VPN. Multi-level page tables fix the second, allocating only the sub-tables a process uses; x86-64 walks four levels with a 9+9+9+9+12 split. We trace one reference end to end through TLB, walk, and cache, and close with the overlap trick that lets the L1 cache start before translation ends.\n",{"path":25706,"title":25707,"module":25708,"summary":25709},"\u002Fcomputer-architecture\u002Fexceptions-and-io\u002Fexceptional-control-flow","Exceptional Control Flow","Exceptions & I\u002FO","Beyond the sequential, branch, and call flow a program controls itself, the hardware can divert the processor in response to events. We sort these into four classes — interrupts (asynchronous, from devices), traps (intentional syscalls), faults (recoverable, like a page fault), and aborts (unrecoverable) — then take the mechanism apart: exception numbers and the table dispatch, what the hardware pushes and why it differs from a procedure call, the divide-error \u002F page-fault \u002F general-protection trio on x86-64, the full syscall round trip with a worked write in assembly, and processes and signals as the abstractions ECF makes possible.\n",{"path":25711,"title":25712,"module":25708,"summary":25713},"\u002Fcomputer-architecture\u002Fexceptions-and-io\u002Finterrupts-and-the-kernel","Interrupts and the Kernel","An I\u002FO device signals completion by raising an interrupt, crossing the privilege boundary from user mode into the kernel. We fix that boundary, follow an interrupt from device through the interrupt controller to its vectored handler, and use the timer interrupt to drive preemptive scheduling and the context switch. Then the I\u002FO mechanics: polling versus interrupt-driven I\u002FO with a cycle count, device registers and memory-mapped I\u002FO versus port I\u002FO, DMA's full transfer walkthrough and its cache hazard, and a disk read traced end to end, from the read syscall to the completion interrupt.\n",{"path":25715,"title":25716,"module":25717,"summary":25718},"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fprocesses-threads-and-parallelism","Processes, Threads, and Parallelism","Multithreading & Multicore","Around 2004 the single core stopped getting faster, and the industry's answer was to hand programmers more cores instead. This lesson builds the vocabulary that shift demands: process versus thread and exactly which hardware state each one owns, concurrency versus parallelism, the three kinds of parallelism a machine can exploit, why Dennard scaling ended and forced the multicore turn, and Amdahl's law — the arithmetic that bounds the speedup those cores can deliver.\n",{"path":25720,"title":25721,"module":25717,"summary":25722},"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fhardware-multithreading","Hardware Multithreading","A pipeline spends much of its life waiting — on cache misses, on dependences, on branches. Hardware multithreading fills the dead cycles with instructions from another thread. We compare coarse-grained switching (change threads on a long stall), fine-grained interleaving (change every cycle), and simultaneous multithreading (mix threads inside a single cycle), work out exactly which hardware a second thread context duplicates and which it shares, and weigh when SMT pays off and when two threads just fight over one cache.\n",{"path":25724,"title":25725,"module":25717,"summary":25726},"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fcache-coherence","Cache Coherence","Give each core its own cache and the same address can live in two places at once, with copies that disagree. We reproduce the stale-copy bug with a two-core trace, then fix it the way hardware does: snooping caches that watch a shared bus and keep every line in a protocol state. We build MSI in full, upgrade it to MESI, contrast invalidation with updating, add coherence misses as the fourth C, and end with false sharing: the performance bug where cores fight over a line while never touching the same byte.\n",{"path":25728,"title":25729,"module":25717,"summary":25730},"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fmemory-consistency-and-synchronization","Memory Consistency and Synchronization","Coherence keeps cores agreeing about one location; consistency is the contract about many. We define sequential consistency, then watch real hardware break it: the store buffer lets a load slip ahead of an older store, and the classic two-thread litmus test ends with both sides reading zero. We state x86-TSO precisely, restore order with mfence, build atomic read-modify-write from the lock prefix, xchg, and cmpxchg, and write a spinlock twice — once naively, once bus-friendly — closing with what lock-free progress actually guarantees.\n",{"path":25732,"title":25733,"module":25717,"summary":25734},"\u002Fcomputer-architecture\u002Fmultithreading-and-multicore\u002Fmulticore-organization","Multicore Organization","Where everything sits on the die. A modern die gives each core private L1 and L2 caches, spreads a shared last-level cache across slices, and wires it all together with a ring or mesh; multi-socket servers add NUMA, where memory is local to one socket and every remote access pays a latency penalty. We walk the floorplan, put numbers on local versus remote latency, meet thread affinity, and account for the two shared resources — coherence traffic and LLC capacity — that decide how far a parallel program scales.\n",{"path":25736,"title":25737,"module":25738,"summary":25739},"\u002Fcomputer-architecture\u002Fcapstone\u002Fthe-whole-machine","The Whole Machine","Capstone","We take one line of C down the whole tower the course built — compiler to assembly, assembly to machine-code bytes, the bytes into the fetch–decode–execute datapath — then trace one load and one add through the pipelined, cached, translated, interruptible machine, each step cross-linked to the lesson that built it. We close with the map of the course as a stack of layers and an accounting of what we simplified: out-of-order execution, superscalar issue, and speculation past the branch predictor.\n",{"path":25741,"title":25742,"module":25738,"summary":25743},"\u002Fcomputer-architecture\u002Fcapstone\u002Fassembling-a-complete-cpu","Assembling a Complete CPU","We bolt the parts the course built — PC, instruction memory and its fetch logic, register file, ALU, condition codes, data memory, and the control unit — into one complete CPU, name the lesson that built each, wire them in a deliberate order, and power the machine on from reset. Then we assemble a real test program (sum a four-element array through a call\u002Fret procedure), give its exact bytes and memory layout, and trace it cycle by cycle to the answer 0xabcdabcdabcd. We close with how to validate such a machine, and what it takes to put two of them on one die.\n",{"path":25745,"title":25746,"module":6,"summary":6},"\u002Fcomputer-architecture","Computer Architecture",{"path":25748,"title":25749,"module":24340,"summary":25750},"\u002Fdifferential-equations\u002Ffoundations\u002Fmodels-and-direction-fields","Models, Direction Fields, and Solution Curves","A differential equation relates an unknown function to its own rates of change. Three first-order models — a falling body, a cooling object, a population under predation — share the form dy\u002Fdt = ay - b; the slope field fixes their equilibria and long-run behavior before any formula is found. Solving the linear case gives the general solution, its integral curves, and the particular solution selected by an initial condition.\n",{"path":25752,"title":25753,"module":24340,"summary":25754},"\u002Fdifferential-equations\u002Ffoundations\u002Fclassification-and-terminology","Classifying Equations: Order, Linearity, ODE vs. PDE","Every solution method targets a specific class of equation, so the first question about any differential equation is which classes it belongs to. Four independent axes sort them: ordinary versus partial, order, linear versus nonlinear, and homogeneous versus nonhomogeneous. Systems, verification of a solution by substitution, and the split between initial and boundary value problems complete the vocabulary.\n",{"path":25756,"title":25757,"module":25758,"summary":25759},"\u002Fdifferential-equations\u002Ffirst-order\u002Flinear-first-order-integrating-factors","Linear Equations and Integrating Factors","First-Order Equations","A first-order linear equation has the unknown and its derivative to the first power only. Multiplying by an integrating factor collapses the left side into a single derivative, and one integration gives the general solution in closed form. The solution exists wherever the coefficients are continuous, and for a constant coefficient it splits into a decaying transient and a steady state set by the forcing.\n",{"path":25761,"title":25762,"module":25758,"summary":25763},"\u002Fdifferential-equations\u002Ffirst-order\u002Fseparable-and-exact","Separable and Exact Equations","Two nonlinear first-order classes solve by direct integration. A separable equation splits so that each variable can be integrated on its own side, giving an implicit relation. An exact equation is the total differential of a hidden potential function, recognized by a symmetry test on its coefficients; when the test fails, an integrating factor can sometimes restore exactness. A change of variable brings homogeneous equations into the separable class.\n",{"path":25765,"title":25766,"module":25758,"summary":25767},"\u002Fdifferential-equations\u002Ffirst-order\u002Fmodeling-first-order","Modeling with First-Order Equations","A rate law is a differential equation. Each first-order model starts from one governing principle: conservation of mass for a mixing tank, proportional change for interest and radioactive decay, Newton's law of cooling, a force balance for a body falling against drag, and Kirchhoff's law for a series circuit. Setting the derivative to zero recovers the steady state, and the transient records how the initial condition relaxes toward it.\n",{"path":25769,"title":25770,"module":25758,"summary":25771},"\u002Fdifferential-equations\u002Ffirst-order\u002Fautonomous-and-population-dynamics","Autonomous Equations, Phase Lines, and Population Dynamics","An autonomous equation y' = f(y) can be analyzed qualitatively without being solved. Its constant solutions are the zeros of f, and the sign of f between them fixes whether nearby solutions rise or fall, which the phase line records as a column of arrows. The logistic and threshold models, constant- and effort-proportional harvesting, and the properties nonlinear equations lose all follow from this reading.\n",{"path":25773,"title":25774,"module":25758,"summary":25775},"\u002Fdifferential-equations\u002Ffirst-order\u002Fexistence-uniqueness-euler","Existence, Uniqueness, and Euler's Method","Existence and uniqueness can be settled before any attempt to solve. The existence-uniqueness theorem gives sufficient conditions on f, and a standard example shows what fails when they do not hold. Picard's successive approximations build the solution as the limit of an iteration, and Euler's method turns the same tangent-line idea into a numerical procedure for the equations no formula reaches.\n",{"path":25777,"title":25778,"module":25758,"summary":25779},"\u002Fdifferential-equations\u002Ffirst-order\u002Ffirst-order-difference-equations","First-Order Difference Equations","A difference equation advances a sequence one index at a time by a rule y_{n+1} = f(y_n). The linear case y_{n+1} = rho*y_n + b solves in closed form and converges to its equilibrium exactly when the ratio has magnitude below one, which underlies compound-interest and loan calculations. The logistic difference equation shows the nonlinear counterpart: an exchange of stability, a cascade of period doublings, and the onset of chaos.\n",{"path":25781,"title":25782,"module":25783,"summary":25784},"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fhomogeneous-constant-coefficients","Homogeneous Equations, the Wronskian, and Real Roots","Second-Order Linear Equations","A second-order linear homogeneous equation with constant coefficients is solved by guessing an exponential and reducing to the quadratic characteristic equation. Two solutions span every solution exactly when their Wronskian is nonzero; that condition, superposition, and Abel's formula give the full structure of the general solution for the case of two distinct real roots.\n",{"path":25786,"title":25787,"module":25783,"summary":25788},"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fcomplex-and-repeated-roots","Complex Roots, Repeated Roots, and Reduction of Order","When the characteristic equation has complex conjugate roots, Euler's formula converts the complex exponentials into a real fundamental set of decaying or growing oscillations. When it has a repeated root, one exponential is lost and reduction of order recovers the missing second solution as $t\\,e^{rt}$. The same substitution $y = v(t)y_1(t)$ finds a second solution from any known one.\n",{"path":25790,"title":25791,"module":25783,"summary":25792},"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fnonhomogeneous-undetermined-coefficients","Nonhomogeneous Equations: Undetermined Coefficients","The general solution of a nonhomogeneous linear equation is a complementary solution plus any one particular solution. When the forcing term is a polynomial, exponential, sine, or cosine, a particular solution can be found by assuming a trial form of the same shape with unknown coefficients and solving for them. The one complication is resonance, handled by multiplying the trial by a power of $t$.\n",{"path":25794,"title":25795,"module":25783,"summary":25796},"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fvariation-of-parameters","Variation of Parameters","Variation of parameters finds a particular solution of any nonhomogeneous linear equation from a fundamental set of the homogeneous one. Replacing the constants in the complementary solution by functions and imposing one convenient constraint reduces the problem to a two-by-two linear system whose solution is expressed through the Wronskian, giving an integral formula that works for forcing terms undetermined coefficients cannot touch.\n",{"path":25798,"title":25799,"module":25783,"summary":25800},"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fmechanical-electrical-vibrations","Mechanical and Electrical Vibrations","A spring-mass-damper obeys a second-order linear equation, and so does a series RLC circuit, with the same mathematics governing both. Free undamped motion is a pure sinusoid; damping adds a decaying envelope with three regimes; periodic forcing produces a transient that dies out and a steady-state oscillation whose amplitude peaks sharply near the natural frequency, the phenomenon of resonance.\n",{"path":25802,"title":25803,"module":25783,"summary":25804},"\u002Fdifferential-equations\u002Fsecond-order-linear\u002Fhigher-order-linear","Higher-Order Linear Equations","The second-order theory extends directly to order $n$: the solution space is $n$-dimensional, spanned by any $n$ solutions with nonzero Wronskian. For constant coefficients the characteristic polynomial has degree $n$, and its roots (counted with multiplicity, real and complex) build the basis by the same rules as before. Coupled oscillators are the natural application that raises the order.\n",{"path":25806,"title":25807,"module":25808,"summary":25809},"\u002Fdifferential-equations\u002Fseries-solutions\u002Fpower-series-ordinary-points","Power Series Solutions Near Ordinary Points","Series Solutions and Special Functions","A linear equation with variable coefficients has no characteristic equation. A power series substituted into the equation matches coefficients to a recurrence relation, which near an ordinary point yields two independent analytic solutions. The radius of convergence is at least the distance from the expansion point to the nearest singular point in the complex plane.\n",{"path":25811,"title":25812,"module":25808,"summary":25813},"\u002Fdifferential-equations\u002Fseries-solutions\u002Fregular-singular-frobenius","Euler Equations, Regular Singular Points, and Frobenius","The Euler equation x^2 y'' + a x y' + b y = 0 is solved outright by y = x^r, and its three root cases fix the behavior at any regular singular point. The Frobenius method multiplies x^r by a power series; the indicial equation chooses the exponents, and equal or integer-separated roots force a logarithm in the second solution. Gauss's hypergeometric equation is the archetype containing most classical functions as special cases.\n",{"path":25815,"title":25816,"module":25808,"summary":25817},"\u002Fdifferential-equations\u002Fseries-solutions\u002Fbessel-and-special-functions","Bessel's Equation, Legendre Polynomials, and Special Functions","Bessel's equation puts the Frobenius machinery through all three of its cases and produces the functions J and Y that govern anything vibrating or diffusing with circular symmetry. The gamma function extends the factorial so that Bessel functions of every order make sense; Legendre's equation, run through the hypergeometric form, yields the polynomials that play the same role in spherical geometry. Orthogonality ties both families to the eigenfunction expansions of Sturm–Liouville theory.\n",{"path":25819,"title":25820,"module":25821,"summary":25822},"\u002Fdifferential-equations\u002Flaplace\u002Flaplace-definition-ivps","The Laplace Transform: Definition, Properties, and Solving IVPs","The Laplace Transform","The Laplace transform sends a function of time to a function of a complex frequency by integrating it against the kernel e^{-st}. Differentiation in t becomes multiplication by s, so a linear constant-coefficient initial value problem turns into an algebraic equation. Existence rests on piecewise continuity and exponential order; the derivative rule folds in the initial data; and inversion runs through a transform table and partial fractions.\n",{"path":25824,"title":25825,"module":25821,"summary":25826},"\u002Fdifferential-equations\u002Flaplace\u002Fstep-impulse-convolution","Step Functions, Discontinuous Forcing, Impulses, and Convolution","The Heaviside step function and the second shifting theorem transform switches and discontinuous forcing into exponential factors on the transform. The Dirac delta idealizes an instantaneous impulse and transforms to a pure exponential. The convolution theorem inverts a product of transforms, writes the forced response as the impulse response convolved with the input, and solves Abel's tautochrone by transform.\n",{"path":25828,"title":25829,"module":25830,"summary":25831},"\u002Fdifferential-equations\u002Fsystems\u002Fmatrices-eigenvalues-review","Matrices, Linear Systems, and the Eigenvalue Toolkit","Systems of First-Order Linear Equations","Any nth-order linear equation, and any coupled collection of them, rewrites as a single first-order system x' = P(t)x + g(t). The matrix and vector algebra behind that form, the eigenvalue problem det(A - λI) = 0 that drives every solution method, and the fundamental theory — superposition, the Wronskian, Abel's theorem — together establish that n independent solutions span all solutions.\n",{"path":25833,"title":25834,"module":25830,"summary":25835},"\u002Fdifferential-equations\u002Fsystems\u002Fconstant-coefficient-systems-phase-portraits","Homogeneous Constant-Coefficient Systems and Phase Portraits","For x' = Ax with A constant, the trial x = ξe^{rt} turns the differential equation into the eigenvalue problem Aξ = rξ. The eigenvalues fix the geometry of the phase plane: real opposite signs give a saddle, real same sign a node, complex a spiral, purely imaginary a center. Worked in the plane, these cases form the eigenvalue-type classification of equilibria.\n",{"path":25837,"title":25838,"module":25830,"summary":25839},"\u002Fdifferential-equations\u002Fsystems\u002Frepeated-eigenvalues-fundamental-matrices","Repeated Eigenvalues, Fundamental Matrices, and Nonhomogeneous Systems","When a repeated eigenvalue supplies too few eigenvectors, a generalized eigenvector supplies the missing solution as ξte^{ρt} + ηe^{ρt}, giving an improper node. A fundamental set packaged as a matrix Φ(t) yields the matrix exponential e^{At}, the propagator mapping initial states to later ones. Variation of parameters solves the nonhomogeneous system x' = Ax + g(t).\n",{"path":25841,"title":25842,"module":25843,"summary":25844},"\u002Fdifferential-equations\u002Fnumerical\u002Feuler-and-runge-kutta","Euler, Improved Euler, and Runge–Kutta","Numerical Methods","Most initial value problems have no closed-form solution, so the solution is approximated on a grid. Euler's method steps along the tangent line, the improved Euler method averages two slopes, and the classical Runge–Kutta method averages four. Each added stage raises the order of accuracy at the cost of more evaluations per step, measured by how the local and global truncation errors scale with the step size.\n",{"path":25846,"title":25847,"module":25843,"summary":25848},"\u002Fdifferential-equations\u002Fnumerical\u002Fmultistep-systems-stability","Multistep Methods, Systems, and Stability","One-step methods discard everything but the last point. Multistep methods fit a polynomial to several past values and integrate it forward: the explicit Adams–Bashforth formulas, the implicit and more accurate Adams–Moulton formulas, and predictor–corrector pairs that combine them. The same rules extend verbatim to systems in vector form. A separate concern is stability: round-off can dominate truncation, and stiff equations force a tiny step for stability even when accuracy would allow a large one.\n",{"path":25850,"title":25851,"module":25852,"summary":25853},"\u002Fdifferential-equations\u002Fnonlinear\u002Fphase-plane-autonomous-stability","The Phase Plane, Critical Points, and Stability","Nonlinear Systems and Stability","Most nonlinear systems cannot be solved in closed form, so they are studied geometrically. The phase plane turns an autonomous planar system into a family of trajectories; the five archetypes of critical point follow from the eigenvalues of the coefficient matrix; the trace-determinant plane reads off type and stability directly; and epsilon-delta definitions make stability, asymptotic stability, and instability precise.\n",{"path":25855,"title":25856,"module":25852,"summary":25857},"\u002Fdifferential-equations\u002Fnonlinear\u002Flocally-linear-and-liapunov","Locally Linear Systems and Liapunov's Method","Near a critical point a nonlinear system looks linear, and the linear part is the Jacobian. The linearization fixes the type and stability of the nonlinear critical point in every case except a center or a repeated eigenvalue. Liapunov's direct method settles those cases and bounds the basin of attraction by constructing an energy-like function, without solving the system.\n",{"path":25859,"title":25860,"module":25852,"summary":25861},"\u002Fdifferential-equations\u002Fnonlinear\u002Fcompeting-species-predator-prey-limit-cycles","Population Models, Limit Cycles, and Chaos","The phase-plane methods apply directly to interacting-population models. Competing species either coexist or drive one another to extinction, decided by a single inequality among the interaction constants; the Lotka-Volterra predator-prey system produces closed population cycles. Limit cycles and the Poincaré-Bendixson theorem, the van der Pol oscillator, and the Lorenz equations with their strange attractor carry the theory into chaos.\n",{"path":25863,"title":25864,"module":25865,"summary":25866},"\u002Fdifferential-equations\u002Fpdes-fourier-bvp\u002Ffourier-series","Fourier Series and Convergence","PDEs, Fourier Series, and Boundary Value Problems","A two-point boundary value problem has nontrivial solutions only at a discrete set of eigenvalues, the same trichotomy that governs a singular linear system. For y'' + lambda y = 0 with zero endpoints the eigenfunctions are sines and cosines, and their orthogonality gives the Euler-Fourier coefficient formulas. The convergence theorem fixes when the series returns the function, the Gibbs phenomenon measures the overshoot at a jump, and even\u002Fodd symmetry produces half-range sine and cosine series.\n",{"path":25868,"title":25869,"module":25865,"summary":25870},"\u002Fdifferential-equations\u002Fpdes-fourier-bvp\u002Fheat-wave-laplace-equations","Separation of Variables: Heat, Wave, and Laplace Equations","Separation of variables replaces a partial differential equation by a pair of ordinary ones joined through a shared separation constant. Applied to the heat equation it produces the eigenvalue problem X'' + lambda X = 0, and the solution assembles as a Fourier series in the eigenfunctions. The same steps solve the wave equation, whose modes are standing waves, and Laplace's equation, the steady-state limit posed on a region rather than an interval.\n",{"path":25872,"title":25873,"module":25865,"summary":25874},"\u002Fdifferential-equations\u002Fpdes-fourier-bvp\u002Fsturm-liouville","Sturm-Liouville Theory","The eigenvalue problem behind separation of variables generalizes to the self-adjoint Sturm-Liouville form. Lagrange's identity makes the operator symmetric, and from that one fact follow real eigenvalues, orthogonal eigenfunctions, and eigenfunction expansions that behave like Fourier series. Singular problems admit Bessel and Legendre functions, and Sturm's separation and comparison theorems describe how the eigenfunctions oscillate.\n",{"path":25876,"title":25877,"module":25878,"summary":25879},"\u002Fdifferential-equations\u002Fhistory-variations\u002Fcalculus-of-variations","The Calculus of Variations","Historical Notes and the Calculus of Variations","Ordinary calculus finds the point where a function is stationary; the calculus of variations finds the whole curve where an integral is stationary. Euler's differential equation is the necessary condition for an extremal, and it becomes integrable in three cases, solving the shortest-path, minimal-surface, and brachistochrone problems. Lagrange multipliers extend the method to isoperimetric constraints, and Hamilton's principle recovers Newton's law from a single stationary integral.\n",{"path":25881,"title":25882,"module":25878,"summary":25883},"\u002Fdifferential-equations\u002Fhistory-variations\u002Fhistorical-notes","Great Problems and the People Who Solved Them","Differential equations grew out of specific problems, not a plan: the invention of calculus by Newton and Leibniz, the Bernoulli brachistochrone challenge, Euler's flood of methods, Lagrange's analytical mechanics, Gauss and Riemann's rigor, Laplace's celestial mechanics, and Poincaré's qualitative theory. Each method descends from a named problem, and reading the subject forward from those problems explains why its parts fit together.\n",{"path":25885,"title":25886,"module":6,"summary":6},"\u002Fdifferential-equations","Differential Equations",{"path":25888,"title":25889,"module":25890,"summary":25891},"\u002Frelativity\u002Ffoundations\u002Fspecial-relativity-postulates","The Postulates of Special Relativity","Foundations of Relativity","Newton's laws are the same in every inertial frame, but Maxwell's are not: the equations of electromagnetism single out one speed, c, and the nineteenth century read that as the speed of light relative to a medium, the ether. The Michelson-Morley experiment looked for Earth's motion through that medium and found nothing. Einstein's two postulates replace the ether, and their first consequence is that simultaneity is frame-dependent.\n",{"path":25893,"title":25894,"module":25890,"summary":25895},"\u002Frelativity\u002Ffoundations\u002Florentz-transformation-spacetime","The Lorentz Transformation and Spacetime","Requiring that a light sphere stay a light sphere in every inertial frame fixes the coordinate change between frames uniquely: the Lorentz transformation, with its factor gamma. Differentiating it gives relativistic velocity addition, which caps composed speeds at c. Plotting the same events on skewed spacetime axes turns the algebra into geometry, with calibration hyperbolae, an invariant interval, and a light cone that sorts events into past, future, and elsewhere.\n",{"path":25897,"title":25898,"module":25890,"summary":25899},"\u002Frelativity\u002Ffoundations\u002Ftime-dilation-length-contraction","Time Dilation, Length Contraction, and Paradoxes","A light clock and the constancy of c give the two headline effects directly: a moving clock runs slow by gamma, and a moving rod is short by the same factor. Cosmic-ray muons reaching sea level are the standing experimental proof. The relativistic Doppler effect adds the time-dilation factor to the classical shift, and the twin and pole-barn paradoxes dissolve once the relativity of simultaneity is taken seriously.\n",{"path":25901,"title":25902,"module":25890,"summary":25903},"\u002Frelativity\u002Ffoundations\u002Frelativistic-momentum-energy","Relativistic Momentum and Energy","Conserving momentum in every inertial frame forces the redefinition p = gamma m u, which diverges as the speed approaches c. Integrating the corresponding force gives the total energy E = gamma m c-squared, whose rest term m c-squared is Einstein's mass-energy equivalence. Energy and momentum join into a four-vector whose invariant length is the rest energy, giving E-squared = (pc)-squared + (m c-squared)-squared, massless particles, and nuclear binding energy.\n",{"path":25905,"title":25906,"module":25890,"summary":25907},"\u002Frelativity\u002Ffoundations\u002Fgeneral-relativity","A Taste of General Relativity","Einstein's happiest thought was that a freely falling observer feels no gravity: a uniform gravitational field is locally indistinguishable from an accelerating frame. That equivalence principle predicts that light bends near a mass, that clocks run slow deep in a gravitational well, that Mercury's orbit precesses, and that radar echoes are delayed. Every prediction has been confirmed, and pushing the redshift to its limit gives the black hole.\n",{"path":25909,"title":25910,"module":25911,"summary":25912},"\u002Frelativity\u002Fspacetime-and-the-lorentz-group\u002Fminkowski-spacetime-and-the-interval","Minkowski Spacetime and the Interval","Spacetime and the Lorentz Group","The Lorentz transformation of the foundations module is repackaged as the geometry of a four-dimensional space whose invariant is not a distance but the spacetime interval. Events, worldlines, and the metric signature define a causal structure that every observer shares. Proper time is the length of a timelike worldline, and the twin paradox becomes the statement that a straight worldline accumulates the most proper time.\n",{"path":25914,"title":25915,"module":25911,"summary":25916},"\u002Frelativity\u002Fspacetime-and-the-lorentz-group\u002Ffour-vectors-and-index-notation","Four-Vectors and Index Notation","The index calculus that the rest of the course runs on. Contravariant and covariant components, the Minkowski metric as the machine that raises and lowers indices, and the Einstein summation convention are assembled into scalar products that are the same in every frame. The four-velocity and four-acceleration follow, together with the identity that the four-velocity has constant invariant length.\n",{"path":25918,"title":25919,"module":25911,"summary":25920},"\u002Frelativity\u002Fspacetime-and-the-lorentz-group\u002Fthe-lorentz-group-and-rapidity","The Lorentz Group and Rapidity","The Lorentz transformations are the linear maps that preserve the Minkowski metric, and they form the group O(1,3). Boosts are hyperbolic rotations parametrized by rapidity, which adds along a line where velocity does not. The boost and rotation generators fix the group's local structure; its four disconnected components are set by two signs; and two non-collinear boosts compose into a boost plus a rotation, the Wigner rotation behind Thomas precession.\n",{"path":25922,"title":25923,"module":25911,"summary":25924},"\u002Frelativity\u002Fspacetime-and-the-lorentz-group\u002Fdoppler-aberration-and-appearance","Doppler, Aberration, and Appearance","Light carries a null four-momentum, and boosting it produces every optical effect of relativity at once. The covariant Doppler formula follows from the transformation of frequency, aberration from the transformation of direction, and the headlight effect from the resulting concentration of light forward. The Terrell-Penrose result shows that a fast object photographs as rotated, not contracted.\n",{"path":25926,"title":25927,"module":25928,"summary":25929},"\u002Frelativity\u002Frelativistic-dynamics\u002Ffour-momentum-force-and-accelerated-motion","Four-Momentum, Four-Force, and Accelerated Motion","Relativistic Dynamics","The four-momentum packages energy and momentum into a single vector whose invariant length is the rest mass. Its proper-time derivative is the four-force, always orthogonal to the four-velocity, and a constant orthogonal four-force produces hyperbolic motion. Constant proper acceleration gives rapidity linear in proper time, the relativistic rocket equation, and the Rindler horizon behind an eternally accelerating observer.\n",{"path":25931,"title":25932,"module":25928,"summary":25933},"\u002Frelativity\u002Frelativistic-dynamics\u002Fparticle-decays-and-two-body-kinematics","Particle Decays and Two-Body Kinematics","Conservation of four-momentum fixes the kinematics of a decay from the masses alone. In the center-of-momentum frame a parent breaks into two daughters with equal and opposite momenta and energies set by the Kallen triangle function. Boosting to the lab opens the decay into a cone, and the invariant mass built from the daughters reconstructs the parent as a peak. Worked cases: the two-photon decay of the neutral pion and a heavy two-body hadronic decay.\n",{"path":25935,"title":25936,"module":25928,"summary":25937},"\u002Frelativity\u002Frelativistic-dynamics\u002Fcollisions-thresholds-and-the-cm-frame","Relativistic Collisions and Threshold Energies","Two-body collisions run on the same conserved four-momentum as decays. The invariant s sets the total energy available in the center-of-momentum frame and therefore the threshold for producing new particles. Fixed-target energy grows only as the square root of beam energy while a collider grows linearly, which is why colliders reach high energy. Compton scattering follows as a worked photon-electron collision giving the wavelength shift.\n",{"path":25939,"title":25940,"module":25928,"summary":25941},"\u002Frelativity\u002Frelativistic-dynamics\u002Fmandelstam-variables-and-invariants","Mandelstam Variables and Lorentz Invariants","For a two-to-two process the three Mandelstam invariants s, t, and u encode all the kinematics in frame-independent form. They obey a single linear constraint, the sum of the four squared masses, so only two are independent. s is the center-of-momentum energy squared, t and u are momentum transfers tied to the scattering angle, and crossing symmetry relates one amplitude across three channels through these variables.\n",{"path":25943,"title":25944,"module":25945,"summary":25946},"\u002Frelativity\u002Fcovariant-electrodynamics\u002Ffour-current-and-the-four-potential","The Four-Current and Four-Potential","Covariant Electromagnetism","Charge density and current combine into a single four-vector whose divergence is charge conservation. The scalar and vector potentials combine likewise into the four-potential, whose gauge freedom fixes to the Lorenz condition, reducing Maxwell's equations for the potentials to a single wave equation sourced by the four-current.\n",{"path":25948,"title":25949,"module":25945,"summary":25950},"\u002Frelativity\u002Fcovariant-electrodynamics\u002Fthe-electromagnetic-field-tensor","The Electromagnetic Field Tensor","The antisymmetric derivative of the four-potential is the field-strength tensor F, gauge invariant by construction, with the electric and magnetic fields as its components. Its dual exchanges E and B, and its two contractions form the Lorentz invariants that classify a field as electric, magnetic, or radiative in every frame.\n",{"path":25952,"title":25953,"module":25945,"summary":25954},"\u002Frelativity\u002Fcovariant-electrodynamics\u002Ftransformation-of-electric-and-magnetic-fields","How E and B Transform","Transforming the field tensor under a boost gives explicit rules for the electric and magnetic fields: components along the motion are unchanged, transverse components mix and pick up a gamma. The field of a uniformly moving charge compresses transversely, and the force between a current and a moving charge shows that magnetism is the relativistic shadow of electrostatics.\n",{"path":25956,"title":25957,"module":25945,"summary":25958},"\u002Frelativity\u002Fcovariant-electrodynamics\u002Fcovariant-maxwell-and-the-stress-energy-tensor","Covariant Maxwell and the Stress–Energy Tensor","Maxwell's four equations collapse into two tensor equations, one sourced by the four-current and one an identity on the field strength, with charge conservation automatic. The Lorentz force becomes a four-vector law, and the field's energy, momentum, and stress assemble into a symmetric, conserved stress–energy tensor — the object that will source gravity.\n",{"path":25960,"title":25961,"module":25962,"summary":25963},"\u002Frelativity\u002Fcurved-spacetime\u002Fthe-equivalence-principle-formalized","The Equivalence Principle","Curved Spacetime","The equality of gravitational and inertial mass promotes to a physical principle in three graded strengths — weak, Einstein, and strong. A freely falling laboratory is locally indistinguishable from an inertial frame, but the qualifier \"locally\" is essential: the size of the patch over which gravity vanishes is set by the tidal field, which no change of frame can remove. Tidal forces are the true, coordinate-independent signature of gravity, and they are what curvature will measure.\n",{"path":25965,"title":25966,"module":25962,"summary":25967},"\u002Frelativity\u002Fcurved-spacetime\u002Fmanifolds-vectors-and-the-metric","Manifolds, Vectors, and the Metric","A manifold is a space that looks locally like flat space, described by overlapping coordinate charts. Tangent vectors are directional derivatives with the coordinate basis vectors as partial-derivative operators; one-forms live in the dual space; and the metric tensor turns a coordinate line element into an invariant length. The 2-sphere and Rindler metrics serve as worked examples, including the coordinate singularities that are artefacts of the chart, not of the geometry.\n",{"path":25969,"title":25970,"module":25962,"summary":25971},"\u002Frelativity\u002Fcurved-spacetime\u002Fcovariant-derivative-and-christoffel-symbols","Parallel Transport and the Covariant Derivative","The ordinary derivative of a vector field is not a tensor, because it subtracts vectors living in different tangent spaces. A connection supplies the missing comparison: the covariant derivative adds Christoffel-symbol correction terms that cancel the coordinate artefacts. Requiring the connection to be torsion-free and to preserve the metric fixes the Christoffel symbols uniquely in terms of derivatives of the metric, giving the Levi-Civita connection that general relativity uses.\n",{"path":25973,"title":25974,"module":25962,"summary":25975},"\u002Frelativity\u002Fcurved-spacetime\u002Fgeodesics-and-the-geodesic-equation","Geodesics and the Newtonian Limit","Free fall is geodesic motion: a freely falling particle follows the straightest possible worldline, obtained either by parallel-transporting its own tangent vector or by extremizing proper time. Both routes give the geodesic equation. Affine parameters, and conserved quantities from symmetries via Killing vectors, make it solvable. In the weak-field slow-motion limit the geodesic equation reproduces Newton's law of gravity, fixing the time-time metric component as the Newtonian potential.\n",{"path":25977,"title":25978,"module":25962,"summary":25979},"\u002Frelativity\u002Fcurved-spacetime\u002Fcurvature-riemann-and-geodesic-deviation","Curvature and the Riemann Tensor","Curvature is the failure of parallel transport to commute: carrying a vector around an infinitesimal loop returns it rotated, and the rotation per unit area is the Riemann tensor. Its symmetries cut the components to twenty in four dimensions. Geodesic deviation makes it the equation of tidal forces, and its contractions — the Ricci tensor, the Ricci scalar, and the divergence-free Einstein tensor — assemble the objects the field equation is built from.\n",{"path":25981,"title":25982,"module":25962,"summary":25983},"\u002Frelativity\u002Fcurved-spacetime\u002Fthe-einstein-field-equations","The Einstein Field Equations","The field equation is assembled from a short list of requirements: a symmetric, divergence-free, second-order geometric tensor set proportional to the stress–energy tensor, with the coefficient fixed by the Newtonian limit. The cosmological constant is the one extra term the requirements allow. The Einstein–Hilbert action gives the same equation from a variational principle, and the coupled system closes the logic of the module: matter curves spacetime, and spacetime tells matter how to move.\n",{"path":25985,"title":25986,"module":25987,"summary":25988},"\u002Frelativity\u002Fthe-schwarzschild-solution\u002Fthe-schwarzschild-metric","The Schwarzschild Metric","The Schwarzschild Solution","The first exact solution of Einstein's equation follows from two assumptions, staticity and spherical symmetry, imposed on the vacuum outside a mass. Solving the vacuum field equations fixes two metric functions and produces the Schwarzschild geometry, whose one length scale is the Schwarzschild radius $r_s = 2GM\u002Fc^2$. Birkhoff's theorem shows this is the only spherical vacuum, and the far field reduces to Newtonian gravity.\n",{"path":25990,"title":25991,"module":25987,"summary":25992},"\u002Frelativity\u002Fthe-schwarzschild-solution\u002Fgeodesics-and-orbits-in-schwarzschild","Orbits in the Schwarzschild Geometry","The two Killing symmetries of the Schwarzschild metric give a conserved energy and angular momentum per unit mass, reducing geodesic motion to a one-dimensional problem in an effective potential. The potential carries an extra attractive $1\u002Fr^3$ term absent from Newton's, which caps the centrifugal barrier, produces an innermost stable circular orbit at $6GM\u002Fc^2$, and makes bound orbits precess instead of closing.\n",{"path":25994,"title":25995,"module":25987,"summary":25996},"\u002Frelativity\u002Fthe-schwarzschild-solution\u002Flight-bending-and-null-geodesics","Null Geodesics and the Photon Sphere","Light follows null geodesics, governed by a photon effective potential with a single unstable maximum at $3GM\u002Fc^2$, the photon sphere. The impact parameter sorts rays into those that escape with a deflection and those captured, with the critical value $b_c = 3\\sqrt{3}\\,GM\u002Fc^2$ dividing them. A grazing ray bends by $4GM\u002F(c^2 b)$, twice the naive Newtonian value, and the critical impact parameter sets the edge of a black hole's shadow.\n",{"path":25998,"title":25999,"module":26000,"summary":26001},"\u002Frelativity\u002Ftests-of-general-relativity\u002Fperihelion-precession-of-mercury","The Perihelion Precession of Mercury","Tests of General Relativity","A single extra term in the Schwarzschild orbit equation, cubic in the inverse radius, keeps a bound orbit from closing. The perturbation advances the perihelion by 6πGM\u002F(c²a(1−e²)) per revolution, which for Mercury is 43 arcseconds per century — exactly the anomaly left after Newtonian planetary perturbations are subtracted. A note on frame dragging closes the lesson.\n",{"path":26003,"title":26004,"module":26000,"summary":26005},"\u002Frelativity\u002Ftests-of-general-relativity\u002Fdeflection-of-light-and-gravitational-lensing","Light Deflection and Gravitational Lensing","A light ray grazing the Sun bends by 4GM\u002F(c²b), exactly twice the value a Newtonian corpuscle would give; the extra factor is the curvature of space. The 1919 eclipse confirmed it. The same bending focuses light from distant sources into Einstein rings, multiple images, and microlensing brightenings, making lensing a direct probe of mass, including mass that emits no light.\n",{"path":26007,"title":26008,"module":26000,"summary":26009},"\u002Frelativity\u002Ftests-of-general-relativity\u002Fgravitational-redshift-and-shapiro-delay","Gravitational Redshift and the Shapiro Delay","A clock deeper in a gravitational well ticks slower, and a photon climbing out loses frequency by the ratio of the metric's time-time components. Pound and Rebka measured the 2.5×10⁻¹⁵ shift over a 22.5-metre tower. Radar signals grazing the Sun return late by about 250 microseconds, the Shapiro delay. Both probe the time part of the metric directly.\n",{"path":26011,"title":26012,"module":26000,"summary":26013},"\u002Frelativity\u002Ftests-of-general-relativity\u002Frelativity-in-technology-gps","Relativity and the Global Positioning System","A GPS satellite clock runs slow by 7 microseconds a day from its orbital speed and fast by 46 from its higher gravitational potential, a net gain of about 38 microseconds a day. Left uncorrected, the timing error would grow into kilometres of position error within a day and exceed navigation tolerance within minutes. The satellites carry a pre-launch frequency offset to cancel it.\n",{"path":26015,"title":26016,"module":26017,"summary":26018},"\u002Frelativity\u002Fblack-holes\u002Fhorizons-and-coordinate-singularities","Horizons and Coordinate Singularities","Black Holes","The Schwarzschild radius is a coordinate singularity, not a curvature singularity: the metric blows up there only because the static coordinates fail, while the geometry stays finite. Eddington–Finkelstein and Kruskal– Szekeres coordinates cross the horizon smoothly and show the light cones tipping toward the center. A freely falling observer reaches the true singularity at r=0 in finite proper time, while a distant observer sees the infall freeze and redden at the horizon.\n",{"path":26020,"title":26021,"module":26017,"summary":26022},"\u002Frelativity\u002Fblack-holes\u002Frotating-and-charged-black-holes","Rotating and Charged Black Holes","A stationary black hole is fixed by three numbers: mass, angular momentum, and charge. The Reissner–Nordström metric adds charge and splits the horizon in two; the Kerr metric adds rotation, drags inertial frames, and wraps the horizon in an ergosphere where nothing can stay still. Inside the ergosphere the Penrose process extracts rotational energy, and the no-hair theorem states that no other detail of the collapsed matter survives.\n",{"path":26024,"title":26025,"module":26017,"summary":26026},"\u002Frelativity\u002Fblack-holes\u002Fblack-hole-thermodynamics","Black-Hole Thermodynamics","The four laws of black-hole mechanics mirror the four laws of thermodynamics term for term, with horizon area playing the role of entropy and surface gravity the role of temperature. Hawking's calculation makes the analogy literal: a black hole radiates at a temperature set by its surface gravity, carries a real entropy proportional to its horizon area, and slowly evaporates. The thermal spectrum raises the information paradox.\n",{"path":26028,"title":26029,"module":26030,"summary":26031},"\u002Frelativity\u002Fgravitational-waves\u002Flinearized-gravity-and-wave-solutions","Linearized Gravity and Wave Solutions","Gravitational Waves","Weak gravity is a small perturbation of flat spacetime, and the linearized Einstein equation in the Lorenz gauge is an ordinary wave equation propagating at the speed of light. The trace-reversed perturbation carries the dynamics, residual gauge freedom fixes the transverse-traceless form, and the two physical polarizations deform a ring of freely falling masses into oscillating ellipses whose fractional size change is the strain.\n",{"path":26033,"title":26034,"module":26030,"summary":26035},"\u002Frelativity\u002Fgravitational-waves\u002Fgeneration-and-the-quadrupole-formula","The Quadrupole Formula","The retarded solution of the linearized field equation gives the field of a moving source, and conservation of mass and momentum forbids monopole and dipole radiation, leaving the mass quadrupole as the leading emitter. The quadrupole formula fixes the strain and the radiated luminosity, and applied to a compact binary it predicts the inspiral chirp of rising frequency and amplitude. The Hulse-Taylor pulsar's orbital decay confirmed it to a fraction of a percent.\n",{"path":26037,"title":26038,"module":26030,"summary":26039},"\u002Frelativity\u002Fgravitational-waves\u002Fdetection-ligo-and-the-first-events","LIGO and the First Detections","A gravitational wave is measured as a differential length change of the two arms of a kilometre-scale Michelson interferometer, a strain of order ten to the minus twenty-one that moves the mirrors by a fraction of a proton radius. GW150914 recorded the inspiral, merger, and ringdown of two black holes, fixing their masses and the energy radiated, and GW170817 with its coincident gamma-ray burst and kilonova opened multimessenger astronomy.\n",{"path":26041,"title":26042,"module":26043,"summary":26044},"\u002Frelativity\u002Fcosmological-bridge\u002Fthe-cosmological-principle-and-flrw-metric","The Cosmological Principle and the FLRW Metric","A Bridge to Cosmology","Homogeneity and isotropy restrict the spacetime of the universe to a single family of metrics: a flat cosmic-time slicing of spatial sections of constant curvature, scaled by a time-dependent factor a(t). This lesson builds the Friedmann–Lemaître–Robertson–Walker metric from those symmetries, separates comoving from proper distance, and derives cosmological redshift as the stretching of wavelengths with the scale factor.\n",{"path":26046,"title":26047,"module":26043,"summary":26048},"\u002Frelativity\u002Fcosmological-bridge\u002Ffriedmann-equations-and-cosmic-dynamics","The Friedmann Equations and Cosmic Dynamics","The Einstein equation applied to the FLRW metric with a perfect-fluid source yields the two Friedmann equations and the conservation law that ties them together. This lesson derives them, defines the critical density and the density parameters that fix the spatial geometry, works out how matter, radiation, and a cosmological constant dilute and drive the expansion, and hands off to a dedicated cosmology subject.\n",{"path":26050,"title":26051,"module":6,"summary":6},"\u002Frelativity","Relativity",{"path":26053,"title":26054,"module":6,"summary":6},"\u002Fphysical-computing","Physical Computing",{"path":26056,"title":26057,"module":26058,"summary":26059},"\u002Fquantum-mechanics\u002Fold-quantum-theory\u002Fblackbody-radiation-and-the-planck-quantum","Blackbody Radiation and the Planck Quantum","Origins of the Quantum","Millikan's oil-drop experiment fixed the electron charge as an indivisible unit, and the spectrum of thermal radiation forced a second, deeper quantum. Classical physics predicts an infinite energy density at short wavelengths; Planck removed the divergence by allowing a cavity oscillator to hold only energies that are integer multiples of hf, the first appearance of the quantum of action.\n",{"path":26061,"title":26062,"module":26058,"summary":26063},"\u002Fquantum-mechanics\u002Fold-quantum-theory\u002Fthe-photoelectric-effect-and-the-photon","The Photoelectric Effect and the Photon","Light shone on a clean metal ejects electrons, but the details defied the wave theory: the electrons' maximum energy depends on the light's frequency, not its brightness, and there is a sharp threshold frequency below which nothing happens. Einstein resolved every anomaly by treating light as a stream of energy quanta hf, each absorbed whole by one electron, and Millikan's measurement of the stopping-potential slope confirmed h to a decade before anyone expected.\n",{"path":26065,"title":26066,"module":26058,"summary":26067},"\u002Fquantum-mechanics\u002Fold-quantum-theory\u002Fx-rays-and-the-compton-effect","X-Rays and the Compton Effect","X-rays are short-wavelength electromagnetic waves produced when fast electrons are braked in a target, and their diffraction by crystals lets Bragg's law measure atomic spacings. Compton then scattered X-rays off electrons and found the wavelength shifted by an amount that only a photon carrying momentum hf\u002Fc could explain, closing the case for the particle nature of light.\n",{"path":26069,"title":26070,"module":26058,"summary":26071},"\u002Fquantum-mechanics\u002Fold-quantum-theory\u002Fthe-old-quantum-theory-bohr-and-sommerfeld","The Old Quantum Theory: Bohr, Sommerfeld, and Correspondence","Between Bohr's 1913 atom and Schrödinger's 1926 equation, physics ran on a provisional recipe: keep classical orbits, but admit only those whose action integral is a whole multiple of Planck's constant. This lesson develops the Wilson-Sommerfeld phase-integral rule, applies it to the oscillator and to the elliptical Kepler orbits of hydrogen, derives Sommerfeld's relativistic fine structure and the quantization of orbit orientation, and shows how the correspondence principle fixed intensities and selection rules. The systematic failures — helium, line intensities, the anomalous Zeeman effect — mark exactly where a theory of orbits had to give way to a theory of waves.\n",{"path":26073,"title":26074,"module":26075,"summary":26076},"\u002Fquantum-mechanics\u002Fmatter-waves\u002Fde-broglie-waves-and-electron-diffraction","De Broglie Waves and Electron Diffraction","The Wave Nature of Matter","In 1924 de Broglie proposed that every particle carries a wave of wavelength h\u002Fp. The hypothesis explains Bohr's quantized orbits as standing waves, and Davisson and Germer, then G. P. Thomson, confirmed it by diffracting electrons from crystals exactly as X-rays diffract. We derive the electron wavelength, work the Bragg analysis of the data, and give the relativistic form.\n",{"path":26078,"title":26079,"module":26075,"summary":26080},"\u002Fquantum-mechanics\u002Fmatter-waves\u002Fwave-packets-and-the-probability-interpretation","Wave Packets and the Probabilistic Wave Function","A single de Broglie wave fills all space, but a particle is localized. Adding many waves of nearby wavelength builds a wave packet that is confined and moves at the group velocity, which equals the particle velocity. Born's rule reads the squared amplitude of the wave function as a probability density, the meaning confirmed by electron interference building up one detection at a time.\n",{"path":26082,"title":26083,"module":26075,"summary":26084},"\u002Fquantum-mechanics\u002Fmatter-waves\u002Fthe-uncertainty-principle","The Uncertainty Principle and Wave-Particle Duality","The packet relations delta-k delta-x about 1 become Heisenberg's principle once momentum is hbar times wave number: position and momentum cannot both be sharp, nor energy and time. The gamma-ray microscope shows the limit is physical, not technical. It fixes the zero-point energy of a confined particle, the size of the hydrogen atom, and the natural width of spectral lines, and it frames the wave-particle duality of all matter and radiation.\n",{"path":26086,"title":26087,"module":26088,"summary":26089},"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fthe-schrodinger-equation-in-one-dimension","The Schrödinger Equation in One Dimension","Wave Mechanics in One Dimension","The wave equation for matter cannot be derived; it is postulated and judged by experiment. We build the time-dependent Schrödinger equation from the de Broglie relations, read Born's probability rule off the complex wave function, and separate the time and space dependence to get the time-independent equation whose bound-state solutions are the stationary states. The five acceptability conditions on the wave function are what force energy to be quantized.\n",{"path":26091,"title":26092,"module":26088,"summary":26093},"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fthe-free-particle-and-wave-packet-dynamics","The Free Particle and Wave-Packet Dynamics","The free particle has no bound states: its stationary solutions are non-normalizable plane waves forming a continuum. Physical states are wave packets built by superposing them, and the superposition is a Fourier transform. We delta-normalize the plane waves, assemble a Gaussian packet, solve for its exact time evolution, and read off the two facts that reconcile the wave picture with mechanics: the packet moves at the group velocity ħk\u002Fm, the classical velocity, and it spreads because its component momenta travel at different speeds.\n",{"path":26095,"title":26096,"module":26088,"summary":26097},"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fparticle-in-infinite-and-finite-square-wells","Particle in Infinite and Finite Square Wells","The infinite square well is the simplest bound-state problem: two boundary conditions quantize the energy into a ladder E_n = n² E_1, and the eigenfunctions are the standing waves of a string fixed at both ends. Relaxing the walls to a finite depth lets the wave function leak into the classically forbidden region, keeps the number of bound states finite, and turns the eigenvalue condition into a transcendental equation solved graphically.\n",{"path":26099,"title":26100,"module":26088,"summary":26101},"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Foperators-expectation-values-and-the-harmonic-oscillator","Operators, Expectation Values, and the Harmonic Oscillator","Measurable quantities are extracted from the wave function as expectation values, and each observable is represented by an operator that acts between Ψ* and Ψ — position by multiplication, momentum by a derivative, energy by the Hamiltonian. Applied to the harmonic oscillator, the machinery yields evenly spaced levels E_n = (n+½)ℏω, Gaussian- times-Hermite eigenfunctions of definite parity, and the selection rule Δn = ±1.\n",{"path":26103,"title":26104,"module":26088,"summary":26105},"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fthe-dirac-delta-potential","The Dirac-Delta Potential: A Single Bound State and Scattering","A potential concentrated at a single point is solvable in closed form and isolates the physics of matching a wave function across a discontinuity. Integrating the Schrödinger equation across the spike gives a jump condition on the derivative; the attractive delta well then supports exactly one bound state, of energy set by the strength alone, while the same spike scatters an incoming beam with a transmission that rises from zero to one. The attractive well and the repulsive barrier scatter identically yet only the well binds.\n",{"path":26107,"title":26108,"module":26088,"summary":26109},"\u002Fquantum-mechanics\u002Fwave-mechanics-1d\u002Fbarrier-penetration-and-quantum-tunneling","Barrier Penetration and Quantum Tunneling","Unbound states scatter rather than bind. A particle meeting a step is partly reflected even when it has more than enough energy to pass, and a particle meeting a barrier taller than its energy has a nonzero chance of appearing on the far side. Matching the wave function across the boundaries gives the reflection and transmission coefficients and the exponential tunneling probability that explains alpha decay, the scanning tunneling microscope, and the ammonia clock.\n",{"path":26111,"title":26112,"module":26113,"summary":26114},"\u002Fquantum-mechanics\u002Fformalism\u002Fhilbert-space-and-dirac-notation","Hilbert Space and Dirac Bra–Ket Notation","The Formalism of Quantum Mechanics","Wave mechanics is one representation of a deeper structure: quantum states are vectors in a complex inner-product space, and observables act on them as linear operators. We build that space from the axioms, introduce Dirac's kets and bras as vectors and the linear functionals that measure them, and identify the wavefunction as the components of an abstract state in the position basis. The resolution of the identity is the single algebraic tool that ties every basis, expansion, and matrix element together.\n",{"path":26116,"title":26117,"module":26113,"summary":26118},"\u002Fquantum-mechanics\u002Fformalism\u002Fobservables-hermitian-operators-and-eigenvalues","Observables, Hermitian Operators, and the Spectral Theorem","Every measurable quantity is represented by a Hermitian operator, and the reason is forced: a measurement needs real eigenvalues, orthogonal eigenvectors, and a complete eigenbasis, and Hermiticity delivers precisely those. We derive those properties from self-adjointness, state the spectral theorem, handle degeneracy, and show that two observables share an eigenbasis precisely when they commute — the algebraic condition behind compatible and incompatible measurements.\n",{"path":26120,"title":26121,"module":26113,"summary":26122},"\u002Fquantum-mechanics\u002Fformalism\u002Fthe-postulates-and-quantum-measurement","The Postulates and Quantum Measurement","With states as vectors and observables as Hermitian operators, the physical content of quantum mechanics reduces to a short list of postulates. We state them precisely, derive the Born probability rule for discrete and continuous spectra, work out projective collapse and its idempotence, compute expectation values and their variance, and state the measurement problem cleanly — the one place the postulates split unitary evolution from measurement without explaining the seam.\n",{"path":26124,"title":26125,"module":26113,"summary":26126},"\u002Fquantum-mechanics\u002Fformalism\u002Fposition-momentum-and-continuous-spectra","Position, Momentum, and Continuous Spectra","Position and momentum are the observables with no normalizable eigenstates: their spectra are continuous, their eigenkets are delta-normalized, and the two are Fourier conjugates. We derive the canonical commutator from the momentum operator, build the continuous-basis machinery (Dirac deltas replacing Kronecker deltas), show the position and momentum wavefunctions are a Fourier-transform pair, and compute expectation values in either representation.\n",{"path":26128,"title":26129,"module":26113,"summary":26130},"\u002Fquantum-mechanics\u002Fformalism\u002Fcommutators-and-the-generalized-uncertainty-principle","Commutators and the Generalized Uncertainty Principle","The commutator of two observables measures the obstruction to sharing an eigenbasis, and it bounds how sharply both can be known at once. We derive the generalized uncertainty relation from the Schwarz inequality, recover the position–momentum bound as a special case, characterize the minimum-uncertainty states that saturate it as Gaussians, and give the energy–time relation its correct reading as a lifetime bound rather than a commutator relation.\n",{"path":26132,"title":26133,"module":26113,"summary":26134},"\u002Fquantum-mechanics\u002Fformalism\u002Ftime-evolution-schrodinger-and-heisenberg-pictures","Time Evolution, Propagators, and the Heisenberg Picture","Time evolution is generated by the Hamiltonian and implemented by a unitary operator that preserves probability. We build that operator, expand a state in stationary states to see why probability densities freeze while phases wind, introduce the propagator, transfer the time dependence onto operators in the Heisenberg picture, and derive Ehrenfest's theorem — which recovers classical equations of motion for expectation values and identifies conserved quantities as observables commuting with the Hamiltonian.\n",{"path":26136,"title":26137,"module":26138,"summary":26139},"\u002Fquantum-mechanics\u002Foscillator-and-symmetry\u002Fladder-operators-and-the-number-states","Ladder Operators and the Number States","The Oscillator Algebraically, and Symmetry","The harmonic oscillator can be solved without touching a differential equation. Factoring the Hamiltonian into a lowering operator and its adjoint turns the spectrum into pure algebra: the commutator relation fixes the ladder, the vacuum condition fixes the ground state, and the energies fall out as equally spaced rungs. The same operators give the matrix elements of position and momentum for free.\n",{"path":26141,"title":26142,"module":26138,"summary":26143},"\u002Fquantum-mechanics\u002Foscillator-and-symmetry\u002Fcoherent-and-squeezed-states","Coherent and Squeezed States","A single number state never moves — its position expectation is pinned at the origin. The superposition that oscillates like a classical particle is the eigenstate of the annihilation operator: the coherent state. It is a displaced vacuum, carries Poissonian photon statistics, saturates the uncertainty bound, and traces a rigid Gaussian orbit in phase space. Squeezing deforms that circle, trading precision in one quadrature for noise in the other.\n",{"path":26145,"title":26146,"module":26138,"summary":26147},"\u002Fquantum-mechanics\u002Foscillator-and-symmetry\u002Fsymmetries-generators-and-conservation-laws","Symmetries, Generators, and Conservation Laws","Every continuous symmetry of a quantum system is a unitary operator built by exponentiating a Hermitian generator: momentum generates translations, angular momentum generates rotations, the Hamiltonian generates time evolution. When a generator commutes with the Hamiltonian, the transformation leaves the dynamics unchanged and the generator is conserved — the quantum form of Noether's theorem — and any symmetry that mixes states within a level forces degeneracy.\n",{"path":26149,"title":26150,"module":26138,"summary":26151},"\u002Fquantum-mechanics\u002Foscillator-and-symmetry\u002Fparity-time-reversal-and-discrete-symmetries","Parity, Time Reversal, and Discrete Symmetries","Parity and time reversal are symmetries no continuous generator can reach. Parity is a unitary involution whose eigenvalues label states even or odd, fixing the dipole selection rules. Time reversal is antiunitary: it conjugates i, flips momenta and spins, and for half-integer spin squares to minus one, which by Kramers' theorem makes every level of a time-reversal-invariant Hamiltonian at least doubly degenerate.\n",{"path":26153,"title":26154,"module":24993,"summary":26155},"\u002Fquantum-mechanics\u002Fangular-momentum\u002Forbital-angular-momentum-and-spherical-harmonics","Orbital Angular Momentum and Spherical Harmonics","Orbital angular momentum is the operator triple built from position and momentum. Its components fail to commute, so no state carries sharp values of more than one of them, but each commutes with the total square. Solving the common eigenvalue problem in spherical coordinates quantizes both the magnitude and the projection and produces the spherical harmonics, the angular part of every central-force wavefunction.\n",{"path":26157,"title":26158,"module":24993,"summary":26159},"\u002Fquantum-mechanics\u002Fangular-momentum\u002Fthe-angular-momentum-algebra","The Angular-Momentum Algebra and Ladder Operators","The eigenvalues of angular momentum follow from the commutation relations alone, with no reference to coordinates or wavefunctions. Raising and lowering operators built from the components generate finite multiplets, force the total quantum number to be a non-negative integer or half-integer, and fix the matrix elements of every component. The half-integer values excluded by orbital motion appear here, and they are what spin realizes.\n",{"path":26161,"title":26162,"module":24993,"summary":26163},"\u002Fquantum-mechanics\u002Fangular-momentum\u002Faddition-of-angular-momenta-and-clebsch-gordan","Addition of Angular Momenta and Clebsch–Gordan Coefficients","Two angular momenta combine into a total whose allowed magnitudes run from the difference to the sum of the parts in integer steps. The change from the uncoupled product basis to the coupled total-angular-momentum basis is carried out with the lowering operator and orthogonality, and its matrix of overlaps is the table of Clebsch–Gordan coefficients. Two spin-halves split into a triplet and a singlet, the prototype for every composite spin.\n",{"path":26165,"title":26166,"module":26167,"summary":26168},"\u002Fquantum-mechanics\u002Fcentral-potentials\u002Fthe-schrodinger-equation-in-three-dimensions","The Schrödinger Equation in Three Dimensions","Central Potentials","A central potential depends only on the distance from a force center, so the three-dimensional Schrödinger equation separates in spherical coordinates. The angular factor is a spherical harmonic; the radial factor obeys a one-dimensional equation with an effective potential whose centrifugal barrier depends on the angular-momentum quantum number. The free particle and the spherical box fix the two limiting cases through the spherical Bessel functions.\n",{"path":26170,"title":26171,"module":26167,"summary":26172},"\u002Fquantum-mechanics\u002Fcentral-potentials\u002Fthe-hydrogen-atom","The Hydrogen Atom","The Coulomb potential turns the radial equation into one whose bound states exist only for a discrete set of energies. A power-series solution truncated to keep the wavefunction normalizable forces the principal quantum number, and the energy comes out proportional to minus one over its square, recovering the Rydberg spectrum. The bound states are the associated Laguerre functions times spherical harmonics, and their energy depends on the principal number alone, giving an n-squared degeneracy larger than rotational symmetry can explain.\n",{"path":26174,"title":26175,"module":26167,"summary":26176},"\u002Fquantum-mechanics\u002Fcentral-potentials\u002Fthe-isotropic-oscillator-and-hidden-symmetry","The Isotropic Oscillator and Hidden Symmetry","The three-dimensional isotropic harmonic oscillator solves in both Cartesian and spherical bases, and the two solutions must agree on the degeneracy of every level. That agreement, and the accidental degeneracy of hydrogen, both come from a symmetry larger than rotation: the oscillator carries an SU(3) invariance built from a conserved quadrupole tensor, and the Coulomb problem carries an SO(4) invariance built from the conserved Runge–Lenz vector. These hidden symmetries pin the degeneracies that rotational invariance alone leaves unexplained.\n",{"path":26178,"title":26179,"module":26180,"summary":26181},"\u002Fquantum-mechanics\u002Fspin\u002Fspin-half-pauli-matrices-and-stern-gerlach","Spin-½, the Pauli Matrices, and Stern–Gerlach","Spin","A silver atom passing through an inhomogeneous magnetic field splits into two beams, not a smear. That single fact fixes the internal angular momentum of the electron to a two-valued quantity with no spatial wavefunction. We build the two-dimensional spin space, the Pauli matrices and their algebra, the spinor for measurement along an arbitrary axis, and the sequential Stern–Gerlach filters that expose measurement disturbance.\n",{"path":26183,"title":26184,"module":26180,"summary":26185},"\u002Fquantum-mechanics\u002Fspin\u002Fspin-in-a-magnetic-field-precession-and-resonance","Spin in a Magnetic Field: Precession and Resonance","A spin coupled to a magnetic field is the simplest nontrivial quantum dynamics. A static field makes the spin expectation precess on a cone at the Larmor frequency while the energy levels split linearly. Adding a weak oscillating field and passing to the rotating frame produces Rabi oscillations and a resonance lineshape — the physics of NMR and ESR, and the driven qubit.\n",{"path":26187,"title":26188,"module":26180,"summary":26189},"\u002Fquantum-mechanics\u002Fspin\u002Ftwo-level-systems-and-the-bloch-sphere","Two-Level Systems and the Bloch Sphere","Every two-state quantum system is a spin-½ in disguise. Its Hamiltonian is an effective magnetic field, its pure states are points on the Bloch sphere, and its unitary evolution is a rigid rotation of that sphere. The same structure produces avoided level crossings, the ammonia inversion doublet and its maser, and the qubit.\n",{"path":26191,"title":26192,"module":26193,"summary":26194},"\u002Fquantum-mechanics\u002Fidentical-particles\u002Fidentical-particles-and-exchange-symmetry","Identical Particles and Exchange Symmetry","Identical Particles","Two electrons carry no label that distinguishes one from the other, and that bare fact reshapes the state space. The exchange operator that swaps particle labels commutes with any Hamiltonian built from identical particles, so its eigenvalue is conserved, and nature admits only its two extremes: totally symmetric states for bosons and totally antisymmetric states for fermions. The antisymmetry forces a statistical correlation, the exchange \"force,\" that keeps fermions apart and draws bosons together even with no interaction between them.\n",{"path":26196,"title":26197,"module":26193,"summary":26198},"\u002Fquantum-mechanics\u002Fidentical-particles\u002Fthe-pauli-principle-atoms-and-the-periodic-table","The Pauli Principle, Atoms, and the Periodic Table","Antisymmetry packaged as a Slater determinant turns the exclusion principle into an operating rule for building atoms. Helium shows the machinery in full: the electron-electron repulsion splits into a direct Coulomb integral and an exchange integral, and the exchange term alone pushes the spin-triplet (orthohelium) below the spin-singlet (parahelium) with no magnetic interaction in sight. Screening, the aufbau order, and Hund's rules then assemble the whole periodic table from the same antisymmetry.\n",{"path":26200,"title":26201,"module":26202,"summary":26203},"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Ftime-independent-perturbation-theory","Time-Independent Perturbation Theory","Approximation Methods for Bound States","Almost no realistic Hamiltonian can be solved exactly. Perturbation theory treats a hard Hamiltonian as a solvable one plus a small correction and expands the eigenvalues and eigenstates in powers of that correction. We derive the first- and second-order energy shifts and the first-order state correction for a nondegenerate level, expose the small-denominator failure that degeneracy forces, and fix it by diagonalizing the perturbation inside the degenerate subspace to find the \"good\" zeroth-order states.\n",{"path":26205,"title":26206,"module":26202,"summary":26207},"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Ffine-structure-and-the-real-hydrogen-atom","Fine Structure and the Real Hydrogen Atom","The Bohr spectrum is only the leading term. Two relativistic corrections of order alpha-squared — the relativistic kinetic-energy correction and spin–orbit coupling, joined by the Darwin term for s states — split the hydrogen levels into fine structure that depends on the total angular momentum j. We derive each shift as a first-order perturbation, combine them into a formula depending only on n and j, and continue down the energy ladder to the Lamb shift and the hyperfine 21 cm line.\n",{"path":26209,"title":26210,"module":26202,"summary":26211},"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Fthe-zeeman-and-stark-effects","The Zeeman and Stark Effects","An atom in an external field is a perturbation problem whose good basis depends on which interaction wins. A magnetic field competes with the internal spin–orbit coupling: the weak-field limit gives the anomalous Zeeman splitting set by the Landé g-factor, the strong-field limit gives the Paschen–Back pattern in the uncoupled basis, and the intermediate regime is a matrix diagonalization. An electric field gives a quadratic shift for the nondegenerate ground state and a linear splitting for the degenerate n = 2 level.\n",{"path":26213,"title":26214,"module":26202,"summary":26215},"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Fthe-variational-method","The Variational Method","The expectation of the Hamiltonian in any trial state is an upper bound on the true ground-state energy. Minimizing that expectation over a parametrized family of trial functions turns the ground-state problem into ordinary calculus and needs no small parameter. We prove the bound, apply it to the helium atom with a screened effective charge, use a two-center trial to predict binding in the hydrogen molecular ion, and extend the method to excited states through orthogonality.\n",{"path":26217,"title":26218,"module":26202,"summary":26219},"\u002Fquantum-mechanics\u002Fapproximation-methods\u002Fthe-wkb-approximation","The WKB Approximation","When the potential varies slowly on the scale of the de Broglie wavelength, the wavefunction is locally a plane wave with a position-dependent wavelength. This semiclassical picture builds the wavefunction from the classical momentum, breaks down at the turning points where the momentum vanishes, and is repaired there by connection formulas. The result recovers the Bohr–Sommerfeld quantization rule with its half-integer correction and gives the exponential tunneling rate through a smooth barrier, the Gamow factor.\n",{"path":26221,"title":26222,"module":6,"summary":6},"\u002Fquantum-mechanics","Quantum Mechanics",{"path":26224,"title":26225,"module":26226,"summary":26227},"\u002Freal-analysis\u002Ffoundations\u002Fsets-logic-functions","Sets, Logic, and Functions","Foundations and the Real Number System","The working language of analysis: quantifiers and the proof patterns (contrapositive, contradiction, induction), sets and their operations, relations and equivalence classes, and functions with their images, injections, surjections, and bijections. Cardinality is measured by bijection, and Cantor's theorem that no set surjects onto its power set forces uncountable sets to exist.\n",{"path":26229,"title":26230,"module":26226,"summary":26231},"\u002Freal-analysis\u002Ffoundations\u002Fordered-fields-completeness","Ordered Fields and the Completeness Axiom","The real numbers are the unique ordered field with the least-upper-bound property. The field and order axioms, the exact failure of the rationals (no supremum for the set of rationals below √2), and completeness as the defining axiom of ℝ lead to the first consequences: the existence of √2, the Archimedean property, and the density of ℚ in ℝ.\n",{"path":26233,"title":26234,"module":26226,"summary":26235},"\u002Freal-analysis\u002Ffoundations\u002Fabsolute-value-bounds","Absolute Value, Bounded Sets, and Inequalities","The absolute value turns the order on ℝ into a notion of distance, with the triangle inequality as the estimate underlying most later proofs. Covered: its algebra, the triangle and reverse-triangle inequalities, and the extension of the sup\u002Finf vocabulary from sets to bounded functions.\n",{"path":26237,"title":26238,"module":26226,"summary":26239},"\u002Freal-analysis\u002Ffoundations\u002Fintervals-uncountability","Intervals, Uncountability, and Decimals","Intervals are classified, and ℝ is proved uncountable two ways: a nested-interval construction and the decimal diagonal argument. Decimal expansions are built as suprema of truncations, which pins the source of their non-uniqueness (the 0.4999… equals 0.5000… identity) and the identification of the rationals with the eventually-repeating expansions. The middle-thirds Cantor set is an uncountable set of measure zero.\n",{"path":26241,"title":26242,"module":26243,"summary":26244},"\u002Freal-analysis\u002Fsequences-series\u002Fsequences-limits","Sequences and Their Limits","Sequences and Series","A sequence is a function on the natural numbers; it converges to a limit when its terms eventually stay within any prescribed tolerance of that number. The epsilon-M definition fixes the order of the quantifiers, and from it the limit is unique, every convergent sequence is bounded, and only the tail matters. Divergence to plus or minus infinity records terms that outgrow every bound.\n",{"path":26246,"title":26247,"module":26243,"summary":26248},"\u002Freal-analysis\u002Fsequences-series\u002Flimit-laws-monotone","Limit Laws and Monotone Convergence","Limits commute with sums, products, quotients, roots, and absolute values and preserve non-strict inequalities, so a limit can be assembled from the limits of its parts without returning to epsilon and M. The squeeze lemma transfers a limit through two envelopes; the monotone convergence theorem produces a limit from boundedness alone; and the ratio test settles the geometric and factorial standard limits.\n",{"path":26250,"title":26251,"module":26243,"summary":26252},"\u002Freal-analysis\u002Fsequences-series\u002Flimsup-bolzano-weierstrass","Subsequences, Limit Superior, and Bolzano–Weierstrass","A bounded sequence need not converge, but it always has convergent subsequences, and its terms cluster between two extreme values. The limit superior and inferior are the limits of the tail suprema and infima; they always exist for a bounded sequence, coincide exactly when it converges, and are its largest and smallest subsequential limits. Bolzano–Weierstrass extracts a convergent subsequence from boundedness alone.\n",{"path":26254,"title":26255,"module":26243,"summary":26256},"\u002Freal-analysis\u002Fsequences-series\u002Fcauchy-completeness","Cauchy Sequences and the Completeness of the Reals","The Cauchy criterion tests convergence without knowing the limit: a sequence converges exactly when its terms eventually all lie within any tolerance of one another. Cauchy sequences are bounded, in the reals Cauchy and convergent are equivalent, and this completeness property is interchangeable with the least-upper-bound axiom — the single feature that separates the real line from the rationals.\n",{"path":26258,"title":26259,"module":26243,"summary":26260},"\u002Freal-analysis\u002Fsequences-series\u002Fseries-convergence","Series and Convergence Tests","A series converges when its sequence of partial sums does, so every fact about sequences transfers. Geometric and telescoping series sum in closed form; the n-th term test rejects series whose terms miss zero, though the harmonic series shows the converse fails; and the comparison test against the geometric and p-series benchmarks settles most nonnegative-term series.\n",{"path":26262,"title":26263,"module":26243,"summary":26264},"\u002Freal-analysis\u002Fsequences-series\u002Fabsolute-conditional-rearrangement","Absolute Convergence, the Ratio and Root Tests, and Rearrangements","Absolute convergence is the strong form of convergence that permits free manipulation; conditional convergence is fragile. Absolute convergence implies convergence, and the ratio and root tests detect it by comparison with the geometric series. The alternating series test supplies conditionally convergent series, Riemann's theorem rearranges any of them to any sum, and Mertens' theorem multiplies series when at least one converges absolutely.\n",{"path":26266,"title":26267,"module":26268,"summary":26269},"\u002Freal-analysis\u002Fmetric-spaces\u002Fmetric-spaces-norms","Metric Spaces, Norms, and Examples","Metric Spaces and Topology","A metric is a function $d(x,y)$ obeying four axioms: nonnegativity, identity of indiscernibles, symmetry, and the triangle inequality. The Euclidean, taxicab, sup, discrete, and great-circle metrics all qualify, as does the sup metric on $C[a,b]$. Every norm induces a metric, and strongly equivalent metrics share the same open sets.\n",{"path":26271,"title":26272,"module":26268,"summary":26273},"\u002Freal-analysis\u002Fmetric-spaces\u002Fopen-closed-sets","Open and Closed Sets, Interior, Closure","Open sets are those in which every point has room to move; closed sets are their complements. From the single ball construction come the topology axioms (arbitrary unions, finite intersections), the interior, closure, and boundary of a set, and the fact that openness is always relative to the ambient space.\n",{"path":26275,"title":26276,"module":26268,"summary":26277},"\u002Freal-analysis\u002Fmetric-spaces\u002Fconvergence-completeness","Convergence, Cauchy Sequences, and Completeness","The $\\varepsilon$-$N$ definition of a limit transfers verbatim to any metric space once $|x-y|$ is replaced by $d(x,y)$. Convergent sequences characterize closed sets and closures; Cauchy sequences and completeness capture spaces with no missing limits, with $\\mathbb{R}^n$ and $C[a,b]$ complete and $\\mathbb{Q}$ and $(0,1]$ not.\n",{"path":26279,"title":26280,"module":26268,"summary":26281},"\u002Freal-analysis\u002Fmetric-spaces\u002Fcompactness","Compactness","A set is compact if every open cover has a finite subcover. In a metric space this is equivalent to sequential compactness and to being complete and totally bounded. Compact sets are closed and bounded; the Heine–Borel theorem gives the converse in $\\mathbb{R}^n$ but nowhere else in general.\n",{"path":26283,"title":26284,"module":26268,"summary":26285},"\u002Freal-analysis\u002Fmetric-spaces\u002Fconnectedness","Connectedness","A space is connected when it cannot be split into two nonempty open pieces. The connected subsets of $\\mathbb{R}$ are precisely the intervals, path- connectedness gives a constructive sufficient condition, and connectedness is a topological invariant preserved by continuous maps, the fact behind the intermediate value theorem.\n",{"path":26287,"title":26288,"module":24687,"summary":26289},"\u002Freal-analysis\u002Fcontinuity\u002Flimits-of-functions","Limits of Functions","The limit of a function at a point is an epsilon–delta condition pinning one value L as the target of f(x) as x approaches c, mirroring the sequence definition with distance replacing index. It is stated only at cluster points of the domain, is unique when it exists, and reduces to sequential limits through the Heine criterion. The algebra of limits and one-sided limits follow from that reduction.\n",{"path":26291,"title":26292,"module":24687,"summary":26293},"\u002Freal-analysis\u002Fcontinuity\u002Fcontinuous-functions","Continuous Functions","A function is continuous at c when its limit there equals its own value, lim f(x) = f(c). The epsilon–delta and sequential forms agree; sums, products, quotients, and compositions of continuous functions are continuous; and the failures split into jump, Dirichlet, popcorn, and removable types. The topological reading is that preimages of open sets are open.\n",{"path":26295,"title":26296,"module":24687,"summary":26297},"\u002Freal-analysis\u002Fcontinuity\u002Fevt-ivt","Extreme and Intermediate Value Theorems","On a closed bounded interval a continuous function attains an absolute maximum and minimum (the extreme value theorem, compactness preserved by continuity) and takes every value between its endpoint values (the intermediate value theorem, connectedness preserved). Both proofs run through Bolzano–Weierstrass and bisection, and yield root-finding, existence of k-th roots, and fixed-point theorems.\n",{"path":26299,"title":26300,"module":24687,"summary":26301},"\u002Freal-analysis\u002Fcontinuity\u002Funiform-continuity","Uniform Continuity","Uniform continuity strengthens continuity by demanding one delta that works at every point of the domain, not a delta re-chosen at each point. It separates x^2 on a compact interval from x^2 on the whole line and from 1\u002Fx near zero; continuity on a closed bounded interval is automatically uniform; uniformly continuous functions preserve Cauchy sequences and extend to endpoints; and Lipschitz continuity is the strongest of the three, through its secant-slope bound.\n",{"path":26303,"title":26304,"module":24687,"summary":26305},"\u002Freal-analysis\u002Fcontinuity\u002Fcontinuity-metric-spaces","Continuity on Metric Spaces","The epsilon–delta definition used only distances, so continuity transfers to maps between metric spaces by replacing absolute values with the two metrics. In this generality continuity still admits a sequential form, preserves compactness and connectedness, is uniform on a compact domain, and reads topologically as preimages of open sets being open, the formulation that defines homeomorphisms.\n",{"path":26307,"title":26308,"module":24687,"summary":26309},"\u002Freal-analysis\u002Fcontinuity\u002Flimits-infinity-monotone","Limits at Infinity and Monotone Functions","Treating infinity as a cluster point extends the epsilon–delta limit to x approaching plus or minus infinity, giving horizontal asymptotes and infinite limits. For monotone functions the one-sided limits always exist as suprema and infima, the discontinuities are jumps and at most countably many, the continuity is equivalent to the image being an interval, and a strictly monotone function always has a continuous inverse.\n",{"path":26311,"title":26312,"module":26313,"summary":26314},"\u002Freal-analysis\u002Fdifferentiation\u002Fthe-derivative","The Derivative","Differentiation","The derivative is the limit of the difference quotient, the slope the secant lines approach as the second point slides into the first. Differentiability forces continuity; linearity and the product, quotient, and chain rules follow from the definition; and a continuous function can fail to be differentiable, as the absolute value does at the origin.\n",{"path":26316,"title":26317,"module":26313,"summary":26318},"\u002Freal-analysis\u002Fdifferentiation\u002Fmean-value-theorem","The Mean Value Theorem","A relative extremum in the interior forces the derivative to vanish; Rolle's theorem and the mean value theorem turn that local fact into global control. The sign of the derivative fixes monotonicity, a bounded derivative yields a Lipschitz bound, and Darboux's theorem shows derivatives have the intermediate value property even where they are discontinuous.\n",{"path":26320,"title":26321,"module":26313,"summary":26322},"\u002Freal-analysis\u002Fdifferentiation\u002Ftaylors-theorem","Taylor's Theorem","Taylor's theorem generalizes the mean value theorem: an n-times differentiable function is matched near a point by a degree-n polynomial, with a Lagrange remainder that names the error exactly through one higher derivative. Iterating the mean value theorem proves it; the second-derivative test is the order-one case; and a smooth non-analytic bump separates a Taylor series from the function it fails to represent.\n",{"path":26324,"title":26325,"module":26313,"summary":26326},"\u002Freal-analysis\u002Fdifferentiation\u002Finverse-function-1d","The Inverse Function Theorem in One Variable","A nonzero derivative certifies a local inverse and fixes its slope. A strictly monotone differentiable function has a differentiable inverse whose derivative is the reciprocal of the original; the inverse function theorem removes the monotonicity hypothesis, and the reciprocal formula constructs nth roots and the logarithm's derivative, failing exactly where the derivative vanishes.\n",{"path":26328,"title":26329,"module":26330,"summary":26331},"\u002Freal-analysis\u002Friemann-integration\u002Fdarboux-integral","Partitions, Darboux Sums, and Integrability","The Riemann Integral","The Riemann integral is defined by trapping the area under a bounded function between under- and over-estimates. Partitions cut the domain into strips; lower and upper Darboux sums bracket the area; refining a partition tightens the bracket. A function is integrable exactly when the bracket can be made arbitrarily thin, and the tagged Riemann-sum limit gives the same number.\n",{"path":26333,"title":26334,"module":26330,"summary":26335},"\u002Freal-analysis\u002Friemann-integration\u002Fintegrability-classes","Which Functions Are Integrable","The Cauchy criterion certifies whole classes of functions as integrable. Continuous functions are integrable because uniform continuity makes every oscillation cap small; monotone functions are integrable because their caps telescope to a single total jump; bounded functions with finitely many discontinuities are integrable by isolating the bad points. The Dirichlet function fails, and the Lebesgue criterion names the exact boundary.\n",{"path":26337,"title":26338,"module":26330,"summary":26339},"\u002Freal-analysis\u002Friemann-integration\u002Fproperties-of-the-integral","Properties of the Integral","The integral is a linear, order-preserving, additive operator on the integrable functions. It splits across subintervals, respects inequalities, bounds the size of a function by the integral of its absolute value, and preserves products. The mean value theorem for integrals identifies the integral with an attained average height on a fixed rectangle.\n",{"path":26341,"title":24742,"module":26330,"summary":26342},"\u002Freal-analysis\u002Friemann-integration\u002Ffundamental-theorem","The fundamental theorem ties the integral to the derivative in two forms. The evaluation form computes a definite integral from any antiderivative; the differentiation form shows the area function has derivative equal to the integrand at points of continuity. Together they make differentiation and integration inverse operations, and yield integration by parts and change of variables.\n",{"path":26344,"title":26345,"module":26330,"summary":26346},"\u002Freal-analysis\u002Friemann-integration\u002Flog-exp-improper","The Logarithm, Exponential, and Improper Integrals","The integral defines transcendental functions. The logarithm is the area under 1\u002Ft, the exponential is its inverse, and their calculus properties follow from the fundamental theorem. Improper integrals extend integration to unbounded intervals and unbounded integrands as limits of proper integrals, with a p-test, a comparison test, absolute versus conditional convergence, and the integral test linking integrals to series.\n",{"path":26348,"title":26349,"module":26350,"summary":26351},"\u002Freal-analysis\u002Ffunction-sequences\u002Fpointwise-uniform-convergence","Pointwise and Uniform Convergence","Sequences and Series of Functions","A sequence of functions has two natural notions of limit. Pointwise convergence fixes each input and takes the limit of numbers; uniform convergence demands one rate that works for every input at once. The uniform norm turns the second into a statement about a single sequence of numbers, and the uniform Cauchy criterion and the Weierstrass M-test let us certify it.\n",{"path":26353,"title":26354,"module":26350,"summary":26355},"\u002Freal-analysis\u002Ffunction-sequences\u002Finterchange-of-limits","Interchange of Limits: Continuity, Integration, Differentiation","Passing to a limit inside a continuity statement, an integral, or a derivative is an interchange of two limits, and the two limits do not always commute. Uniform convergence licenses the first two swaps: the uniform limit of continuous functions is continuous, and the limit of the integrals is the integral of the limit. Differentiation needs uniform convergence of the derivatives, and counterexamples show why each hypothesis is required.\n",{"path":26357,"title":26358,"module":26350,"summary":26359},"\u002Freal-analysis\u002Ffunction-sequences\u002Fpower-series-weierstrass","Power Series and the Weierstrass Approximation Theorem","A power series converges uniformly on every closed subinterval inside its radius of convergence, together with all of its derivatives. That makes it continuous, differentiable, and integrable term by term, so a power series defines an infinitely differentiable function. The Weierstrass approximation theorem then shows that polynomials come uniformly close to any continuous function on a closed bounded interval.\n",{"path":26361,"title":26362,"module":26350,"summary":26363},"\u002Freal-analysis\u002Ffunction-sequences\u002Fpicard-ode","Picard's Existence and Uniqueness Theorem","The Banach fixed-point theorem says a contraction of a complete metric space has exactly one fixed point, found by iterating from any start. Applied to the space of continuous functions with the uniform norm, it proves Picard's theorem: a first-order differential equation with a Lipschitz right-hand side has a unique local solution. Picard iteration constructs that solution explicitly, and worked examples show the Lipschitz condition is not optional.\n",{"path":26365,"title":26366,"module":26367,"summary":26368},"\u002Freal-analysis\u002Fseveral-variables\u002Fdifferentiability-rn","The Derivative of a Map ℝⁿ → ℝᵐ","Functions of Several Variables (Introduction)","The derivative of a map between Euclidean spaces is the linear transformation of vanishing relative error, unique when it exists and represented in coordinates by the Jacobian matrix of partial derivatives. Differentiability forces continuity through a local Lipschitz bound. Existence of the partial derivatives alone does not suffice; continuity of the partials does.\n",{"path":26370,"title":26371,"module":26367,"summary":26372},"\u002Freal-analysis\u002Fseveral-variables\u002Fgradient-chain-rule","Directional Derivatives, the Gradient, and the Chain Rule","The directional derivative measures the rate of change of a scalar field along a chosen heading and equals the derivative applied to that direction. The gradient collects these into a vector that points along steepest ascent and sits orthogonal to level sets. The chain rule composes derivatives by multiplying Jacobians, and a mean value theorem holds for scalar fields but fails for vector-valued maps.\n",{"path":26374,"title":26375,"module":26367,"summary":26376},"\u002Freal-analysis\u002Fseveral-variables\u002Fhigher-derivatives-taylor-extrema","Higher Derivatives, Taylor's Theorem, and Extrema","Iterating the derivative gives a symmetric second derivative, the Hessian, whose mixed partials agree when they are continuous. Taylor's theorem expands a smooth map to any order with a Lagrange-type remainder, and at a critical point the definiteness of the Hessian decides between a local minimum, a local maximum, and a saddle.\n",{"path":26378,"title":26379,"module":26367,"summary":26380},"\u002Freal-analysis\u002Fseveral-variables\u002Finverse-implicit-theorems","The Inverse and Implicit Function Theorems","A nonlinear map with a nonsingular Jacobian is locally invertible, with the inverse's derivative given by the inverse matrix. The contraction mapping principle supplies the local inverse; the implicit function theorem then solves a system for some variables in terms of the rest whenever the relevant Jacobian block is invertible. Worked coordinate changes show both theorems in use.\n",{"path":26382,"title":26383,"module":26367,"summary":26384},"\u002Freal-analysis\u002Fseveral-variables\u002Fmultiple-integrals","Multiple Integrals","The Riemann integral of a bounded function over a closed rectangle in Euclidean space is built from Darboux upper and lower sums on a grid of subrectangles, with the same squeeze criterion that governs the one-variable integral. Continuous integrands are integrable, and a set of content zero can be ignored. Fubini's theorem evaluates a multiple integral as an iterated one in either order, and the indicator trick extends the theory to regions bounded by curves.\n",{"path":26386,"title":26387,"module":6,"summary":6},"\u002Freal-analysis","Real Analysis",{"path":26389,"title":26390,"module":24340,"summary":26391},"\u002Fabstract-algebra\u002Ffoundations\u002Fsets-functions-relations","Sets, Functions, and Equivalence Relations","Algebra is built on three prior notions: the set, the map between sets, and the equivalence relation that reorganizes a set into disjoint classes. Sets, maps (injective, surjective, bijective), fibers and preimages, and the correspondence between equivalence relations and partitions — the one structural fact reused in every later quotient construction.\n",{"path":26393,"title":26394,"module":24340,"summary":26395},"\u002Fabstract-algebra\u002Ffoundations\u002Fintegers-and-modular-arithmetic","The Integers and Modular Arithmetic","The integers carry the template every ring later imitates: well-ordering drives induction, induction drives the division algorithm, and division drives the Euclidean algorithm, gcd, Bézout's identity, and unique factorization into primes. Quotienting by congruence mod n builds the first finite arithmetic, Z\u002FnZ, whose invertible elements form the group of units.\n",{"path":26397,"title":26398,"module":26399,"summary":26400},"\u002Fabstract-algebra\u002Fgroups-and-symmetry\u002Fgroup-axioms-and-first-examples","Group Axioms and First Examples","Groups and Symmetry","A group is a set with one associative operation that has an identity and inverses. We state the axioms, prove that the identity, inverses, and cancellation behave as expected, define the order of a group and of an element, and catalogue the running examples: the integers, the additive group of residues mod n, and the multiplicative group of units mod n.\n",{"path":26402,"title":26403,"module":26399,"summary":26404},"\u002Fabstract-algebra\u002Fgroups-and-symmetry\u002Fdihedral-and-symmetric-groups","Dihedral and Symmetric Groups","The dihedral group D_{2n} is the symmetries of a regular n-gon, generated by a rotation r and a reflection s subject to three relations. The symmetric group S_n is all permutations of n objects, written in cycle notation. Orders, generators and relations, cycle decomposition, the order of a permutation from its cycle type, and the parity that splits S_n in half.\n",{"path":26406,"title":26407,"module":26399,"summary":26408},"\u002Fabstract-algebra\u002Fgroups-and-symmetry\u002Fmatrix-and-quaternion-groups","Matrix and Quaternion Groups","Invertible matrices over a field form the general linear group GL_n(F), with the determinant-one matrices as the subgroup SL_n(F). Over a finite field the order of GL_n(F) has a clean product formula. The quaternion group Q_8 is a second small nonabelian group, distinct from the dihedral group of the same order; its multiplication and subgroup structure sharpen the contrast between the two.\n",{"path":26410,"title":26411,"module":26399,"summary":26412},"\u002Fabstract-algebra\u002Fgroups-and-symmetry\u002Fhomomorphisms-and-group-actions","Homomorphisms, Isomorphisms, and Actions","A homomorphism is a map between groups that respects the operation; an isomorphism is a bijective one, making two groups the same up to relabeling. The kernel and image measure how far a homomorphism is from injective and surjective. A group action realizes a group as permutations of a set, and actions correspond exactly to homomorphisms into a symmetric group, with orbits and stabilizers as the first tools for counting.\n",{"path":26414,"title":26415,"module":26416,"summary":26417},"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fsubgroups-and-substructures","Subgroups and Their Substructures","Subgroups and Quotients","A subgroup is a subset that is a group under the inherited operation. One test decides it: nonempty and closed under the map $(x,y) \\mapsto xy^{-1}$. From an arbitrary subset $A$ we build the centralizer, normalizer, and center, and from an action the stabilizer and kernel, all of them subgroups nested in a fixed chain inside $G$.\n",{"path":26419,"title":26420,"module":26416,"summary":26421},"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fcyclic-groups","Cyclic Groups","A cyclic group is generated by one element. Two facts organize the whole theory: the order of an element equals the order of the subgroup it generates, and cyclic groups of equal order are isomorphic, so $\\mathbb{Z}$ and $\\mathbb{Z}\u002Fn\\mathbb{Z}$ are the only ones. From there the generators ($\\varphi(n)$ of them), the subgroups (one per divisor of $n$), and a fast exponentiation algorithm all follow.\n",{"path":26423,"title":26424,"module":26416,"summary":26425},"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fgeneration-and-subgroup-lattices","Generation and the Lattice of Subgroups","The subgroup generated by a subset $A$ is the smallest subgroup containing it, described top-down as an intersection and bottom-up as the set of words in $A$ and its inverses. Collecting all subgroups and ordering them by containment produces the subgroup lattice, whose Hasse diagram shows the joins, meets, and containment relations among all subgroups.\n",{"path":26427,"title":26428,"module":26416,"summary":26429},"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fcosets-lagrange-and-normal-subgroups","Cosets, Lagrange, and Normal Subgroups","The left cosets of a subgroup partition a group into equal-sized blocks, so the order of a subgroup divides the order of the group: Lagrange's theorem. When the blocks can be multiplied consistently — exactly when the subgroup is normal — they form the quotient group $G\u002FN$. Fermat's and Euler's theorems fall out as index computations.\n",{"path":26431,"title":26432,"module":26416,"summary":26433},"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fisomorphism-theorems","The Isomorphism Theorems","Four theorems relate homomorphisms, quotients, and subgroup lattices. The first identifies the image of a homomorphism with the quotient by its kernel; the second and third compute quotients built from two subgroups and quotients of quotients; the fourth matches the subgroups of $G\u002FN$ with the subgroups of $G$ lying above $N$. Together they make quotient groups computable.\n",{"path":26435,"title":26436,"module":26416,"summary":26437},"\u002Fabstract-algebra\u002Fsubgroups-and-quotients\u002Fcomposition-series-and-the-alternating-group","Composition Series and the Alternating Group","A composition series breaks a finite group into simple quotient factors, and Jordan-Hölder says those factors are unique up to order. This turns classification into two problems: list the simple groups, and describe how to reassemble them. The sign homomorphism splits $S_n$ into even and odd permutations, defining the alternating group $A_n$, simple for $n \\ge 5$.\n",{"path":26439,"title":26440,"module":26441,"summary":26442},"\u002Fabstract-algebra\u002Fgroup-actions-and-sylow\u002Factions-and-cayleys-theorem","Actions, Orbits, and Cayley's Theorem","Group Actions and Sylow Theory","A group action turns abstract elements into permutations of a set. The action splits the set into orbits, and the orbit-stabilizer theorem ties each orbit's size to the index of a stabilizer. Applied to a group acting on itself by left multiplication, this gives Cayley's theorem: every group is a group of permutations.\n",{"path":26444,"title":26445,"module":26441,"summary":26446},"\u002Fabstract-algebra\u002Fgroup-actions-and-sylow\u002Fconjugation-and-the-class-equation","Conjugation and the Class Equation","A group acts on itself by conjugation, and the orbits are the conjugacy classes. Orbit-stabilizer turns the resulting partition into the class equation, which forces every group of prime-power order to have a nontrivial center. Conjugacy in the symmetric group is cycle type, and Burnside's lemma counts orbits by averaging fixed points.\n",{"path":26448,"title":26449,"module":26441,"summary":26450},"\u002Fabstract-algebra\u002Fgroup-actions-and-sylow\u002Fsylow-theorems","The Sylow Theorems","Lagrange's theorem forbids subgroups whose order fails to divide the group order; Sylow's theorems supply a partial converse for prime powers. A Sylow p-subgroup always exists, all of them are conjugate, and their count satisfies two congruence-and-divisibility constraints tight enough to prove many groups non-simple from their order alone.\n",{"path":26452,"title":26453,"module":26441,"summary":26454},"\u002Fabstract-algebra\u002Fgroup-actions-and-sylow\u002Fautomorphisms-and-simple-groups","Automorphisms and Simplicity of Aₙ","Conjugation makes a group act on itself and on its normal subgroups by automorphisms, giving the inner automorphism group G\u002FZ(G) and the embedding of N(H)\u002FC(H) into Aut(H). Characteristic subgroups are those every automorphism fixes, and the automorphism group of a cyclic group is its unit group. The lesson closes by proving the alternating group Aₙ is simple for n ≥ 5.\n",{"path":26456,"title":26457,"module":26458,"summary":26459},"\u002Fabstract-algebra\u002Fproducts-and-group-structure\u002Fdirect-products-and-finite-abelian-groups","Direct Products and Finite Abelian Groups","Products and Group Structure","The direct product assembles a larger group from componentwise copies of smaller ones, and a recognition theorem reverses the process when two normal subgroups meet trivially and span the group. The Fundamental Theorem of Finitely Generated Abelian Groups then classifies every such group by two equivalent invariants, invariant factors and elementary divisors.\n",{"path":26461,"title":26462,"module":26458,"summary":26463},"\u002Fabstract-algebra\u002Fproducts-and-group-structure\u002Fsemidirect-products","Semidirect Products","The semidirect product relaxes the direct product by requiring only one factor to be normal, with the other acting on it through a homomorphism into its automorphism group. This single twisting map lets abelian pieces assemble into non-abelian groups, realizes the dihedral groups as $\\mathbb{Z}_n \\rtimes \\mathbb{Z}_2$, and, with a recognition theorem, classifies groups of several small orders.\n",{"path":26465,"title":26466,"module":26458,"summary":26467},"\u002Fabstract-algebra\u002Fproducts-and-group-structure\u002Fnilpotent-and-solvable-groups","p-Groups, Nilpotent, and Solvable Groups","Finite p-groups have nontrivial center, and iterating the center upward builds the nilpotent groups, which decompose as the direct product of their Sylow subgroups. Iterating the commutator downward builds the solvable groups, whose factors are abelian. The chain cyclic, abelian, nilpotent, solvable orders these classes, and A_5 breaks the last link.\n",{"path":26469,"title":26470,"module":26458,"summary":26471},"\u002Fabstract-algebra\u002Fproducts-and-group-structure\u002Fclassifying-small-groups","Classifying Groups of Small Order","With Sylow's theorem to force normal subgroups, direct and semidirect products to assemble them, and presentations to name the result, every group up to order fifteen can be listed explicitly. Free groups make presentations precise: generators with no relations, from which any group is a quotient by the normal closure of its relations.\n",{"path":26473,"title":26474,"module":26475,"summary":26476},"\u002Fabstract-algebra\u002Fring-theory\u002Frings-definitions-and-examples","Rings: Definitions and Examples","Ring Theory","A ring carries two operations: an abelian group under addition and an associative multiplication linked by the distributive laws. The named special cases — commutative rings, integral domains, division rings, and fields — differ only in how their multiplication behaves. Standard examples include quadratic integer rings, polynomial rings, matrix rings, and group rings.\n",{"path":26478,"title":26479,"module":26475,"summary":26480},"\u002Fabstract-algebra\u002Fring-theory\u002Fideals-quotients-and-homomorphisms","Ideals, Quotient Rings, and Homomorphisms","Ring homomorphisms have kernels that absorb multiplication; such subsets are ideals, and every ideal is the kernel of the projection onto a quotient ring. The quotient construction yields the ring isomorphism theorems and classifies ideals by their quotients: R\u002FI is a field exactly when I is maximal, an integral domain exactly when I is prime.\n",{"path":26482,"title":26483,"module":26475,"summary":26484},"\u002Fabstract-algebra\u002Fring-theory\u002Ffractions-and-the-chinese-remainder-theorem","Fields of Fractions and the CRT","Rings of fractions invert a multiplicatively closed set, enlarging an integral domain into its field of fractions the way Z becomes Q. The Chinese Remainder Theorem splits a quotient by comaximal ideals into a direct product, generalizing Z\u002FmnZ ≅ Z\u002FmZ × Z\u002FnZ and explaining why the Euler function is multiplicative.\n",{"path":26486,"title":26487,"module":26488,"summary":26489},"\u002Fabstract-algebra\u002Ffactorization-and-polynomials\u002Feuclidean-domains-pids-ufds","Euclidean Domains, PIDs, and UFDs","Factorization and Polynomial Rings","Three classes of integral domain, ordered by how much of elementary arithmetic survives: Euclidean domains carry a division algorithm, principal ideal domains make every ideal a single multiple, and unique factorization domains factor every element into irreducibles in one way. We prove the chain ED implies PID implies UFD, the classes are separated by explicit counterexamples, and irreducible and prime coincide exactly in a UFD.\n",{"path":26491,"title":26492,"module":26488,"summary":26493},"\u002Fabstract-algebra\u002Ffactorization-and-polynomials\u002Fpolynomial-rings-over-fields","Polynomial Rings over Fields","When the coefficients form a field, polynomial long division works exactly as it does over the rationals, and it works with a unique quotient and remainder. That single fact makes F[x] a Euclidean domain, hence a PID and a UFD: every ideal is the multiples of one polynomial, roots correspond to linear factors, and F[x]\u002F(f) is a field precisely when f is irreducible.\n",{"path":26495,"title":26496,"module":26488,"summary":26497},"\u002Fabstract-algebra\u002Ffactorization-and-polynomials\u002Fgauss-lemma-and-unique-factorization","Gauss's Lemma and Unique Factorization","A UFD is not a field, so its polynomial ring is not a PID — yet unique factorization survives the passage from R to R[x]. Gauss's lemma supplies the passage: a polynomial that factors over the fraction field already factors over R, once content is factored out. This gives the theorem that R[x] is a UFD whenever R is, so Z[x] and Q[x,y] factor uniquely even though neither is a PID.\n",{"path":26499,"title":26500,"module":26488,"summary":26501},"\u002Fabstract-algebra\u002Ffactorization-and-polynomials\u002Firreducibility-criteria-and-groebner","Irreducibility Criteria and Gröbner Bases","Deciding whether a given polynomial is irreducible, and computing in multivariate polynomial rings. In one variable: the rational root test, reduction modulo a prime, and Eisenstein's criterion. In several variables, where division fails, a monomial order gives leading terms, a Gröbner basis restores a well-defined remainder, and Buchberger's algorithm computes it.\n",{"path":26503,"title":26504,"module":26505,"summary":26506},"\u002Fabstract-algebra\u002Fmodule-theory\u002Fintroduction-to-modules","Introduction to Modules","Module Theory","A module is an abelian group on which a ring acts, generalizing both vector spaces (when the ring is a field) and abelian groups (when the ring is the integers). Submodules, homomorphisms, quotients, and the isomorphism theorems carry over from groups, and an F[x]-module is the same datum as a vector space with a chosen linear operator — the correspondence behind the canonical forms.\n",{"path":26508,"title":26509,"module":26505,"summary":26510},"\u002Fabstract-algebra\u002Fmodule-theory\u002Ffree-modules-and-direct-sums","Generation, Direct Sums, and Free Modules","A generating set spans a module by R-linear combinations; a direct sum decomposes it into independent pieces; a free module has a basis and the universal property that a homomorphism is determined by arbitrary values on that basis. Rank is well defined over a commutative ring, torsion blocks a basis, and every module is a quotient of a free one — a presentation by generators and relations.\n",{"path":26512,"title":26513,"module":26505,"summary":26514},"\u002Fabstract-algebra\u002Fmodule-theory\u002Ftensor-products-and-exact-sequences","Tensor Products and Exact Sequences","The tensor product builds a module in which elements of two modules can be multiplied, characterized by a universal property turning bilinear maps into linear ones; extension of scalars is its guiding case. Exact sequences track how a module is assembled from a submodule and a quotient, when that assembly splits, and which modules — projective, injective, flat — make the Hom and tensor functors preserve exactness.\n",{"path":26516,"title":26517,"module":26505,"summary":26518},"\u002Fabstract-algebra\u002Fmodule-theory\u002Fvector-spaces-and-linear-maps","Vector Spaces and Linear Maps","A vector space is a module over a field, and the field hypothesis removes every pathology a general module can have: every vector space is free, so it has a basis, a well-defined dimension, and a coordinate isomorphism with F^n. Linear maps become matrices, change of basis becomes similarity, every space pairs with a dual of the same dimension, and the determinant is the unique alternating multilinear normalized form.\n",{"path":26520,"title":26521,"module":26522,"summary":26523},"\u002Fabstract-algebra\u002Fmodules-over-pids\u002Fstructure-theorem-over-pids","The Structure Theorem for Modules over a PID","Modules over PIDs and Canonical Forms","Every finitely generated module over a principal ideal domain splits as a free part plus a direct sum of cyclic torsion pieces, in two canonical ways: invariant factors, tied together by a divisibility chain, and elementary divisors, one prime power at a time. Existence follows from the stacked-basis theorem, both lists are unique, and the case $R = \\mathbb{Z}$ is the classification of finitely generated abelian groups.\n",{"path":26525,"title":26526,"module":26522,"summary":26527},"\u002Fabstract-algebra\u002Fmodules-over-pids\u002Frational-canonical-form","Rational Canonical Form","A linear operator turns its vector space into a module over the polynomial ring $F[x]$, with $x$ acting as the operator. The structure theorem's invariant factors then become polynomials, each cyclic summand becomes a companion matrix, and the block-diagonal assembly is the rational canonical form. It is unique, it is computed inside the base field, and two matrices are similar exactly when their rational canonical forms agree.\n",{"path":26529,"title":26530,"module":26522,"summary":26531},"\u002Fabstract-algebra\u002Fmodules-over-pids\u002Fjordan-canonical-form","Jordan Canonical Form","When the base field contains all the eigenvalues, the elementary divisors of an operator are powers of linear polynomials, and each cyclic summand becomes a Jordan block: an eigenvalue on the diagonal with ones just above it. Stacking the blocks gives the Jordan canonical form, unique up to reordering, as close to diagonal as the operator allows. Diagonalizability reads off the minimal polynomial, and the block sizes are counted by ranks of powers of the operator minus the eigenvalue.\n",{"path":26533,"title":26534,"module":26535,"summary":26536},"\u002Fabstract-algebra\u002Ffield-theory\u002Ffield-extensions-and-algebraic-elements","Field Extensions and Algebraic Elements","Field Theory","A field extension makes a larger field K into a vector space over a smaller field F, and its degree [K:F] is that dimension. Adjoining a root of an irreducible polynomial builds a simple extension F(α) isomorphic to F[x]\u002F(m), whose degree is the degree of the minimal polynomial. The tower law makes these degrees multiply, which turns algebra over fields into bookkeeping with integers.\n",{"path":26538,"title":26539,"module":26535,"summary":26540},"\u002Fabstract-algebra\u002Ffield-theory\u002Fstraightedge-and-compass-constructions","Straightedge-and-Compass Constructions","The lengths a straightedge and compass can build from a unit form a field closed under square roots, and every constructible number lies in a tower of quadratic extensions. So its degree over the rationals is a power of two. That single obstruction settles three problems the Greeks left open: doubling the cube, trisecting a general angle, and squaring the circle are all impossible.\n",{"path":26542,"title":26543,"module":26535,"summary":26544},"\u002Fabstract-algebra\u002Ffield-theory\u002Fsplitting-fields-and-algebraic-closure","Splitting Fields and Algebraic Closure","The splitting field of a polynomial is the smallest extension in which it factors into linear pieces, obtained by adjoining all its roots. Every polynomial has one, its degree is at most n factorial, and any two splitting fields are isomorphic. Pushing this to all polynomials at once gives the algebraic closure, a field in which every polynomial splits and which is unique up to isomorphism.\n",{"path":26546,"title":26547,"module":26535,"summary":26548},"\u002Fabstract-algebra\u002Ffield-theory\u002Fseparable-and-cyclotomic-extensions","Separable Extensions and Cyclotomic Fields","A polynomial is separable when its roots are distinct, detected by whether it shares a factor with its formal derivative. Over perfect fields — characteristic zero and finite fields — every irreducible is separable, and the existence and uniqueness of the finite fields follow. Cyclotomic polynomials package the roots of unity by order, are irreducible over the rationals, and give the cyclotomic field its degree phi(n).\n",{"path":26550,"title":26551,"module":26552,"summary":26553},"\u002Fabstract-algebra\u002Fgalois-theory\u002Fthe-galois-correspondence","The Galois Correspondence","Galois Theory","Galois theory attaches to a field extension its group of symmetries and shows that, for the right extensions, the subgroups of that group are in exact order-reversing correspondence with the intermediate fields. The automorphism group, Artin's theorem, the characterization of Galois extensions, and the Fundamental Theorem together turn questions about fields into questions about finite groups.\n",{"path":26555,"title":26556,"module":26552,"summary":26557},"\u002Fabstract-algebra\u002Fgalois-theory\u002Ffinite-fields","Finite Fields","Every finite field has prime-power order, is the splitting field of $x^{p^n} - x$, and is unique up to isomorphism. Its extension over the prime field is Galois with cyclic group generated by the Frobenius map $x \\mapsto x^p$, so the Galois correspondence reduces the subfield lattice to the divisor lattice of $n$. Möbius inversion counts the irreducible polynomials of each degree, and cyclic error-correcting codes are one application.\n",{"path":26559,"title":26560,"module":26552,"summary":26561},"\u002Fabstract-algebra\u002Fgalois-theory\u002Fcyclotomic-and-abelian-extensions","Cyclotomic and Abelian Extensions","The Galois group of the $n$th cyclotomic field over $\\mathbb{Q}$ is the unit group $(\\mathbb{Z}\u002Fn\\mathbb{Z})^\\times$, which makes cyclotomic fields the worked catalogue of abelian extensions of $\\mathbb{Q}$. The isomorphism identifies subfields with subgroups, realizes every finite abelian group as a Galois group over $\\mathbb{Q}$, and leads to Kronecker–Weber. Composites of Galois extensions and the primitive element theorem supply the machinery.\n",{"path":26563,"title":26564,"module":26552,"summary":26565},"\u002Fabstract-algebra\u002Fgalois-theory\u002Fgalois-groups-of-polynomials","Galois Groups of Polynomials","Ordering the roots of a separable polynomial embeds its Galois group in the symmetric group $S_n$, and the group is transitive exactly when the polynomial is irreducible. The discriminant decides membership in $A_n$; for cubics and quartics the resolvent cubic pins the group down; and reduction modulo a prime produces elements of prescribed cycle type, the standard tool for computing Galois groups over $\\mathbb{Q}$.\n",{"path":26567,"title":26568,"module":26552,"summary":26569},"\u002Fabstract-algebra\u002Fgalois-theory\u002Fsolvability-by-radicals-and-the-quintic","Solvability by Radicals and the Quintic","A polynomial is solvable by radicals exactly when its Galois group is solvable. Cyclic extensions are radical extensions once roots of unity are present, which turns a radical tower into a solvable subnormal series. Since $S_n$ is solvable only for $n \\le 4$, the general quintic has no radical formula, and an explicit quintic with Galois group $S_5$ has roots provably not expressible in radicals.\n",{"path":26571,"title":26572,"module":26573,"summary":26574},"\u002Fabstract-algebra\u002Fcapstone\u002Fcommutative-algebra-and-algebraic-geometry","A Glimpse of Commutative Algebra and Algebraic Geometry","Capstone: Where Algebra Goes Next","Commutative algebra reads geometry off the ring of polynomial functions. The dictionary runs through Noetherian rings and the ascending chain condition, Hilbert's Basis Theorem, affine algebraic sets and the two maps connecting ideals to zero sets, radicals, the Zariski topology, and Hilbert's Nullstellensatz, which over an algebraically closed field makes radical ideals and algebraic sets the same object.\n",{"path":26576,"title":26577,"module":26573,"summary":26578},"\u002Fabstract-algebra\u002Fcapstone\u002Frepresentation-and-character-theory","A Glimpse of Representation and Character Theory","Representation theory studies a group by the ways it can act linearly on a vector space. Representations are equivalent to modules over the group ring; Maschke's theorem gives complete reducibility, the Wedderburn consequences bound the irreducible degrees, and character theory reduces a representation to a trace invariant governed by the orthogonality relations and displayed in the character table of a small group.\n",{"path":26580,"title":26581,"module":6,"summary":6},"\u002Fabstract-algebra","Abstract Algebra",{"path":26583,"title":26584,"module":26585,"summary":26586},"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fatomic-spectra-rutherford","Atomic Spectra and Rutherford's Nucleus","Early Atomic Models and the Old Quantum Theory","Atoms emit light only at sharp, reproducible wavelengths, and by 1890 those wavelengths were captured by the Rydberg-Ritz formula. Neither empirical regularity had a mechanical explanation. Rutherford's alpha-scattering experiment supplied the missing structure: the atom's positive charge and nearly all its mass sit in a tiny central nucleus, with the electrons far outside.\n",{"path":26588,"title":26589,"module":26585,"summary":26590},"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fbohr-model-hydrogen","The Bohr Model of Hydrogen","Bohr grafted three quantum postulates onto Rutherford's nuclear atom: certain orbits do not radiate, radiation accompanies a jump between them, and quantization must match classical physics for large orbits. Quantizing the angular momentum fixes the orbit radii and energies, reproduces the Rydberg-Ritz formula, and predicts the Rydberg constant from fundamental constants alone.\n",{"path":26592,"title":26593,"module":26585,"summary":26594},"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fx-ray-spectra-franck-hertz","X-Ray Spectra and the Franck-Hertz Experiment","Two 1913-14 experiments confirmed the Bohr-Rutherford atom independently of optical spectra. Moseley found that the square root of a characteristic X-ray frequency is linear in atomic number, fixing Z as nuclear charge and ordering the periodic table. Franck and Hertz measured discrete atomic energy levels directly by scattering electrons through a mercury vapor.\n",{"path":26596,"title":26597,"module":26585,"summary":26598},"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fbohr-sommerfeld-old-quantum-theory","The Bohr-Sommerfeld Old Quantum Theory","Bohr fixed the hydrogen levels with a single quantum number by quantizing angular momentum. Sommerfeld replaced that ad hoc rule with a general prescription: quantize the action of each separable coordinate. The rule produces elliptical orbits, a second (azimuthal) quantum number, space quantization, and — once the relativistic mass variation is included — a fine-structure splitting that matches experiment to order alpha squared.\n",{"path":26600,"title":26601,"module":26585,"summary":26602},"\u002Fatomic-physics\u002Fearly-models-and-old-quantum-theory\u002Fold-quantum-theory-limits-wkb","Limits of the Old Quantum Theory and the WKB Bridge","The old quantum theory works only where the classical motion is separable into independent periodic coordinates. It fails for helium, forbids the correct zero angular momentum of the hydrogen ground state, and misses the half-integer in the oscillator and in molecular spectra. The WKB quantization condition, derived from the Schrodinger equation, is the modern descendant of the Sommerfeld rule and repairs the half-integer through the Maslov correction.\n",{"path":26604,"title":26605,"module":26606,"summary":26607},"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fschrodinger-3d-hydrogen","The Schrödinger Equation in Three Dimensions and Hydrogen","The Quantum Hydrogen Atom","Extending the Schrödinger equation to three dimensions and separating it in spherical coordinates produces three ordinary differential equations, one per coordinate. Their boundary conditions generate the quantum numbers n, ℓ, and mℓ, quantize the angular momentum to √(ℓ(ℓ+1))ℏ with projections mℏ, and fix the bound-state energies of hydrogen at −Z²(13.6 eV)\u002Fn².\n",{"path":26609,"title":26610,"module":26606,"summary":26611},"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fhydrogen-wave-functions","Hydrogen Wave Functions and Orbitals","The hydrogen wave functions factor into a radial part Rₙℓ(r) and an angular spherical harmonic Yℓm(θ,φ). Squaring gives the probability cloud; the radial distribution P(r) = r²|ψ|² peaks at the Bohr radius for the ground state and at the Bohr orbits for excited states. The angular part fixes the s, p, and d orbital shapes that govern chemical bonding.\n",{"path":26613,"title":26614,"module":26606,"summary":26615},"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fradial-equation-in-full","Solving the Radial Equation in Full","The hydrogen radial equation is solved from the differential equation up. The substitution u = rR turns it into a one-dimensional problem with a centrifugal barrier; matching the asymptotic behaviour at the origin and at infinity peels off the factors r^(ℓ+1) and e^(−r\u002Fna₀); a Frobenius series for the remainder must terminate, and that termination condition yields the quantization n ≥ ℓ+1 with E = −Z²Ry\u002Fn². The surviving polynomials are the associated Laguerre functions, whose degree n−ℓ−1 counts the radial nodes.\n",{"path":26617,"title":26618,"module":26606,"summary":26619},"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fsymmetry-degeneracy-runge-lenz","Accidental Degeneracy and the Runge-Lenz Symmetry","Hydrogen energies depend only on n, so states of different ℓ at the same n are degenerate. This is not a coincidence but the mark of a hidden symmetry: the quantum Runge-Lenz vector is conserved for the 1\u002Fr potential alone, and together with angular momentum it generates the group SO(4). The Casimir invariant of that group reproduces E = −Z²Ry\u002Fn² and its representations count the n² states. Any departure from 1\u002Fr breaks the symmetry and lifts the ℓ-degeneracy.\n",{"path":26621,"title":26622,"module":26606,"summary":26623},"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fexpectation-values-virial","Expectation Values, the Virial Theorem, and Scaling","The radial matrix elements ⟨r^k⟩ of hydrogenic states are the raw material of every later correction. This lesson derives ⟨1\u002Fr⟩ from the virial theorem, builds the full family ⟨r⟩, ⟨r²⟩, ⟨1\u002Fr²⟩, ⟨1\u002Fr³⟩ from Kramers' recursion and the Feynman-Hellmann theorem, and reads off their scaling with n, ℓ, and Z. The virial balance ⟨T⟩ = −½⟨V⟩ = −E fixes the energy budget of every bound state.\n",{"path":26625,"title":26626,"module":26606,"summary":26627},"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Fquantum-defects-alkali-spectra","Quantum Defects and Alkali Spectra","An alkali atom is one valence electron outside a closed-shell core, and to a good approximation it is hydrogen with a modified quantum number. Core penetration makes low-ℓ states more bound than the Coulomb formula predicts, and the shortfall is captured by a single number per ℓ, the quantum defect δℓ. The spectrum then follows the Rydberg formula with n replaced by the effective n − δℓ, and the sodium D-line doublet is the worked case.\n",{"path":26629,"title":26630,"module":26606,"summary":26631},"\u002Fatomic-physics\u002Fquantum-hydrogen-atom\u002Frydberg-atoms","Rydberg Atoms","A Rydberg atom is an atom excited to a very high principal quantum number, and every hydrogenic property becomes exaggerated by a power of n. Size grows as n², binding falls as n⁻², radiative lifetime lengthens as n³, and the static polarizability explodes as n⁷. The levels crowd toward the ionization limit, and the enormous dipole interaction between two Rydberg atoms produces the blockade that underlies neutral-atom quantum computing.\n",{"path":26633,"title":26634,"module":26635,"summary":26636},"\u002Fatomic-physics\u002Ffine-structure-and-the-dirac-atom\u002Frelativistic-kinetic-correction","The Relativistic Kinetic-Energy Correction","Fine Structure and the Dirac Atom","The Bohr energies treat the electron as slowly moving, but its speed is of order αc, so the kinetic energy needs a relativistic correction. Expanding √(p²c²+m²c⁴) to order (v\u002Fc)² produces the perturbation −p⁴\u002F8m³c², whose first-order shift on a hydrogenic state is evaluated with the trick p²=2m(E−V). The result depends on n and ℓ, is smaller than the gross structure by α²≈5×10⁻⁵, and is one of the three pieces that combine into the fine-structure formula.\n",{"path":26638,"title":26639,"module":26635,"summary":26640},"\u002Fatomic-physics\u002Ffine-structure-and-the-dirac-atom\u002Fspin-orbit-thomas-precession","Spin-Orbit Coupling and Thomas Precession","In the electron's rest frame the nucleus orbits it, and the resulting current produces a magnetic field that couples to the electron's spin moment. The interaction is ξ(r) L·S, with ξ built from the Coulomb potential and the radial expectation ⟨1\u002Fr³⟩. A relativistic subtlety, Thomas precession, halves the naive coefficient because the electron's rest frame is accelerating. The result splits each ℓ≥1 level into a j=ℓ±½ doublet and makes (n, ℓ, j, mⱼ) the good quantum numbers.\n",{"path":26642,"title":26643,"module":26635,"summary":26644},"\u002Fatomic-physics\u002Ffine-structure-and-the-dirac-atom\u002Fdarwin-term-fine-structure-formula","The Darwin Term and the Fine-Structure Formula","The third fine-structure correction, the Darwin term, is a contact interaction proportional to ∇²V that acts only on s-states, physically a smearing of the electron over a Compton wavelength. Adding the relativistic, spin-orbit, and Darwin shifts, the ℓ-dependence cancels and the total collapses to a formula in n and j alone. The n=2 shell splits into 2S₁\u002F₂, 2P₁\u002F₂, 2P₃\u002F₂, with the two j=½ levels exactly degenerate, a coincidence the Dirac theory explains.\n",{"path":26646,"title":26647,"module":26635,"summary":26648},"\u002Fatomic-physics\u002Ffine-structure-and-the-dirac-atom\u002Fdirac-equation-hydrogen","The Dirac Equation for Hydrogen","The fine-structure formula was assembled from three perturbations; the Dirac equation produces it in one stroke and exactly. A first-order relativistic wave equation forces a four-component spinor, from which spin s=½, the g-factor of 2, the spin-orbit term, and antiparticles all emerge automatically. Its exact Coulomb spectrum depends only on n and j, and expanding in Zα reproduces the perturbative result, including the 2S₁\u002F₂–2P₁\u002F₂ degeneracy that sets up the Lamb shift.\n",{"path":26650,"title":26651,"module":26652,"summary":26653},"\u002Fatomic-physics\u002Fqed-corrections-and-hyperfine-structure\u002Flamb-shift-qed","The Lamb Shift and QED Radiative Corrections","QED Corrections and Hyperfine Structure","The Dirac equation makes the 2S₁\u002F₂ and 2P₁\u002F₂ levels of hydrogen exactly degenerate. Lamb and Retherford measured a splitting of about 1058 MHz that the Dirac theory cannot produce. The gap comes from the electron's coupling to the quantized electromagnetic field: self-energy, vacuum polarization, and the anomalous magnetic moment. Welton's vacuum-fluctuation estimate reproduces the size and shows why the effect lands almost entirely on s-states, and the same radiative corrections make hydrogen the most stringent test of QED.\n",{"path":26655,"title":26656,"module":26652,"summary":26657},"\u002Fatomic-physics\u002Fqed-corrections-and-hyperfine-structure\u002Fhyperfine-structure-21cm","Hyperfine Structure and the 21 cm Line","The proton carries a magnetic moment, and it interacts with the magnetic field the electron produces at the nucleus. For s-states that interaction is the Fermi contact term, proportional to the electron density at the origin and to the dot product of the nuclear and electronic spins. Coupling I and J into F = I + J splits each level by a Landé interval rule; in hydrogen's ground state it produces the F = 0\u002FF = 1 doublet whose 1420 MHz, 21 cm transition maps neutral hydrogen across the galaxy.\n",{"path":26659,"title":26660,"module":26652,"summary":26661},"\u002Fatomic-physics\u002Fqed-corrections-and-hyperfine-structure\u002Fnuclear-effects-isotope-shift","Nuclear Size, Moments, and Isotope Shifts","A real nucleus has a finite size, a mass that changes between isotopes, and, when its spin is at least one, an electric quadrupole moment. Each leaves a fingerprint in the atomic spectrum: the volume shift from s-electrons sampling the charge distribution, the mass and field isotope shifts that separate on a King plot, the quadrupole interaction that breaks the Landé interval rule, and the hyperfine anomaly from the magnetization distribution. Atomic spectroscopy reads nuclear properties out of these shifts.\n",{"path":26663,"title":26664,"module":26665,"summary":26666},"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fperiodic-table-atomic-spectra","The Periodic Table and Atomic Spectra","Many-Electron Atoms","Identical electrons demand antisymmetric wave functions, which is the Pauli exclusion principle: no two electrons share all four quantum numbers. Filling shells in order of increasing energy — shifted by penetration and shielding — builds the periodic table and its recurring ionization pattern. Selection rules govern optical spectra, and an external field splits lines by the Zeeman effect.\n",{"path":26668,"title":26669,"module":26665,"summary":26670},"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fcentral-field-self-consistent","The Central-Field Approximation and the Self-Consistent Field","The N-electron Hamiltonian does not separate because every pair of electrons repels. The central-field approximation replaces that pairwise repulsion with an averaged spherical potential each electron feels, restoring hydrogen-like orbitals labelled by n and ℓ. The Thomas-Fermi statistical model fixes the shape of the screened charge from Fermi-gas thermodynamics; the Hartree self-consistent field determines it exactly by iterating orbitals against the potential they generate until the two agree.\n",{"path":26672,"title":26673,"module":26665,"summary":26674},"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fidentical-particles-hartree-fock","Exchange, Slater Determinants, and Hartree-Fock","A product wave function ignores that electrons are identical fermions. Enforcing antisymmetry writes the state as a Slater determinant, which vanishes whenever two electrons share a spin-orbital — the exclusion principle made algebraic. The energy of a determinant carries a new term with no classical analogue, the exchange integral, nonzero only for parallel spins; it lowers the energy of aligned electrons and carves a Fermi hole around each one. Adding the exchange operator to the mean field gives the Hartree-Fock equations, and what they still miss defines the correlation energy.\n",{"path":26676,"title":26677,"module":26665,"summary":26678},"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fhelium-two-electron-atom","Helium: the Prototype Two-Electron Atom","Helium is the smallest atom the Schrödinger equation cannot solve exactly, and the smallest that shows every many-electron effect. Ignoring the electron repulsion overbinds the ground state by 30 eV; first-order perturbation theory and a one-parameter variational calculation with an effective charge close most of the gap. The excited configurations split into para (singlet) and ortho (triplet) states separated by the exchange integral, with the triplet lower — and the absence of a 1s² triplet is the Pauli principle in its plainest form.\n",{"path":26680,"title":26681,"module":26665,"summary":26682},"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fls-jj-coupling-term-symbols","LS and jj Coupling; Term Symbols","A configuration is not a single energy level. The residual electrostatic repulsion and the spin-orbit interaction split it, and which one dominates fixes the coupling scheme. In light atoms the electrostatic term wins: orbital and spin angular momenta couple separately into L and S, then into J, giving Russell- Saunders term symbols. In heavy atoms spin-orbit wins and each electron's j forms first. The Pauli principle prunes the allowed terms of equivalent electrons, the Landé interval rule spaces the fine-structure multiplet, and the scheme crosses over from LS to jj down a column.\n",{"path":26684,"title":26685,"module":26665,"summary":26686},"\u002Fatomic-physics\u002Fmany-electron-atoms\u002Fhund-rules-ground-terms","Hund's Rules and Ground-State Terms","A configuration allows several terms; Hund's three rules pick the ground one. Maximize the spin S first, then the orbital L, then set J to |L−S| for a less-than-half shell and L+S for a more-than-half shell. The first two rules come from exchange lowering the energy of apart-kept electrons; the third comes from the sign of the spin-orbit coupling, which flips as a shell passes half-filling and turns the multiplet from normal to inverted. Worked ground terms for carbon, nitrogen, oxygen, and iron show the rules in action.\n",{"path":26688,"title":26689,"module":26690,"summary":26691},"\u002Fatomic-physics\u002Fatoms-in-external-fields\u002Fzeeman-effect","The Zeeman Effect","Atoms in External Fields","A magnetic field couples to the atom through its magnetic moment, splitting each level into equally spaced sublevels labelled by the projection of the total angular momentum. When spin is present the spacing is not the classical one: it carries the Landé g-factor, a projection of the spin and orbital moments onto the total angular momentum. We derive the weak-field Hamiltonian from minimal coupling, evaluate the shift with the projection theorem, and read off the polarization of the emitted components.\n",{"path":26693,"title":26694,"module":26690,"summary":26695},"\u002Fatomic-physics\u002Fatoms-in-external-fields\u002Fpaschen-back-intermediate","The Paschen-Back and Intermediate-Field Regimes","When the magnetic interaction grows past the fine-structure coupling, spin and orbital angular momentum decouple and precess independently about the field. The anomalous Zeeman pattern reverts to a simple triplet, the Paschen-Back effect. Between the two limits neither coupling dominates and the level positions follow from diagonalizing the combined spin-orbit and Zeeman Hamiltonian. We build the two-by-two problem for a single valence electron, solve it in closed form, and show both limits emerge from one expression.\n",{"path":26697,"title":26698,"module":26690,"summary":26699},"\u002Fatomic-physics\u002Fatoms-in-external-fields\u002Fstark-effect-polarizability","The Stark Effect and Field Ionization","An electric field shifts atomic levels by coupling to the electron's position. Parity forbids a first-order shift for a non-degenerate state, so most atoms respond only at second order through their polarizability, a quadratic Stark shift. Hydrogen is the exception: its accidental degeneracy admits a permanent dipole and a linear shift, cleanest in parabolic coordinates. At large fields the Coulomb well develops a saddle, and Rydberg states field-ionize at a threshold that falls as the fourth power of the principal quantum number.\n",{"path":26701,"title":26702,"module":26703,"summary":26704},"\u002Fatomic-physics\u002Fradiative-transitions-and-line-shapes\u002Ftime-dependent-perturbation-golden-rule","Time-Dependent Perturbation Theory and the Golden Rule","Radiative Transitions and Spectral Lines","An atom in a weak oscillating field makes transitions between its stationary states. First-order time-dependent perturbation theory gives the transition amplitude as a Fourier component of the perturbation at the Bohr frequency, and the resulting probability is a sinc-squared resonance that sharpens as the field acts longer. For a two-level system the same coupling produces Rabi oscillations; for a transition into a continuum the long-time limit collapses the sinc-squared into a delta function and yields Fermi's golden rule, a constant transition rate set by the coupling strength and the density of final states.\n",{"path":26706,"title":26707,"module":26703,"summary":26708},"\u002Fatomic-physics\u002Fradiative-transitions-and-line-shapes\u002Fdipole-approximation-einstein-coefficients","The Dipole Approximation and Einstein Coefficients","The coupling between an atom and light is the interaction of the electron with the electromagnetic field. Because an optical wavelength dwarfs the atom, the spatial variation of the field across the atom can be dropped, leaving the electric-dipole interaction and its matrix element. That matrix element defines the oscillator strength, which obeys the Thomas-Reiche-Kuhn sum rule. Einstein's three rate coefficients (absorption, stimulated emission, spontaneous emission) follow from detailed balance with thermal radiation, fixing the ratio of spontaneous to stimulated rates and its steep growth with frequency.\n",{"path":26710,"title":26711,"module":26703,"summary":26712},"\u002Fatomic-physics\u002Fradiative-transitions-and-line-shapes\u002Fselection-rules-forbidden-transitions","Selection Rules and Forbidden Transitions","The dipole matrix element vanishes for most pairs of states, and the pattern of which survive is the set of selection rules. Parity forces the orbital angular momentum to change by one; the angular integral of three spherical harmonics restricts the magnetic quantum number to change by zero or one; the photon's spin restricts the total angular momentum. When the dipole element vanishes, higher multipoles (magnetic dipole and electric quadrupole) can still drive the transition at rates smaller by powers of the fine-structure constant, and states with no allowed decay become metastable.\n",{"path":26714,"title":26715,"module":26703,"summary":26716},"\u002Fatomic-physics\u002Fradiative-transitions-and-line-shapes\u002Flifetimes-and-line-shapes","Lifetimes, Line Widths, and Line Shapes","A spectral line is never infinitely sharp. The finite lifetime of the excited state gives every line a natural Lorentzian width set by the total decay rate, the Fourier transform of an exponentially damped emission. Thermal motion adds a Gaussian Doppler width that usually dominates in a gas; collisions add a further Lorentzian pressure width; the observed profile is the Voigt convolution of the Gaussian and Lorentzian parts. Strong driving fields broaden the line further through saturation. Each mechanism has a distinct dependence on temperature, density, and intensity that lets it be identified and, where possible, removed.\n",{"path":26718,"title":26719,"module":26720,"summary":26721},"\u002Fatomic-physics\u002Flasers-and-spectroscopy\u002Flaser-principles","Population Inversion, Gain, and the Laser","Lasers and Spectroscopy","A laser is an optical amplifier placed inside a resonant cavity. Amplification requires that stimulated emission outrun absorption, which requires more atoms in the upper level than the lower one — a population inversion that the Einstein relations forbid in thermal equilibrium and that no two-level pump can produce. Three- and four-level schemes reach it by routing atoms through auxiliary states. The gain coefficient sets how strongly a weak beam grows, the cavity fixes the threshold and selects a comb of longitudinal modes, and gain saturation clamps the steady-state inversion at its threshold value.\n",{"path":26723,"title":26724,"module":26720,"summary":26725},"\u002Fatomic-physics\u002Flasers-and-spectroscopy\u002Fspectroscopy-techniques","Spectroscopic Techniques and Frequency Combs","A tunable laser turns spectroscopy from photographing a spectrum into interrogating a single transition, but at room temperature the Doppler width buries the natural linewidth under a thousandfold-broader Gaussian. Saturated absorption and two-photon spectroscopy defeat the first-order Doppler shift by selecting the zero-velocity class or cancelling the shift between counter-propagating photons, recovering natural-width features. Laser-induced fluorescence pushes sensitivity to single atoms, and the optical frequency comb converts an optical frequency into a countable radio-frequency beat, giving absolute frequency measurement across the visible spectrum.\n",{"path":26727,"title":26728,"module":26720,"summary":26729},"\u002Fatomic-physics\u002Flasers-and-spectroscopy\u002Fline-catalog-nist-asd","Reading Real Spectra with the NIST Database","Every quantity computed in this course — energy levels, transition frequencies, oscillator strengths, lifetimes — is tabulated for real atoms in the NIST Atomic Spectra Database. This lesson reads that data as physics: how levels are labelled by term symbols and energies in wavenumbers, how a transition list encodes wavelength, Einstein coefficient, and line strength, how a Grotrian diagram is reconstructed from the tables, and how a measured spectrum is matched to catalog lines. The residual between computed and tabulated positions is the running score of atomic theory.\n",{"path":26731,"title":26732,"module":26733,"summary":26734},"\u002Fatomic-physics\u002Fmodern-atomic-physics\u002Flaser-cooling-doppler","Laser Cooling and Optical Molasses","Modern Atomic Physics","A near-resonant laser beam pushes an atom because every absorbed photon delivers one unit of momentum and the subsequent spontaneous emission averages to zero. Two counter-propagating red-detuned beams turn that push into friction: the Doppler shift brings a moving atom closer to resonance with the beam it moves against, so the net force opposes the velocity. Six beams give optical molasses in three dimensions. The random recoil of spontaneous emission heats against the friction, and the balance sets the Doppler cooling limit. Adding a magnetic-field gradient makes the force position-dependent as well, giving the magneto-optical trap.\n",{"path":26736,"title":26737,"module":26733,"summary":26738},"\u002Fatomic-physics\u002Fmodern-atomic-physics\u002Fsub-doppler-trapping","Sub-Doppler Cooling and Atom Traps","Optical molasses cools multilevel atoms below the Doppler limit. A polarization gradient plus optical pumping makes an atom repeatedly climb a light-shift hill and be pumped to the valley, losing kinetic energy each cycle — Sisyphus cooling. The floor is the recoil limit, one photon momentum of residual motion. Below it, cooling must avoid scattering photons: conservative magnetic and optical-dipole traps hold the atoms while forced evaporation removes the hot tail, driving the phase-space density up toward quantum degeneracy.\n",{"path":26740,"title":26741,"module":26733,"summary":26742},"\u002Fatomic-physics\u002Fmodern-atomic-physics\u002Fbose-einstein-condensation","Bose-Einstein Condensation of Atomic Gases","Below a critical temperature a gas of identical bosons places a macroscopic fraction of its atoms in the single lowest-energy state. The transition occurs when the thermal de Broglie wavelength grows to the interparticle spacing, so the phase- space density reaches order unity. The critical temperature follows from the Bose-Einstein distribution and the density of states, the condensate fraction grows as one minus (T\u002FTc) to the three-halves, and the condensate reveals itself in time-of-flight as a sharp bimodal peak in the momentum distribution. The 1995 rubidium and sodium experiments realized it in dilute trapped gases.\n",{"path":26744,"title":26745,"module":26733,"summary":26746},"\u002Fatomic-physics\u002Fmodern-atomic-physics\u002Foptical-clocks-precision","Optical Atomic Clocks and Precision Measurement","An atomic clock counts the oscillations of a field locked to an atomic transition. The cesium microwave standard defines the second through the 9.19 GHz ground-state hyperfine transition, interrogated by Ramsey's separated-oscillatory-field method whose fringe width is set by the free-precession time. Optical clocks replace the microwave transition with an optical one five orders of magnitude higher in frequency, raising the quality factor and the fractional stability in proportion. Lattice and single-ion clocks reach fractional uncertainties near ten-to-the-minus- eighteen by trapping the atoms at a magic wavelength that cancels the light shift, and at that level they measure the gravitational redshift over centimetres of height.\n",{"path":26748,"title":26749,"module":6,"summary":6},"\u002Fatomic-physics","Atomic Physics",{"path":26751,"title":26752,"module":6,"summary":6},"\u002Fdatabases","Databases",{"path":26754,"title":26755,"module":24340,"summary":26756},"\u002Fcategory-theory\u002Ffoundations\u002Fwhat-is-a-category","Categories, Objects, and Arrows","A category is objects, arrows between them, a rule for composing arrows, and an identity arrow on every object, subject to associativity and the unit laws. The axioms mention no elements: arrows need not be functions, and an object is known only through the arrows into and out of it. Isomorphism, commutative diagrams, duality, and the terminal object are the first consequences.\n",{"path":26758,"title":26759,"module":24340,"summary":26760},"\u002Fcategory-theory\u002Ffoundations\u002Fexamples-of-categories","A Zoo of Categories","The axioms admit two very different kinds of model: large categories of structured sets and their structure-preserving maps (Set, Mon, Grp, Top, Vect), and small categories that are themselves single algebraic objects — a monoid as a one-object category, a poset as a thin category. The awkward cases Rel and Pfn have sets as objects but relations and partial functions as arrows, and a typed programming language presents its types and programs as a category.\n",{"path":26762,"title":26763,"module":24340,"summary":26764},"\u002Fcategory-theory\u002Ffoundations\u002Fspecial-morphisms","Isomorphisms, Monos, and Epis","Injectivity and surjectivity mention elements, so a general category re-expresses them by cancellation: monomorphisms cancel on the left, epimorphisms on the right. Sections and retractions are the split versions with an explicit one-sided inverse. Mono plus epi does not force an isomorphism, and subobjects are equivalence classes of monos into a fixed object.\n",{"path":26766,"title":26767,"module":24340,"summary":26768},"\u002Fcategory-theory\u002Ffoundations\u002Ffunctors","Functors: Maps Between Categories","A functor sends objects to objects and arrows to arrows while preserving composition and identities. Covariant and contravariant functors, the standard stock (forgetful, free, hom, and powerset), and the classification by faithfulness, fullness, and essential surjectivity all follow. Functors compose, so categories and functors form a category themselves.\n",{"path":26770,"title":26771,"module":24340,"summary":26772},"\u002Fcategory-theory\u002Ffoundations\u002Fnatural-transformations","Natural Transformations and Functor Categories","A natural transformation is a map between two parallel functors: one component arrow per object, subject to a commuting square for every arrow of the source. Naturality is verified for the determinant, the double dual, and list operations; functors and natural transformations form the functor category [C, D]; and vertical and horizontal composition satisfy the Godement interchange law.\n",{"path":26774,"title":26775,"module":24340,"summary":26776},"\u002Fcategory-theory\u002Ffoundations\u002Fsize-and-set-theory","Size: Small, Large, Locally Small","The objects of Set do not form a set, and pretending otherwise reproduces the classical paradoxes. Classes make the small\u002Flarge distinction precise, with locally small and essentially small as the intermediate notions. Cantor's theorem shows Set and its algebraic relatives are large, and the function-based axiomatization of sets is the one category theory prefers to ZFC.\n",{"path":26778,"title":26779,"module":26780,"summary":26781},"\u002Fcategory-theory\u002Funiversal-properties\u002Funiversal-properties","Universal Properties, Initial and Terminal Objects","Universal Properties and Basic Constructions","A universal property characterizes an object by a for-all\u002Fexists-unique condition on the arrows into or out of it, and any two objects satisfying the same property are isomorphic by a unique isomorphism. Initial and terminal objects are the simplest cases; the free vector space, the discrete topology, and the ring of integers show the pattern at work.\n",{"path":26783,"title":26784,"module":26780,"summary":26785},"\u002Fcategory-theory\u002Funiversal-properties\u002Fproducts-and-coproducts","Products and Coproducts","The product of two objects is a wedge of projections through which every other wedge factors uniquely; the coproduct is the dual, built from injections. In Set these are the cartesian product and the disjoint union, in a poset the meet and join, and in abelian groups the two coincide. The mediating-arrow discipline established here is the template for all limits.\n",{"path":26787,"title":26788,"module":26780,"summary":26789},"\u002Fcategory-theory\u002Funiversal-properties\u002Fconstructions-on-categories","Opposite, Product, Slice, and Comma Categories","Categories are themselves mathematical structures, and the standard algebraic constructions apply: opposites, products, subcategories, slices, and the comma category that subsumes them. The opposite category yields the duality principle, halving the subject's proofs; slice and comma categories repackage every universal property as an initial or terminal object.\n",{"path":26791,"title":26792,"module":26793,"summary":26794},"\u002Fcategory-theory\u002Frepresentables-yoneda\u002Frepresentable-functors","Hom-Functors and Representables","Representables and the Yoneda Lemma","Fixing an object A of a locally small category produces a set-valued functor, the hom-functor A(A,-), that records every map out of A. A functor is representable when it is naturally isomorphic to such a hom-functor. We define the covariant and contravariant hom-functors, collect the standard representables (identity, forgetful, powerset), and read maps as generalized elements of varying shape.\n",{"path":26796,"title":26797,"module":26793,"summary":26798},"\u002Fcategory-theory\u002Frepresentables-yoneda\u002Fyoneda-lemma","The Yoneda Lemma","The Yoneda lemma computes the natural transformations out of a representable presheaf: they form a set in natural bijection with X(A). The proof fixes a single degree of freedom, the image of the identity arrow, and shows naturality forces everything else. We prove the bijection, verify naturality in both variables, and read off that a natural transformation out of a representable is just one element.\n",{"path":26800,"title":26801,"module":26793,"summary":26802},"\u002Fcategory-theory\u002Frepresentables-yoneda\u002Fyoneda-consequences","The Yoneda Embedding and Its Uses","Three corollaries turn the Yoneda lemma into working machinery. A representation of a presheaf is the same thing as a universal element; the Yoneda embedding of a category into its presheaf category is full and faithful; and two objects are isomorphic exactly when their representables are. Together they justify constructing arrows by constructing natural transformations between hom-functors, and they contain Cayley's theorem as the one-object case.\n",{"path":26804,"title":26805,"module":26806,"summary":26807},"\u002Fcategory-theory\u002Flimits-colimits\u002Flimits","Cones and Limits","Limits and Colimits","A diagram is a functor from a small shape category; a cone over it is an object with compatible legs to every node; and a limit is the terminal cone, the one every other cone factors through uniquely. Products and terminal objects reappear as limits over particular shapes, and the whole construction is unique up to a single isomorphism.\n",{"path":26809,"title":26810,"module":26806,"summary":26811},"\u002Fcategory-theory\u002Flimits-colimits\u002Fproducts-equalizers-pullbacks","Equalizers and Pullbacks","The equalizer of a parallel pair is the universal arrow that makes the two composites agree; the pullback of a cospan is the universal commutative square. In Set they are solution sets and fibered products, every equalizer is monic, monics are stable under pullback, and products plus equalizers together generate all limits.\n",{"path":26813,"title":26814,"module":26806,"summary":26815},"\u002Fcategory-theory\u002Flimits-colimits\u002Fcolimits","Colimits: Coproducts, Coequalizers, Pushouts","Colimits are limits in the opposite category: cocones replace cones, and the universal cocone is initial rather than terminal. Coproducts glue objects side by side, coequalizers impose relations and produce quotients, pushouts glue along a shared part, and in Set every colimit is a quotient of a disjoint union. Directed colimits admit a clean elementwise description.\n",{"path":26817,"title":26818,"module":26806,"summary":26819},"\u002Fcategory-theory\u002Flimits-colimits\u002Fcomputing-limits","Computing Limits in Concrete Categories","In Set the limit of any diagram is the set of threads: choice functions through the nodes that commute with every edge. In Pos, Mon, and Top the recipe is the same limit downstairs plus the unique structure that makes the projections structure-preserving — pointwise order, componentwise operations, the topology generated by the projections. The pattern is what \"the forgetful functor creates limits\" means concretely.\n",{"path":26821,"title":26822,"module":26806,"summary":26823},"\u002Fcategory-theory\u002Flimits-colimits\u002Flimits-and-functors","Preservation, Reflection, and Creation of Limits","A functor preserves limits if it sends limit cones to limit cones, reflects them if it recognizes them, and creates them if limits downstairs lift uniquely upstairs. Representable functors preserve all limits, forgetful functors from algebra create them, and limits in functor categories are computed pointwise, one evaluation at a time.\n",{"path":26825,"title":26826,"module":26827,"summary":26828},"\u002Fcategory-theory\u002Fadjunctions\u002Fadjunctions","Adjoint Functors via Hom-Set Bijections","Adjunctions","An adjunction is a natural bijection between two hom-sets: maps out of $F(A)$ in one category correspond to maps into $G(B)$ in the other. We give the definition, spell out the naturality axioms that make the correspondence compatible with composition, and work the flagship examples — free vector spaces, free groups, discrete and indiscrete topologies, and currying.\n",{"path":26830,"title":26831,"module":26827,"summary":26832},"\u002Fcategory-theory\u002Fadjunctions\u002Funits-and-counits","Units, Counits, and the Triangle Identities","The whole hom-set bijection of an adjunction is generated by two natural transformations: the unit, obtained by transposing identity maps on one side, and the counit, by transposing them on the other. Two triangle identities are all they must satisfy, and any pair satisfying them determines a unique adjunction. The same correspondence specializes to order-preserving maps between posets and to free constructions.\n",{"path":26834,"title":26835,"module":26827,"summary":26836},"\u002Fcategory-theory\u002Fadjunctions\u002Fadjunctions-via-universal-arrows","Adjunctions from Universal Arrows","The unit component at a single object is an initial object of a comma category, and this universal property alone rebuilds the whole adjunction. A functor has a left adjoint exactly when every object admits such a universal arrow, and the left adjoint is assembled from them one object at a time. We prove the equivalence of all three formulations of adjointness.\n",{"path":26838,"title":26839,"module":26827,"summary":26840},"\u002Fcategory-theory\u002Fadjunctions\u002Ffree-forgetful-adjunctions","Free Constructions and Free–Forgetful Adjunctions","Free monoids, free groups, and free vector spaces are left adjoints to forgetful functors, and the universal mapping property is all one needs to prove it. Some forgetful functors also have right adjoints (co-free constructions like the indiscrete topology), producing three-functor chains. Contravariant adjunctions, symmetric in their two functors, close the lesson with the pattern behind duality and representation theorems.\n",{"path":26842,"title":26843,"module":26844,"summary":26845},"\u002Fcategory-theory\u002Fadjoints-limits\u002Flimits-via-adjoints","Limits as Adjoints and as Representables","Adjoints, Representables, and Limits Together","A cone on a diagram is a natural transformation from a constant diagram, so a limit is a representation of the cone functor and, equivalently, a value of the right adjoint to the diagonal functor. We prove both rephrasings, derive uniqueness and functoriality of limits from them, and record the dual statement that a colimit is the left adjoint to the diagonal.\n",{"path":26847,"title":26848,"module":26844,"summary":26849},"\u002Fcategory-theory\u002Fadjoints-limits\u002Fpresheaf-limits-colimits","Limits and Colimits of Presheaves","Representables preserve limits, and limits in a functor category are computed one object at a time, so a presheaf category is complete and cocomplete with all its structure inherited pointwise from Set. The Yoneda embedding then preserves limits but not colimits, and the density theorem repairs the colimit side: every presheaf is a canonical colimit of representables.\n",{"path":26851,"title":26852,"module":26844,"summary":26853},"\u002Fcategory-theory\u002Fadjoints-limits\u002Fadjoints-preserve-limits","Right Adjoints Preserve Limits (RAPL)","A functor with a left adjoint preserves every limit that exists, and dually a functor with a right adjoint preserves colimits. The proof is a four-line chain of natural isomorphisms through the adjunction and the continuity of representables. The theorem yields product-and-exponential arithmetic in Set, another proof that limits commute with limits, and a standard test for proving that a functor has no adjoint.\n",{"path":26855,"title":26856,"module":26844,"summary":26857},"\u002Fcategory-theory\u002Fadjoints-limits\u002Fadjoint-functor-theorem","The Adjoint Functor Theorem","RAPL makes limit preservation necessary for having a left adjoint; the adjoint functor theorems identify when it is sufficient. For ordered sets no extra hypothesis is needed. In general the candidate adjoint is a limit over a comma category that may be large, and the general adjoint functor theorem tames it with a weakly initial set. We prove GAFT in full and apply it to free groups and, through the special adjoint functor theorem, the Stone–Čech compactification.\n",{"path":26859,"title":26860,"module":26861,"summary":26862},"\u002Fcategory-theory\u002Fmonads-algebras\u002Fmonads","Monads from Adjunctions","Monads and Algebras","A monad on a category is an endofunctor equipped with a unit and a multiplication satisfying associativity and unit laws — the data of a monoid, written internally to the category of endofunctors. Every adjunction induces one, and the list, exception, and state constructions that model computational effects are all monads on Set.\n",{"path":26864,"title":26865,"module":26861,"summary":26866},"\u002Fcategory-theory\u002Fmonads-algebras\u002Falgebras-eilenberg-moore","Algebras for a Monad","An algebra for a monad is an object with a structure map that interacts correctly with the unit and multiplication. The algebras form the Eilenberg–Moore category, whose free–forgetful adjunction induces the monad back; a comparison functor relates any other inducing adjunction to it, and for the list monad the algebras are exactly monoids.\n",{"path":26868,"title":26869,"module":26861,"summary":26870},"\u002Fcategory-theory\u002Fmonads-algebras\u002Fkleisli-and-programming","The Kleisli Category and Monads in Programming","The Kleisli category of a monad has the same objects as the base but takes arrows A to TB, composed by mapping and flattening. These arrows are effectful programs, Kleisli composition is the bind of functional programming, and the Kleisli adjunction is the initial resolution of the monad, with Eilenberg–Moore at the terminal end.\n",{"path":26872,"title":26873,"module":26861,"summary":26874},"\u002Fcategory-theory\u002Fmonads-algebras\u002Falgebras-for-endofunctors","Algebras for an Endofunctor and Recursion","Dropping the monad laws leaves algebras for a bare endofunctor, whose initial objects are the least fixed points of the functor by Lambek's lemma. The natural numbers, lists, and trees are initial algebras; the unique map out of an initial algebra is the fold of functional programming; and the Smyth–Plotkin fixed-point technique builds Scott domains the same way.\n",{"path":26876,"title":26877,"module":26878,"summary":26879},"\u002Fcategory-theory\u002Fcartesian-closed-lambda\u002Fcartesian-closed-categories","Cartesian Closed Categories","Cartesian Closed Categories and Typed Lambda Calculus","A cartesian closed category has a terminal object, binary products, and for every pair of objects an exponential object that internalizes the hom-set as an object of the category. The defining data is an evaluation arrow and a currying operation, packaged by the adjunction between product-with-A and exponential-by-A. Set, Boolean and Heyting algebras, functor categories, and Cat are all cartesian closed.\n",{"path":26881,"title":26882,"module":26878,"summary":26883},"\u002Fcategory-theory\u002Fcartesian-closed-lambda\u002Flambda-calculus-correspondence","Typed Lambda Calculus and CCCs","The typed lambda calculus and the cartesian closed category are two presentations of the same theory. Types become objects, terms with one free variable become arrows, product types become products, and function types become exponentials, with abstraction matching currying and application matching evaluation. Building the category of a lambda theory and the internal language of a category are mutually inverse up to equivalence.\n",{"path":26885,"title":26886,"module":26878,"summary":26887},"\u002Fcategory-theory\u002Fcartesian-closed-lambda\u002Ffixed-points-and-recursion","Fixed Points in Cartesian Closed Categories","The untyped lambda calculus has a fixed-point combinator; the typed calculus cannot, and Lawvere's fixed-point theorem explains why: any point-surjection onto an exponential forces every endomap to have a fixed point, which is the abstract form of Cantor's diagonal argument. Recursion is recovered instead by restricting to omega-complete partially ordered objects, where every continuous endomap has a least fixed point built by iterating from bottom. This gives While loops a semantics.\n",{"path":26889,"title":26890,"module":6,"summary":6},"\u002Fcategory-theory","Category Theory",{"path":26892,"title":25534,"module":26893,"summary":26894},"\u002Fdeep-learning\u002Fmathematical-background\u002Flinear-algebra-for-deep-learning","Mathematical Background","Every quantity a network touches is a tensor, and every layer is a matrix acting on one. This lesson compiles the linear algebra deep learning actually uses: products and norms, the system $Ax=b$ and when it is solvable, the two decompositions (eigen and SVD) that diagonalize a transformation, and the pseudoinverse that solves what cannot be solved exactly. It then derives PCA as the worked example that ties it all together.\n",{"path":26896,"title":26897,"module":26893,"summary":26898},"\u002Fdeep-learning\u002Fmathematical-background\u002Fprobability-and-information-theory","Probability & Information Theory","This lesson assembles the probabilistic vocabulary a network is trained in (random variables, densities, the chain rule, expectation and covariance, the handful of distributions that recur everywhere) and then the information theory that turns a probabilistic model into a loss: self-information, entropy, and the KL divergence whose asymmetry is the cross-entropy objective itself.\n",{"path":26900,"title":26901,"module":26893,"summary":26902},"\u002Fdeep-learning\u002Fmathematical-background\u002Fnumerical-computation","Numerical Computation","Machine learning runs on finite-precision arithmetic, where every number is approximated and every operation rounds. This lesson sets the numerical ground rules: overflow and underflow and the standard stabilizations, the condition number that measures how much a problem amplifies error, and the gradient-based optimization (first and second order, constrained and unconstrained) that every training loop runs.\n",{"path":26904,"title":24893,"module":26893,"summary":26905},"\u002Fdeep-learning\u002Fmathematical-background\u002Fcalculus","This lesson assembles the differential calculus used in training networks: the gradient and directional derivative, the Jacobian and Hessian, and the chain rule in scalar, vector, and matrix form. From the chain rule it derives back-propagation as a single sweep over the computational graph, tabulates the matrix-calculus identities that recur in layer gradients, reads optimization off a second-order Taylor expansion, and ends with why reverse-mode automatic differentiation is the algorithm every framework runs.\n",{"path":26907,"title":26908,"module":24340,"summary":26909},"\u002Fdeep-learning\u002Ffoundations\u002Fwhat-is-deep-learning","What Is Deep Learning?","Deep learning is representation learning by composition: stack simple differentiable layers, define a loss, and let gradient descent discover the features a human would otherwise have to engineer by hand. We set up the whole vocabulary (model, loss, optimizer, data), the training loop that ties them together, and the three reasons the approach became practical.\n",{"path":26911,"title":26912,"module":24340,"summary":26913},"\u002Fdeep-learning\u002Ffoundations\u002Fmachine-learning-refresher","A Machine-Learning Refresher","The statistical framework the networks live in: data drawn from an unknown distribution, a loss to minimize, and the central question of generalization: will it work on data we have not seen? We set up empirical risk, capacity, the bias–variance tradeoff, and maximum likelihood.\n",{"path":26915,"title":26916,"module":24340,"summary":26917},"\u002Fdeep-learning\u002Ffoundations\u002Flinear-models-and-the-perceptron","Linear Models & the Perceptron","The simplest learners (linear regression, logistic regression, the perceptron) already contain the whole template: a weighted sum, a loss, a gradient step. They also fail on the XOR problem, which no linear model can solve — the limitation that motivates deep learning.\n",{"path":26919,"title":26920,"module":26921,"summary":26922},"\u002Fdeep-learning\u002Fneural-networks\u002Fthe-multilayer-perceptron","The Multilayer Perceptron","Neural Networks","Stacking linear layers with a nonlinearity between them removes the limitation that stopped the perceptron. We build the multilayer perceptron in explicit matrix form (the forward pass, its dimensions, a worked XOR network with concrete weights) and prove why the nonlinearity is essential: without it the deepest stack collapses to a single hyperplane.\n",{"path":26924,"title":26925,"module":26921,"summary":26926},"\u002Fdeep-learning\u002Fneural-networks\u002Factivation-functions","Activation Functions","The activation is the only nonlinear part of a layer, and the reason depth adds expressive power. We catalog the standard hidden units (sigmoid, tanh, ReLU and its descendants, plus GELU, softplus, swish and maxout), derive each unit's derivative in full, make the vanishing-gradient problem quantitative with the chain-rule product, work numeric examples, and explain why the saturating units gave way to ReLU and why ReLU's own dead-unit failure gave way to Leaky\u002FPReLU\u002FELU\u002FGELU.\n",{"path":26928,"title":26929,"module":26921,"summary":26930},"\u002Fdeep-learning\u002Fneural-networks\u002Funiversal-approximation","Universal Approximation","One hidden layer with a non-polynomial activation can approximate any continuous function on a compact set to arbitrary accuracy: the universal approximation theorem. We prove it constructively (two sigmoids make a bump; sums of bumps make any curve), then show the limitation: existence is not efficiency. Depth-separation results exhibit functions a deep net represents with $O(n)$ units that a shallow net needs $\\exp(n)$ units to match.\n",{"path":26932,"title":26933,"module":26921,"summary":26934},"\u002Fdeep-learning\u002Fneural-networks\u002Fbackpropagation","Backpropagation","Backpropagation is the chain rule run backward over a computational graph. We formalize the graph, derive the four backprop equations for an MLP, present the forward and backward passes as algorithms, and work a tiny two-layer net by hand with explicit numbers. The result: one scalar loss, reverse-mode autodiff, and a gradient for every parameter at twice the cost of a forward pass.\n",{"path":26936,"title":26937,"module":26921,"summary":26938},"\u002Fdeep-learning\u002Fneural-networks\u002Floss-functions-and-output-units","Loss Functions & Output Units","The last layer is where a network's hidden representation meets the task. Choosing an output unit and a loss is not two independent choices; maximum likelihood fixes the pair. We derive the standard couplings (linear\u002FMSE, sigmoid\u002FBCE, softmax\u002Fcross-entropy), show why softmax and cross-entropy were built to cancel into the residual $\\hat y - y$, and prove why squared error is the wrong loss for a saturating classifier.\n",{"path":26940,"title":26941,"module":26942,"summary":26943},"\u002Fdeep-learning\u002Foptimization\u002Fgradient-descent-and-sgd","Gradient Descent & SGD","Optimization","Training is descent on the empirical risk: step the parameters against the gradient. We derive the minibatch gradient as an unbiased estimator whose variance falls as $1\u002FB$, derive the learning-rate ceiling from the smoothness-stability bound $\\eta \u003C 2\u002FL$, and lay out the schedules (step, exponential, cosine, warmup) that anneal it over training.\n",{"path":26945,"title":26946,"module":26942,"summary":26947},"\u002Fdeep-learning\u002Foptimization\u002Fmomentum-and-adaptive-methods","Momentum & Adaptive Methods","Plain gradient descent zig-zags across ravines and moves slowly along flat valleys, because one global learning rate cannot suit a surface with wildly different curvature in different directions. Two fixes address the two problems: momentum accumulates a velocity that damps the oscillation and accelerates the drift, and adaptive methods give every parameter its own learning rate scaled by the history of its gradients. Adam fuses both, and is the default optimizer of modern deep learning.\n",{"path":26949,"title":26950,"module":26942,"summary":26951},"\u002Fdeep-learning\u002Foptimization\u002Finitialization","Weight Initialization","The initial weights determine whether training can succeed before the first gradient step. Initialize every weight equal and all hidden units compute the same function forever; initialize too small or too large and the signal vanishes or explodes as it crosses depth. A single variance condition, $n_{\\text{in}}\\mathrm{Var}(W)=1$, fixes both, and reading it off the forward and backward passes yields Xavier and He initialization directly.\n",{"path":26953,"title":26954,"module":26942,"summary":26955},"\u002Fdeep-learning\u002Foptimization\u002Fthe-optimization-landscape","The Optimization Landscape","The loss of a deep network is a non-convex surface in millions of dimensions, so local search carries no global guarantee, yet it works. We classify critical points by the eigenvalues of the Hessian, show that in high dimension nearly all of them are saddle points rather than bad local minima, and read off the practical terrain — plateaus, cliffs, ill-conditioning, and the sharp-versus-flat distinction that ties the geometry of a minimum to how well it generalizes.\n",{"path":26957,"title":26958,"module":26942,"summary":26959},"\u002Fdeep-learning\u002Foptimization\u002Fsecond-order-and-approximate-methods","Second-Order & Approximate Methods","Newton's method reads the curvature of the loss off its Hessian and jumps to the minimum of the local quadratic in a single step, rescaling away the ill-conditioning that slows first-order descent. We derive it, then explain the three obstacles that keep it out of deep learning: a $d \\times d$ Hessian for $d$ in the billions, an attraction to saddle points, and minibatch noise. The alternative is approximation (conjugate gradients, BFGS and L-BFGS, the natural gradient and Hessian-free methods), each buying some of Newton's curvature information without ever forming or inverting $H$.\n",{"path":26961,"title":26962,"module":26963,"summary":26964},"\u002Fdeep-learning\u002Fregularization\u002Fregularization-overview","Regularization Overview","Regularization","Regularization is any modification to a learning algorithm meant to lower test error at the possible expense of training error. We derive the bias–variance decomposition that explains why it helps, set up the two parameter-norm penalties, $L^2$ weight decay and $L^1$, derive their update rules and eigenbasis shrinkage, show geometrically why $L^1$ alone produces sparse weights (soft-thresholding), distinguish weight decay from loss-added $L^2$ under AdamW, and read both penalties through the two lenses that recur across the chapter: a norm-ball constraint via KKT, and a prior via MAP estimation.\n",{"path":26966,"title":26967,"module":26963,"summary":26968},"\u002Fdeep-learning\u002Fregularization\u002Fdropout-and-data-augmentation","Dropout & Data Augmentation","Two of the most effective regularizers add no penalty term at all; they perturb the computation instead. Dropout multiplies hidden units by a random Bernoulli mask, training an exponential ensemble of thinned subnetworks that share weights; inverted scaling collapses that ensemble into one cheap forward pass at test time. Data augmentation enlarges the training set with label-preserving transforms, injecting the invariances the task demands, and noise injection (input, weight, label smoothing, Mixup) generalizes the same idea into a continuous family.\n",{"path":26970,"title":26971,"module":26963,"summary":26972},"\u002Fdeep-learning\u002Fregularization\u002Fearly-stopping-and-parameter-sharing","Early Stopping & Parameter Sharing","Two cheap regularizers that cost no extra term in the loss. Early stopping treats training time itself as a hyperparameter (watch the validation curve, halt at its minimum, keep the best checkpoint), and for a quadratic objective it is provably equivalent to $L^2$ weight decay. Parameter sharing goes the other way: it constrains many weights to be _equal_, the prior behind every convolution and every recurrent step, and the reason a CNN has orders of magnitude fewer parameters than the dense net it replaces.\n",{"path":18525,"title":26974,"module":26963,"summary":26975},"Normalization","Normalization layers standardize activations to zero mean and unit variance inside the network, then hand the model a learnable scale and shift to undo the constraint when it pays to. Batch normalization does this across the batch and must keep separate train-time and test-time statistics; layer, instance, and group norm change only the axes they average over. The result is faster, better-conditioned optimization and a free dose of regularizing batch noise.\n",{"path":26977,"title":26978,"module":23403,"summary":26979},"\u002Fdeep-learning\u002Farchitectures\u002Fconvolutional-networks","Convolutional Networks","A convolutional network replaces the dense layer's all-to-all weight matrix with a small kernel slid across the input. Three structural commitments (sparse connectivity, parameter sharing, and translation equivariance) collapse the parameter count by orders of magnitude and bake the right prior for images directly into the architecture. We derive the convolution arithmetic, the output geometry, pooling, and the receptive field, then assemble the canonical stack.\n",{"path":18943,"title":26981,"module":23403,"summary":26982},"CNN Architectures","Six landmark networks, each contributing exactly one idea: LeNet's conv-pool stack, AlexNet's ReLU-and-dropout scale, VGG's $3\\times3$ uniformity, Inception's multi-scale module, ResNet's residual skip, and DenseNet's dense connectivity. The common thread is the degradation problem (why plain deeper nets train worse, not just overfit) and the residual block that solved it by keeping a $+1$ path open for the gradient.\n",{"path":17,"title":26984,"module":23403,"summary":26985},"Recurrent Networks","A recurrent network folds a sequence into a fixed-size hidden state, reusing one set of weights at every time step, the architectural prior that the same rule applies wherever it lands in time. Unrolling the recurrence exposes a deep feed-forward graph; backpropagation through it sums gradient contributions across all steps and chains a product of Jacobians, and that product is why long-range gradients vanish or explode. That failure motivates gated architectures.\n",{"path":154,"title":26987,"module":23403,"summary":26988},"LSTM & GRU","A plain recurrent network propagates its hidden state through a repeated weight-matrix multiply, and the Jacobian product that results vanishes or explodes long before a useful gradient can reach the early steps. Gated RNNs fix this with an additive memory path: a cell state that is carried forward almost unchanged, past which the gradient flows along a near-identity highway. We derive that highway, give the full LSTM and GRU equations, and compare the two.\n",{"path":23405,"title":5,"module":23403,"summary":23421},{"path":22458,"title":26991,"module":23403,"summary":26992},"The Transformer Architecture","The Transformer is the architecture built around the attention mechanism. This first part assembles the full encoder–decoder of \"Attention Is All You Need\" — embeddings and positional encoding, stacked self-attention and feed-forward sublayers wrapped in residual connections and LayerNorm, masked decoding and cross-attention — works through causal masking and the three modern families (encoder-only, decoder-only, encoder–decoder), and accounts for where the parameters and the $O(n^2)$ compute actually go.\n",{"path":26994,"title":26995,"module":23403,"summary":26996},"\u002Fdeep-learning\u002Farchitectures\u002Ftransformers-in-practice","Transformers in Practice","The Transformer makes no assumption about what a token represents. This part follows the architecture out of language: image patches feed a plain encoder (the Vision Transformer), the decoder-only half scales into the GPT line of large language models, and one substrate covers translation, retrieval, and multimodal grounding. We work the ViT patch arithmetic and a GPT parameter count by hand, then close on the empirical scaling laws — power-law loss, the Chinchilla compute-optimal balance, and emergent behavior — that made scale the dominant lever.\n",{"path":26998,"title":26999,"module":23403,"summary":27000},"\u002Fdeep-learning\u002Farchitectures\u002Fgraph-neural-networks","Graph Neural Networks","A graph neural network learns on data with no grid and no canonical ordering: atoms in a molecule, users in a social network, road segments in a map. The unifying idea is message passing — each node repeatedly aggregates its neighbors' states and updates its own — built to respect the one symmetry graphs demand, permutation equivariance. We derive the message-passing framework, specialize it into GCN, GraphSAGE, GAT, and GIN, read off graph-level outputs, and bound what message passing can and cannot tell apart.\n",{"path":27002,"title":27003,"module":23403,"summary":27004},"\u002Fdeep-learning\u002Farchitectures\u002Fstate-space-models","State-Space Models and Mamba","A state-space model carries a continuous linear hidden state through a sequence, and that linearity buys two equivalent algorithms from one set of weights: a recurrence that runs in linear time with constant memory, and a global convolution that trains in parallel. Long-range memory comes from how the transition matrix is initialized (HiPPO) and parameterized (S4's diagonal-plus-low-rank form). Mamba breaks the convolution on purpose, making the parameters input-dependent so the model can select what to remember, recovered at speed by a hardware-aware parallel scan.\n",{"path":27006,"title":27007,"module":27008,"summary":27009},"\u002Fdeep-learning\u002Ftheory\u002Fgeneralization-theory","Generalization Theory","Theory & Frontiers","Classical learning theory bounds the gap between training and test error by a model's capacity (VC dimension, Rademacher complexity), and predicts that a model with more parameters than data should overfit catastrophically. Modern networks do the opposite: they interpolate, even fit pure noise, and still generalize. We derive the classical bounds, work the bias-variance decomposition, show why the bounds go vacuous, and survey what replaced them: double descent, the interpolation threshold, margin and norm-based bounds, and the implicit bias of the optimizer itself.\n",{"path":27011,"title":27012,"module":27008,"summary":27013},"\u002Fdeep-learning\u002Ftheory\u002Fadversarial-robustness","Adversarial Robustness","A trained network can be fooled by a perturbation too small for a human to see: add a carefully aimed vector of magnitude $\\epsilon$ to a correctly classified image and the prediction flips. We derive the fast gradient sign method as the first-order-optimal step inside an $L_\\infty$ ball, explain the linearity hypothesis that makes high-dimensional models so easy to push around, build up to projected gradient descent, and frame adversarial training as a min-max robust-optimization problem with its own accuracy cost. Defenses beyond training continue in the next lesson.\n",{"path":27015,"title":27016,"module":27008,"summary":27017},"\u002Fdeep-learning\u002Ftheory\u002Fadversarial-defenses","Adversarial Defenses","Defending a network against an adversary is far harder than attacking one. This lesson covers the defense side: certified guarantees via randomized smoothing, the transferability that makes black-box attacks possible, and the recurring failure of gradient masking, where a defense hides the attacker's gradient instead of moving the decision boundary. It ends with the adaptive-attack discipline (BPDA, EOT, transfer) that every robustness claim must be tested against.\n",{"path":27019,"title":27020,"module":27008,"summary":27021},"\u002Fdeep-learning\u002Ftheory\u002Fbayesian-and-ensemble-methods","Bayesian & Ensemble Methods","A trained network returns a single point prediction and, with the softmax, a confidence, but that confidence is usually miscalibrated, collapsing to near- certainty even on inputs the model has never seen. This lesson covers uncertainty estimation for networks: the two kinds of uncertainty, the Bayesian posterior over weights and its tractable stand-ins (MC dropout, deep ensembles), and how to check whether a model's reported confidences match observed frequencies.\n",{"path":27023,"title":27024,"module":27008,"summary":27025},"\u002Fdeep-learning\u002Ftheory\u002Fdeep-equilibrium-models","Deep Equilibrium Models","A deep network need not be a fixed stack of layers; it can be a single weight-tied layer iterated to convergence, its output defined implicitly as the fixed point $z^\\star = f_\\theta(z^\\star, x)$. The forward pass becomes root-finding and the backward pass becomes implicit differentiation, so training costs O(1) memory regardless of effective depth. We derive both passes from the implicit function theorem and close the course on defining a layer by a fixed-point condition rather than an explicit stack.\n",{"path":27027,"title":27028,"module":27029,"summary":27030},"\u002Fdeep-learning\u002Fgenerative-models\u002Flinear-factor-models","Linear Factor Models","Generative Models","The simplest generative models share one template: a latent variable drawn from a fixed prior, run through a linear decoder, plus noise. Probabilistic PCA, factor analysis, independent component analysis, and sparse coding are all this template with a different prior on the latents and a different noise model. We derive each marginal, see why ICA needs non-Gaussianity to identify its sources, and show how sparse coding learns Gabor-like dictionary atoms.\n",{"path":27032,"title":27033,"module":27029,"summary":27034},"\u002Fdeep-learning\u002Fgenerative-models\u002Fautoencoders","Autoencoders","An autoencoder is a network trained to copy its input to its output through a narrow channel; the useful product is the bottleneck representation $h$, not the reconstruction. We derive the undercomplete autoencoder and prove its linear case recovers PCA, then trade the bottleneck for explicit regularization (sparse, denoising, contractive) and show how a denoising autoencoder learns the low-dimensional manifold the data lives on.\n",{"path":27036,"title":27037,"module":27029,"summary":27038},"\u002Fdeep-learning\u002Fgenerative-models\u002Fvariational-autoencoders","Variational Autoencoders","An autoencoder compresses, but its latent space has gaps: sample a point between two encodings and the decoder produces noise. The variational autoencoder fixes this by training a probabilistic encoder against a prior, so the latent space becomes a smooth, samplable density. We derive the evidence lower bound it maximizes, the reparameterization trick that lets gradients flow through a random sample, and the closed-form Gaussian regularizer that pulls the posterior toward the prior.\n",{"path":27040,"title":27041,"module":27029,"summary":27042},"\u002Fdeep-learning\u002Fgenerative-models\u002Fgenerative-adversarial-networks","Generative Adversarial Networks","A generative adversarial network trains two networks against each other: a generator that turns noise into samples, and a discriminator that tries to tell real data from forgeries. The game has a clean theory: the optimal discriminator is a likelihood ratio, and at equilibrium the generator minimizes the Jensen–Shannon divergence to the data, with a global optimum exactly when its distribution matches the data. We derive that result, fix the saturating loss that breaks training, and catalogue the failure modes (mode collapse, instability, vanishing gradients) and the architectural fixes.\n",{"path":27044,"title":27045,"module":27029,"summary":27046},"\u002Fdeep-learning\u002Fgenerative-models\u002Fautoregressive-and-normalizing-flows","Autoregressive Models & Normalizing Flows","Two families that provide exact likelihoods, each at a cost. Autoregressive models factor the joint by the probability chain rule and learn each conditional with a masked network: exact $\\log p(x)$, but sampling proceeds one coordinate at a time. Normalizing flows push a simple base density through an invertible map and read $\\log p(x)$ off the change-of-variables formula, trading architectural freedom for a cheap Jacobian determinant via triangular coupling layers.\n",{"path":27048,"title":27049,"module":27029,"summary":27050},"\u002Fdeep-learning\u002Fgenerative-models\u002Fenergy-based-and-boltzmann-machines","Energy-Based & Boltzmann Machines","Energy-based models replace an explicit density with a scalar energy and a Boltzmann normalization, $p(x) = e^{-E(x)}\u002FZ$: simple to specify, but with an intractable partition function $Z$. The Boltzmann machine and its restricted variant make the energy bilinear so the hidden units factorize, and contrastive divergence sidesteps $Z$ by replacing the model expectation with a few Gibbs steps started at the data. We close on the undirected deep models (DBNs and DBMs) and how they differ from the directed VAE.\n",{"path":27052,"title":27053,"module":27029,"summary":27054},"\u002Fdeep-learning\u002Fgenerative-models\u002Fdiffusion-and-score-based-models","Diffusion and Score-Based Models","Corrupt a data point with Gaussian noise in small steps until only noise remains, then train a network to undo one step at a time. We derive the forward process and its closed-form marginal, reduce the variational bound to the single noise-prediction objective that makes diffusion trainable, and show the score-matching view that unifies it with Langevin sampling and the continuous SDE. The lesson closes with DDIM fast sampling, classifier-free guidance, and the latent diffusion that powers modern text-to-image systems.\n",{"path":27056,"title":27057,"module":27058,"summary":27059},"\u002Fdeep-learning\u002Fprobabilistic-methods\u002Fstructured-probabilistic-models","Structured Probabilistic Models","Probabilistic Methods","A joint distribution over $n$ variables is a table with exponentially many entries; nobody can store it, fit it, or sample from it directly. Structure fixes this: a graph whose missing edges encode conditional independencies that factor the joint into small local pieces. We build the two dialects, directed (Bayesian networks) and undirected (Markov random fields), read independence off the graph, and connect the machinery to the latent-variable and energy-based models that power deep generative learning.\n",{"path":27061,"title":27062,"module":27058,"summary":27063},"\u002Fdeep-learning\u002Fprobabilistic-methods\u002Fmonte-carlo-and-mcmc","Monte Carlo & MCMC","Most quantities of interest in a probabilistic model are integrals nobody can compute in closed form: expectations, marginals, partition functions. Monte Carlo replaces the integral with an average over samples; importance sampling reweights samples from a tractable proposal; and when even sampling the target is hard, Markov-chain Monte Carlo builds a chain whose stationary distribution _is_ the target. We derive Metropolis–Hastings and Gibbs, analyze mixing, and close on the partition-function gradient that powers energy-based learning.\n",{"path":27065,"title":27066,"module":27058,"summary":27067},"\u002Fdeep-learning\u002Fprobabilistic-methods\u002Fapproximate-inference","Approximate Inference","In a latent-variable model the quantity we need, the posterior $p(h\\mid v)$ over hidden causes, is almost never computable, because its normalizer is an intractable sum over configurations. Approximate inference reframes the problem as optimization: maximize the evidence lower bound, a tractable functional whose gap to the true log-evidence equals a KL divergence. From that single bound fall expectation–maximization, mean-field variational inference, MAP, and the learned encoders behind variational autoencoders.\n",{"path":27069,"title":27070,"module":27071,"summary":27072},"\u002Fdeep-learning\u002Fpractical\u002Fpractical-methodology","Practical Methodology","Practical Deep Learning","Knowing the algorithms is half the job; the other half is a disciplined loop. Fix a goal and a metric, stand up an end-to-end baseline, then read the train\u002Fvalidation gap to decide whether the next move is more data or a bigger model. We detail that loop: choosing metrics under class imbalance, default baselines by data type, extrapolating the data a target needs, and guarding the data pipeline against the leaks and label bugs that corrupt every gradient. Hyperparameter tuning, debugging, and deployment continue in the sequel.\n",{"path":27074,"title":27075,"module":27071,"summary":27076},"\u002Fdeep-learning\u002Fpractical\u002Fhyperparameters-and-debugging","Hyperparameters & Debugging","The tuning half of the methodology loop. The learning rate is the one hyperparameter that dominates, so we tune it first, on a log scale, coarse to fine, and prefer random search to grid when only a few dials matter. Then an ordered debugging playbook — overfit one batch, check the loss at initialization against ln C, watch the gradient norm, gradient-check against centered finite differences — and, after launch, monitoring for train-test skew and distribution drift with confidence-based abstention.\n",{"path":27078,"title":27079,"module":27071,"summary":27080},"\u002Fdeep-learning\u002Fpractical\u002Frepresentation-learning","Representation Learning","A good representation makes a hard task easy by changing coordinates: it disentangles the factors of variation, spends its bits as a distributed code, and respects the low-dimensional manifold the data lives on. We make those three properties precise, recover the manifold hypothesis, and close on the first method that turned them into training practice — greedy layer-wise unsupervised pretraining — before the sequel picks up how the field learned to reuse those features.\n",{"path":27082,"title":27083,"module":27071,"summary":27084},"\u002Fdeep-learning\u002Fpractical\u002Ftransfer-learning","Transfer Learning","A representation learned once can be reused everywhere. We cover the main mechanisms of reuse: feature extraction versus fine-tuning, the generic-to-specific gradient of features that sets the freeze boundary, the learning-rate discipline that keeps borrowed weights from being erased, domain adaptation when only the input distribution shifts, and the modern arc from supervised transfer to self-supervised foundation models.\n",{"path":27086,"title":27087,"module":27071,"summary":27088},"\u002Fdeep-learning\u002Fpractical\u002Fapplications","Applications","We survey large-scale training (the hardware, the two axes of parallelism, mixed precision, and the compression tricks that shrink a model after it is trained), then specialize the same gradient loop to vision, language, speech, and recommendation. Each domain is a different prior bolted onto one optimizer: convolutional invariance for pixels, distributed word vectors for tokens, sequence transduction for audio, low-rank factorization for the user–item matrix.\n",{"path":27090,"title":27091,"module":27071,"summary":27092},"\u002Fdeep-learning\u002Fpractical\u002Fmodel-compression-and-distillation","Model Compression and Distillation","A trained network and a deployable one are rarely the same object. This lesson is the toolkit for closing that gap: knowledge distillation transfers a large teacher's soft, information-rich logits into a small student; pruning deletes the weights that contribute least; quantization swaps 32-bit floats for 8- or 4-bit integers; and low-rank factorization replaces a fat matrix with two thin ones. We derive each method, show what it costs in accuracy, and lay out which combinations win on which hardware.\n",{"path":27094,"title":27095,"module":27071,"summary":27096},"\u002Fdeep-learning\u002Fpractical\u002Fmeta-learning-and-few-shot","Meta-Learning and Few-Shot Learning","A deep network trained on one example per class overfits. Meta-learning targets this few-shot regime by training across a distribution of tasks so that a new task is learnable from a handful of examples. We formalize the $N$-way $K$-shot episode, then derive the two dominant families: metric methods that learn an embedding where distance classifies (Prototypical Networks), and optimization methods that learn an initialization a few gradient steps can adapt (MAML). We close on the link to transfer learning and to the in-context few-shot behavior of large language models.\n",{"path":27098,"title":27099,"module":27100,"summary":27101},"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Flarge-language-models","Large Language Models","Large Models & Agents","A large language model is a decoder-only Transformer trained on one objective, next-token prediction, then scaled until new behavior appears. This first part builds the object itself: the equivalence between next-token prediction and lossless compression, subword tokenization (BPE, WordPiece, Unigram, SentencePiece) worked on a real sentence, the four pretraining objectives and the attention masks that distinguish them, and the three model families (encoder-only, decoder-only, encoder--decoder) with their parameter budgets. Scaling, decoding, the KV cache, and alignment continue in part two.\n",{"path":27103,"title":27104,"module":27100,"summary":27105},"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fscaling-inference-and-alignment","Scaling, Inference, and Alignment of Language Models","Once a language model is built, three questions remain: how does it improve as it grows, how is it decoded and served affordably, and how is a raw next-token predictor turned into an assistant. We derive the Kaplan power laws and the Chinchilla compute-optimal balance, trace emergent abilities and in-context learning, catalog the decoding strategies from greedy to nucleus sampling, work the KV cache that makes generation quadratic instead of cubic, cover parameter-efficient adaptation by low-rank updates (LoRA), and close on the alignment stack: instruction tuning, RLHF, and DPO.\n",{"path":27107,"title":27108,"module":27100,"summary":27109},"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fseq2seq-pretraining-and-bart","Denoising Sequence-to-Sequence Pretraining: BART","BERT corrupts and reconstructs; GPT predicts the next token. Sequence-to-sequence pretraining unifies both by training a full encoder–decoder as a denoising autoencoder: corrupt the text with a noise function, then reconstruct the original through a bidirectional encoder and an autoregressive decoder. This first part derives the denoising objective, catalogs BART's five noise functions (with a worked Poisson-infilling budget), proves BART specializes to both BERT and GPT, and traces a dimension-annotated forward pass through its encoder--decoder. T5, PEGASUS, fine-tuning, and decoding continue in part two.\n",{"path":27111,"title":27112,"module":27100,"summary":27113},"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Ftext-to-text-transfer-and-conditional-generation","Text-to-Text Transfer and Conditional Generation","BART reconstructs a corrupted document; T5 pushes the same denoising idea into a single interface where every task is a string-to-string map. This second part covers T5's span corruption with sentinel tokens (with a worked token budget), PEGASUS's summarization-matched gap sentences and the MASS midpoint, supervised fine-tuning and beam-search decoding with a length penalty, the exposure-bias failure modes of autoregressive decoding, and a theorem showing why a bidirectional encoder--decoder strictly dominates a decoder-only model when the output is conditioned on a full input.\n",{"path":27115,"title":27116,"module":27100,"summary":27117},"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fspeech-and-audio-models","Speech Recognition: Front-Ends and Alignment","Speech is a long, high-rate sequence whose label is short and unaligned, so the whole subject turns on bridging that mismatch. This first part builds the spectral front-ends that compress a waveform into frames (STFT, mel spectrogram, MFCC, with a worked frame-count), derives CTC's marginalization over alignments and its forward-backward recursion with a two-frame numeric example, and contrasts it with attention-based seq2seq (LAS) and the RNN transducer. Self-supervised and weakly-supervised models, and text-to-speech, continue in part two.\n",{"path":27119,"title":27120,"module":27100,"summary":27121},"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fself-supervised-speech-and-synthesis","Self-Supervised Speech Models and Synthesis","The recognition front-ends and alignment losses of part one all need transcribed audio, which is scarce. This second part removes that dependence: wav2vec 2.0 learns speech representations from unlabeled audio by a masked contrastive objective, HuBERT swaps the contrast for masked prediction of clustered units, and Whisper trades curation for scale with weakly-supervised web audio and a multitask token interface. We close with text-to-speech (the same length mismatch run backwards) and a tour of speech foundation models, discrete audio codecs, and neural TTS.\n",{"path":27123,"title":27124,"module":27100,"summary":27125},"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fai-agents","AI Agents: Tools and Reasoning","A language model that only emits text is a function from prompt to prompt; an agent closes the loop, letting that model act on an environment, read back the result, and decide again. This first part formalizes the agent as a policy over interaction histories, builds out tool calling and the executor trust boundary, the ReAct interleaving of reasoning and action (with concrete traces), and search over thoughts: chain-of-thought, self-consistency, least-to-most, and Tree of Thoughts. Memory, retrieval, reflection, and multi-agent orchestration continue in part two.\n",{"path":27127,"title":27128,"module":27100,"summary":27129},"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fagent-memory-retrieval-and-orchestration","Agent Memory, Retrieval, and Orchestration","An agent's reasoning and tool use only matter if it can remember what it learned and coordinate work larger than one context window. This second part builds the systems around the loop: short-term scratchpad versus long-term vector store, retrieval-augmented generation with a worked softmax over passage scores, reflection (Reflexion, Self-Refine), and multi-agent orchestration. It closes on the failure modes that bound agents — invalid tool calls, horizon-error compounding, context overflow, non-terminating loops — and the benchmarks that score the full loop.\n",{"path":27131,"title":27132,"module":27100,"summary":27133},"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fmixture-of-experts","Mixture-of-Experts","A mixture-of-experts layer replaces one feed-forward network with many and a router that sends each token to only a few of them, so the parameter count and the per-token compute become separate dials. We derive the gated output, sparse top-$k$ routing softmax, the load-balancing loss that stops the router from collapsing onto a single expert, and expert\u002Ftoken capacity with dropping, then work the dimension-annotated tensor shapes and FLOP arithmetic. We trace the architectures from the sparsely-gated LSTM through GShard, Switch Transformer, and Mixtral, cover distributed expert parallelism, and close on the training dynamics, failure modes, and serving costs of a sparse model.\n",{"path":27135,"title":27136,"module":27100,"summary":27137},"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Fmultimodal-models","Multimodal Contrastive Learning","A multimodal model places images, text, and audio in one representation space, so a picture and its caption land close together. This first part builds the contrastive route: the shared embedding space and its residual modality gap, the Vision Transformer image encoder (patch embedding, CLS token, position embeddings, with shapes), the symmetric InfoNCE loss that trains the CLIP dual encoder from a batch similarity matrix (with a worked numeric step), and zero-shot classification as a softmax over class-prompt embeddings. Fusion and vision-language models continue in part two.\n",{"path":27139,"title":27140,"module":27100,"summary":27141},"\u002Fdeep-learning\u002Flarge-models-and-agents\u002Ffusion-and-vision-language-models","Fusion and Vision-Language Models","A contrastive model compares modalities but never lets one read another. This second part builds the fusion route: early, late, and cross-attention fusion, then the three designs that connect a frozen vision encoder to a frozen language model — Flamingo's zero-initialized gated cross-attention, BLIP-2's Q-Former, and LLaVA's linear projector. We work the token-budget arithmetic that separates them, name the object-hallucination and fine-detail failure modes, cover the contrastive-then- instruction-tune recipe and its retrieval\u002Fcaptioning\u002FVQA benchmarks, and close on natively multimodal models.\n",{"path":27143,"title":27144,"module":27145,"summary":27146},"\u002Fdeep-learning\u002Freinforcement-learning\u002Ffoundations-of-reinforcement-learning","Foundations of Reinforcement Learning","Reinforcement Learning","Reinforcement learning is the third paradigm: an agent learns to act by interacting with an environment that returns rewards, not labels. We formalize the interaction as a Markov decision process, define the value functions that rank states and actions, and derive the Bellman expectation and optimality equations that every method downstream solves. Dynamic programming gives the exact answer when the model is known, and its convergence rests on a single fact: the Bellman operator is a contraction.\n",{"path":27148,"title":27149,"module":27145,"summary":27150},"\u002Fdeep-learning\u002Freinforcement-learning\u002Fmodel-free-prediction-and-control","Model-Free Prediction and Control","When the dynamics are unknown, an agent cannot plan against a model; it must learn directly from sampled experience. We build prediction and control from two estimators of the same return: Monte Carlo averages whole episodes, while temporal-difference learning bootstraps from its own next estimate. We trace the bias-variance contrast between them, derive SARSA and Q-learning as the on-policy and off-policy forms of control, unify everything through n-step returns and eligibility traces, and close on the deadly triad that makes off-policy bootstrapping with function approximation diverge.\n",{"path":27152,"title":27153,"module":27145,"summary":27154},"\u002Fdeep-learning\u002Freinforcement-learning\u002Fdeep-q-networks","Deep Q-Networks","A Deep Q-Network replaces the tabular action-value function with a neural approximator $Q(s,a;\\theta)$ and trains it by regression toward a bootstrapped target. Naive online Q-learning with a network diverges, so DQN adds two stabilizers: an experience-replay buffer that decorrelates samples, and a periodically-frozen target network that holds the regression target still. We derive the loss, give the full algorithm and the Atari pipeline, and then layer on Double DQN, the dueling split, prioritized replay, and the Rainbow combination.\n",{"path":27156,"title":27157,"module":27145,"summary":27158},"\u002Fdeep-learning\u002Freinforcement-learning\u002Fpolicy-gradients-and-actor-critic","Policy Gradients and Actor-Critic Methods","Value-based reinforcement learning learns what each state is worth and acts greedily; policy-gradient methods skip the detour and optimize a parameterized policy directly by ascending the gradient of expected return. The policy gradient theorem makes this tractable through the log-derivative trick, turning an intractable gradient of an expectation into an expectation of a gradient. REINFORCE realizes the idea but suffers high variance; baselines, the advantage function, and actor-critic learning reduce it, and trust-region methods (TRPO, PPO) keep each update from destroying the policy it just learned.\n",{"path":27160,"title":27161,"module":27145,"summary":27162},"\u002Fdeep-learning\u002Freinforcement-learning\u002Frl-from-human-feedback","Reinforcement Learning from Human Feedback","Many objectives we want from a model, that it be helpful and harmless, are hard to write down but easy to judge by comparison. RLHF turns that asymmetry into a training signal: fit a reward model to pairwise human preferences under the Bradley-Terry likelihood, then fine-tune the policy to maximize that reward under a KL penalty toward a reference. We derive the reward loss, the KL-regularized RL objective and its closed-form optimum, then show how DPO inverts that optimum to collapse the whole pipeline into one supervised log-sigmoid loss, and survey IPO, KTO, RLAIF, and GRPO.\n",{"path":27164,"title":23114,"module":6,"summary":6},"\u002Fdeep-learning",{"path":27166,"title":27167,"module":25084,"summary":27168},"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Fequilibrium-state-variables-zeroth-law","Equilibrium, State Variables, and the Zeroth Law","Thermodynamics describes a many-body system by a handful of macroscopic variables and the equilibrium relations among them. This lesson fixes the vocabulary: systems and the walls that separate them, state variables versus path-dependent process quantities, quasi-static and reversible idealizations, and the zeroth law, whose transitivity of thermal equilibrium is what lets temperature exist as a number. The ideal-gas thermometer turns that number into a scale, and an equation of state ties the variables into a surface.\n",{"path":27170,"title":27171,"module":25084,"summary":27172},"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Ffirst-law-heat-and-work","The First Law: Internal Energy, Heat, and Work","The first law is energy conservation for a system that exchanges energy as both heat and work. Internal energy is a state function with an exact differential; heat and work are path-dependent process quantities. This lesson states $\\d U=\\delta Q+\\delta W$, computes compression work as an area on the $P$–$V$ plane, defines the heat capacities $C_V$ and $C_P$ and the enthalpy that makes $C_P$ natural, and works the isothermal and adiabatic processes of an ideal gas, including the adiabat $PV^\\gamma=\\text{const}$.\n",{"path":27174,"title":27175,"module":25084,"summary":27176},"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Fsecond-law-entropy-and-the-carnot-bound","The Second Law, Carnot Cycles, and Entropy","The second law forbids the free conversion of heat into work. This lesson states the Kelvin and Clausius forms, proves them equivalent, and analyzes the Carnot cycle to get the efficiency bound $1-T_c\u002FT_h$. Carnot's theorem makes that bound universal and defines the thermodynamic temperature scale. The Clausius inequality $\\oint \\delta Q\u002FT\\le 0$ then constructs entropy as a state function, $\\d S=\\delta Q_{\\rm rev}\u002FT$, whose non-decrease in isolated systems is the arrow of time.\n",{"path":27178,"title":27179,"module":25084,"summary":27180},"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Fthermodynamic-potentials-and-maxwell-relations","Thermodynamic Potentials and Maxwell Relations","The fundamental relation $\\d U=T\\,\\d S-P\\,\\d V+\\mu\\,\\d N$ packages the first and second laws into one exact differential. Legendre transforms swap each conjugate pair to produce the Helmholtz, enthalpy, Gibbs, and grand potentials, each minimized under its own natural variables. Equality of mixed second partials of these potentials gives the Maxwell relations, which convert unmeasurable entropy derivatives into measurable ones from the equation of state.\n",{"path":27182,"title":27183,"module":25084,"summary":27184},"\u002Fstatistical-mechanics\u002Fthermodynamics\u002Fstability-response-functions-and-the-third-law","Response Functions, Stability, and the Third Law","Response functions — heat capacities, compressibilities, thermal expansion — are the second derivatives of the potentials and the quantities an experiment actually measures. This lesson derives the general relation $C_P-C_V=TV\\alpha^2\u002F\\kappa_T$, shows that convexity of the potentials forces the stability conditions $C_V>0$ and $\\kappa_T>0$, and states the third law: entropy approaches a constant as $T\\to0$, so heat capacities and expansion coefficients vanish there and absolute zero is unattainable.\n",{"path":27186,"title":27187,"module":27188,"summary":27189},"\u002Fstatistical-mechanics\u002Ffoundations\u002Fclassical-statistics-and-equipartition","Classical Statistics and Equipartition","Microstates, Phase Space, and Statistical Entropy","A liter of gas holds on the order of a trillion trillion molecules, far too many to track by their equations of motion. Classical statistical mechanics replaces the trajectories with a single probability law, the Boltzmann distribution, and reads the measurable properties of matter off it: the Maxwell speed distribution, the average energy per degree of freedom, and the heat capacities of gases and solids — together with the low-temperature failures that forced the quantum revision.\n",{"path":27191,"title":27192,"module":27188,"summary":27193},"\u002Fstatistical-mechanics\u002Ffoundations\u002Fphase-space-and-liouvilles-theorem","Phase Space, Trajectories, and Liouville's Theorem","A classical system of N particles is one point in a 6N-dimensional phase space, and its evolution is a single trajectory driven by Hamilton's equations. This lesson builds that geometric picture, introduces the phase-space density of an ensemble, and proves Liouville's theorem: the density is carried by the flow as an incompressible fluid, so phase-space volume is conserved. The stationary densities of equilibrium follow as functions of the conserved quantities alone.\n",{"path":27195,"title":27196,"module":27188,"summary":27197},"\u002Fstatistical-mechanics\u002Ffoundations\u002Fensembles-and-the-equal-probability-postulate","Ensembles and the Postulate of Equal a Priori Probabilities","An ensemble is a probability distribution over the microstates of a system. This lesson states the single postulate on which equilibrium statistical mechanics rests — that an isolated system in equilibrium is equally likely to be in any of its accessible microstates — and works out its consequences: the accessible phase-space volume, the overwhelming dominance of the most probable macrostate as the particle number grows, and the ergodic hypothesis that lets a time average be replaced by an ensemble average.\n",{"path":27199,"title":27200,"module":27188,"summary":27201},"\u002Fstatistical-mechanics\u002Ffoundations\u002Fstatistical-entropy-boltzmann-and-gibbs","Statistical Entropy: Boltzmann and Gibbs","Entropy is the logarithm of the number of accessible microstates. This lesson builds the two statistical entropies — Boltzmann's S = k ln Omega for an isolated system and Gibbs's S = -k sum p ln p for any ensemble — proves they agree for a uniform distribution, and connects both to Shannon's measure of missing information. The second law emerges as the drift toward maximum multiplicity, and maximizing the Gibbs entropy under constraints previews the canonical distribution.\n",{"path":27203,"title":27204,"module":27205,"summary":27206},"\u002Fstatistical-mechanics\u002Fmicrocanonical\u002Fmicrocanonical-ensemble-and-entropy","The Microcanonical Ensemble and Statistical Entropy","The Microcanonical Ensemble","An isolated system holds its energy, volume, and particle number fixed, and the fundamental postulate assigns equal probability to every microstate on its energy shell. This lesson builds the microcanonical distribution, defines the enclosed phase-space volume $\\Gamma(E)$, the surface density of states $\\omega(E)=\\d\\Gamma\u002F\\d E$, and the shell count $\\Omega(E)$, shows their logarithms agree to $O(\\ln N)$ for large $N$, and reads the Boltzmann entropy $S=k\\ln\\Omega$ off the count. The measure factors $h^{3N}$ and $N!$ enter here and make $S$ extensive.\n",{"path":27208,"title":27209,"module":27205,"summary":27210},"\u002Fstatistical-mechanics\u002Fmicrocanonical\u002Fequilibrium-conditions-temperature-pressure-chemical-potential","Thermal, Mechanical, and Diffusive Equilibrium","Two isolated subsystems that can exchange energy, volume, or particles reach equilibrium at the partition that maximizes their combined entropy. Setting the derivative of the total entropy to zero identifies the statistical definitions $1\u002FT=(\\partial S\u002F\\partial E)$, $P\u002FT=(\\partial S\u002F\\partial V)$, and $-\\mu\u002FT=(\\partial S\u002F\\partial N)$, shows heat flows from hot to cold as an entropy increase, and recovers the fundamental relation $\\d S=(\\d E+P\\,\\d V-\\mu\\,\\d N)\u002FT$ from pure counting.\n",{"path":27212,"title":27213,"module":27205,"summary":27214},"\u002Fstatistical-mechanics\u002Fmicrocanonical\u002Fideal-gas-phase-space-and-the-sackur-tetrode-entropy","The Ideal Gas, Phase-Space Volume, and the Sackur–Tetrode Entropy","The monatomic ideal gas is the first system whose microcanonical count can be done in closed form. The momentum integral is the volume of a $3N$-dimensional ball of radius $\\sqrt{2mE}$, the configuration integral is $V^N$, and together they give the Sackur–Tetrode entropy $S=Nk[\\ln(V\u002FN\\lambda^3)+5\u002F2]$ with the thermal wavelength $\\lambda=h\u002F\\sqrt{2\\pi mkT}$. The formula matches the measured entropy of helium, fixes the classical regime $n\\ll n_Q$, and shows why the $N!$ is needed for extensivity.\n",{"path":27216,"title":27217,"module":27205,"summary":27218},"\u002Fstatistical-mechanics\u002Fmicrocanonical\u002Ftwo-state-systems-paramagnets-and-negative-temperature","Two-State Systems, Paramagnets, and Negative Temperature","The ideal two-state paramagnet has a multiplicity counted by the binomial coefficient, an entropy that is an inverted dome in the energy, and a temperature read from the slope $1\u002FT=\\partial S\u002F\\partial E$. Because the energy is bounded above, the slope changes sign past the entropy maximum: a population-inverted spin system has a negative absolute temperature, which is hotter than any positive temperature. Nuclear-spin experiments and lasers realize the inverted state.\n",{"path":27220,"title":27221,"module":27222,"summary":27223},"\u002Fstatistical-mechanics\u002Fcanonical\u002Fcanonical-ensemble-and-the-boltzmann-distribution","The Canonical Ensemble and the Boltzmann Distribution","The Canonical Ensemble","A system held at fixed temperature by contact with a heat reservoir is described by the canonical ensemble. Expanding the reservoir entropy to first order in the system energy gives the Boltzmann distribution $p_i\\propto e^{-\\beta E_i}$, and the same law follows from maximizing the Gibbs entropy at fixed mean energy. Both routes identify $\\beta=1\u002Fk_BT$ and fix the probability of every microstate from the temperature alone.\n",{"path":27225,"title":27226,"module":27222,"summary":27227},"\u002Fstatistical-mechanics\u002Fcanonical\u002Fpartition-function-and-the-helmholtz-free-energy","The Partition Function and the Helmholtz Free Energy","The normalizing sum of the Boltzmann distribution, the partition function $Z=\\sum_i e^{-\\beta E_i}$, is a generating function for the thermodynamics. The mean energy is $-\\partial\\ln Z\u002F\\partial\\beta$, and the Gibbs entropy of the canonical distribution collapses to the bridge relation $F=-k_BT\\ln Z$. From $F$ every thermodynamic quantity follows by differentiation, and $Z$ factorizes over independent degrees of freedom.\n",{"path":27229,"title":27230,"module":27222,"summary":27231},"\u002Fstatistical-mechanics\u002Fcanonical\u002Fenergy-fluctuations-and-ensemble-equivalence","Energy Fluctuations and the Equivalence of Ensembles","In the canonical ensemble the energy fluctuates, and the second derivative of $\\ln Z$ gives its variance. The fluctuation–response identity $\\langle\\Delta E^2\\rangle = k_BT^2C_V$ ties the spread of the energy to the heat capacity, and the relative fluctuation falls as $1\u002F\\sqrt{N}$. In the thermodynamic limit the canonical energy distribution is a sharp spike, and the canonical and microcanonical ensembles predict the same thermodynamics.\n",{"path":27233,"title":27234,"module":27222,"summary":27235},"\u002Fstatistical-mechanics\u002Fcanonical\u002Fthe-einstein-solid-and-harmonic-systems","Harmonic Systems: The Einstein Solid and Vibrational Heat Capacity","A quantum harmonic oscillator has a geometric partition function summed in closed form, giving a mean energy $\\hbar\\omega(\\tfrac12+\\langle n\\rangle)$ with the Bose occupation factor. Modeling a solid as $3N$ independent oscillators yields a heat capacity that rises from zero and saturates at the Dulong–Petit value $3Nk_B$. The Einstein temperature sets the crossover, and the model's exponential low-temperature falloff, too steep against the observed $T^3$, motivates the Debye theory.\n",{"path":27237,"title":27238,"module":27222,"summary":27239},"\u002Fstatistical-mechanics\u002Fcanonical\u002Fparamagnetism-and-the-schottky-anomaly","Paramagnetism, Two-Level Systems, and the Schottky Anomaly","A magnetic moment in a field is a two-level system whose partition function is a hyperbolic cosine. The magnetization of a spin-$\\tfrac12$ paramagnet is $N\\mu\\tanh(\\mu B\u002Fk_BT)$, generalizing to the Brillouin function for spin $J$; it gives Curie's law $\\chi\\propto 1\u002FT$ at high temperature and saturates at low temperature. A finite level gap produces the Schottky heat-capacity peak, and the temperature dependence of the entropy on the field is the basis of adiabatic demagnetization cooling.\n",{"path":27241,"title":27242,"module":27243,"summary":27244},"\u002Fstatistical-mechanics\u002Fclassical-gas\u002Fideal-gas-partition-function-and-the-gibbs-paradox","The Ideal Gas Partition Function and the Gibbs Paradox","The Classical Ideal Gas","The classical monatomic ideal gas built from the partition function. The single-particle sum is $z_1=V\u002F\\lambda^3$ with the thermal de Broglie wavelength $\\lambda$; the $N$-particle partition function is $z_1^N\u002FN!$, and the $N!$ is forced by indistinguishability. From $Z$ the ideal-gas law, $U=\\tfrac32 Nk_BT$, and the Sackur–Tetrode entropy follow. The $N!$ makes the entropy extensive and resolves the Gibbs paradox: mixing identical gases produces no entropy change.\n",{"path":27246,"title":27247,"module":27243,"summary":27248},"\u002Fstatistical-mechanics\u002Fclassical-gas\u002Fequipartition-and-the-virial-theorem","Equipartition and the Virial Theorem","The equipartition theorem derived from the canonical ensemble: every phase-space coordinate that enters the Hamiltonian quadratically carries a mean energy $\\tfrac12 k_BT$. The generalized form $\\langle x_i\\,\\partial H\u002F\\partial x_j\\rangle = k_BT\\,\\delta_{ij}$ contains equipartition and the classical virial theorem as special cases. Equipartition fixes the classical heat capacities, fails by quantum freeze-out when a level gap exceeds $k_BT$, and shifts for a relativistic gas whose energy is linear rather than quadratic in momentum.\n",{"path":27250,"title":27251,"module":27243,"summary":27252},"\u002Fstatistical-mechanics\u002Fclassical-gas\u002Fmolecular-gases-rotation-and-vibration","Molecular Gases: Rotational and Vibrational Degrees of Freedom","The internal partition function of a diatomic gas factorizes into translational, rotational, vibrational, and electronic parts. The rigid rotor gives a rotational temperature $\\theta_{\\rm rot}$; the harmonic bond gives a vibrational temperature $\\theta_{\\rm vib}$. Each mode contributes to the heat capacity only above its characteristic temperature, producing the diatomic $C_V$ staircase from $\\tfrac32 R$ to $\\tfrac52 R$ to $\\tfrac72 R$. Homonuclear molecules carry a symmetry number, and hydrogen splits into ortho and para species.\n",{"path":27254,"title":27255,"module":27256,"summary":27257},"\u002Fstatistical-mechanics\u002Fgrand-canonical\u002Fgrand-canonical-ensemble-and-the-grand-partition-function","The Grand Canonical Ensemble","Grand Canonical Ensemble","When a system exchanges both energy and particles with a reservoir, the reservoir fixes its temperature and its chemical potential. Expanding the reservoir entropy to first order in the exchanged energy and particle number gives the Gibbs factor $e^{-\\beta(E-\\mu N)}$, and summing it over every microstate of every particle number gives the grand partition function $\\Xi$. The grand potential $\\Phi = -k_BT\\ln\\Xi = -PV$ generates the mean particle number, energy, entropy, and pressure by differentiation.\n",{"path":27259,"title":27260,"module":27256,"summary":27261},"\u002Fstatistical-mechanics\u002Fgrand-canonical\u002Fchemical-potential-fugacity-and-number-fluctuations","Chemical Potential, Fugacity, and Number Fluctuations","The chemical potential is the energy to add one particle at fixed entropy and volume, equal to the slope of the free energy in the particle number. For the classical ideal gas $\\mu=k_BT\\ln(n\\lambda^3)$ is large and negative, and the fugacity $z=n\\lambda^3$ is small. The grand ensemble makes the particle number fluctuate; its variance $\\langle\\Delta N^2\\rangle=k_BT(\\partial N\u002F\\partial\\mu)$ equals $k_BT\\,N^2\\kappa_T\u002FV$, tying density fluctuations to the isothermal compressibility. Equality of $\\mu$ is the condition for diffusive equilibrium and phase coexistence.\n",{"path":27263,"title":27264,"module":27256,"summary":27265},"\u002Fstatistical-mechanics\u002Fgrand-canonical\u002Fensemble-summary-and-the-thermodynamic-web","The Three Ensembles and the Thermodynamic Web","The microcanonical, canonical, and grand canonical ensembles hold different variables fixed and generate different potentials — the entropy $S$, the Helmholtz free energy $F$, and the grand potential $\\Phi$ — linked by Legendre transforms that trade each fixed variable for its conjugate. Each successive ensemble lets one more quantity fluctuate. In the thermodynamic limit the three agree, the relative fluctuations vanishing as $1\u002F\\sqrt{N}$; the ideal gas gives the same equation of state in all three. The choice of ensemble is a matter of convenience, set by which sum is easiest.\n",{"path":27267,"title":27268,"module":27269,"summary":27270},"\u002Fstatistical-mechanics\u002Fquantum-statistics\u002Fquantum-statistics-bose-einstein-and-fermi-dirac","Quantum Statistics — Bose-Einstein and Fermi-Dirac","Quantum Statistics","Quantum particles of the same kind are genuinely indistinguishable: no label survives an overlap of their wave functions. Counting states with that constraint replaces the Boltzmann distribution with two quantum laws — the Bose-Einstein distribution for integer-spin particles, which clump into shared states, and the Fermi-Dirac distribution for half-integer-spin particles, which exclude one another. Both reduce to Boltzmann in the dilute, hot limit, and a de Broglie criterion says exactly when.\n",{"path":27272,"title":27273,"module":27269,"summary":27274},"\u002Fstatistical-mechanics\u002Fquantum-statistics\u002Fderiving-the-quantum-distributions","Deriving the Quantum Distributions from the Grand Ensemble","The Bose-Einstein and Fermi-Dirac distributions follow from one observation: in the occupation-number representation the single-particle modes are independent, so the grand partition function factorizes into one factor per mode. A boson mode sums a geometric series over all occupancies; a fermion mode sums two terms. Differentiating each factor gives the mean occupation $1\u002F(e^{\\beta(\\varepsilon-\\mu)}\\mp 1)$, the Maxwell-Boltzmann limit when occupancies are small, and the occupation fluctuations that distinguish bunching from anti-bunching.\n",{"path":27276,"title":27277,"module":27269,"summary":27278},"\u002Fstatistical-mechanics\u002Fquantum-statistics\u002Fthe-classical-limit-and-quantum-concentration","The Classical Limit and Quantum Concentration","When every single-particle level is nearly empty, both quantum distributions collapse to the Maxwell-Boltzmann form, and the fugacity equals the ratio of the number density to the quantum concentration $n_Q = 1\u002F\\lambda^3$. The gas is classical when $n \\ll n_Q$, degenerate when $n \\gtrsim n_Q$. The chemical potential is large and negative in the classical regime and rises through zero as the gas degenerates. The leading quantum correction to the ideal-gas law is a second virial term that lowers the pressure for bosons and raises it for fermions — a statistical attraction and repulsion with no interaction behind it.\n",{"path":27280,"title":27281,"module":27269,"summary":27282},"\u002Fstatistical-mechanics\u002Fquantum-statistics\u002Fideal-quantum-gases-general-framework","Ideal Quantum Gases: The General Framework","Every ideal quantum gas is handled by one calculation. The sum over single-particle modes becomes an energy integral weighted by a density of states $g(\\varepsilon)\\propto\\varepsilon^{1\u002F2}$, and the number and pressure reduce to the Bose and Fermi functions $g_\\nu(z)$ and $f_\\nu(z)$ of the fugacity. An integration by parts fixes $PV=\\tfrac23 U$ for a nonrelativistic gas and $PV=\\tfrac13 U$ for an ultrarelativistic one, independent of statistics. Specializing the density of states and the chemical potential then produces the photon gas, phonons, the Bose gas, and the Fermi gas as four branches of the same framework.\n",{"path":27284,"title":27285,"module":27286,"summary":27287},"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fbose-einstein-condensation-and-the-fermion-gas","Bose-Einstein Condensation and the Fermion Gas","Bosonic Systems","Below a critical temperature a boson gas drops a macroscopic fraction of its particles into the single ground state — Bose-Einstein condensation, the mechanism behind superfluid helium and the dilute-atom condensates cooled to nanokelvin. The same statistics applied to a photon gas reproduces Planck's blackbody spectrum. Fermions do the opposite: forbidden from sharing states, they fill every level up to the Fermi energy, and that filled sea governs the electrons in metals and the pressure that holds up a white dwarf.\n",{"path":27289,"title":27290,"module":27286,"summary":27291},"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fthe-photon-gas-and-plancks-radiation-law","The Photon Gas and Planck's Radiation Law","Electromagnetic radiation in equilibrium with cavity walls is a gas of non-conserved bosons, and non-conservation forces the chemical potential to zero. Counting standing-wave modes with two polarizations and weighting each by the Bose occupation gives the Planck spectral energy density. Its low-frequency tail reproduces the classical Rayleigh-Jeans law and the ultraviolet catastrophe; the Bose factor cuts the divergence off at high frequency and the peak obeys Wien's displacement law.\n",{"path":27293,"title":27294,"module":27286,"summary":27295},"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fblackbody-thermodynamics-and-radiation-pressure","Blackbody Thermodynamics and Radiation Pressure","Integrating the Planck spectrum over all frequencies gives the total energy density proportional to the fourth power of temperature — the Stefan-Boltzmann law — and the isotropy of a relativistic gas fixes the radiation pressure at one third of the energy density. From the free energy follow the entropy and heat capacity, both proportional to T cubed, and the adiabatic law for radiation. The results govern the pressure inside stars and the cooling of the cosmic microwave background as the universe expands.\n",{"path":27297,"title":27298,"module":27286,"summary":27299},"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fphonons-and-the-debye-model","Phonons and the Debye Model","The vibrations of a crystal lattice are quantized into phonons — bosons of zero chemical potential, counted exactly like cavity photons but with three polarizations, a finite sound speed, and a total of 3N modes. The Debye model replaces the true dispersion by a linear one cut off at a frequency that enforces that count. It gives the correct low-temperature T-cubed heat capacity the Einstein model missed and recovers the Dulong-Petit value at high temperature.\n",{"path":27301,"title":27302,"module":27286,"summary":27303},"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fbose-einstein-condensation-derived","Bose-Einstein Condensation Derived","For a gas of conserved bosons the excited states can hold only a finite number of particles at fixed temperature, set by the Bose function at unit fugacity. When the total exceeds that ceiling the surplus collapses into the single ground state, which the continuum density-of-states integral misses and which must be restored by hand. This fixes the critical temperature, the condensate fraction, and the fact that a uniform gas condenses only in three or more dimensions.\n",{"path":27305,"title":27306,"module":27286,"summary":27307},"\u002Fstatistical-mechanics\u002Fbose-systems\u002Fthermodynamics-of-the-bose-gas-and-superfluidity","Thermodynamics of the Bose Gas and Superfluidity","The energy and pressure of the ideal Bose gas follow from the Bose function at the order above the density, and below the critical temperature the pressure depends on temperature alone because the condensate carries none. The heat capacity rises to a cusp at the transition. Real superfluid helium departs from the ideal gas because interactions matter: the Landau criterion ties frictionless flow to the phonon-roton excitation spectrum, and the two-fluid model carries a second sound.\n",{"path":27309,"title":27310,"module":27311,"summary":27312},"\u002Fstatistical-mechanics\u002Ffermi-gas\u002Fthe-ideal-fermi-gas-at-zero-temperature","The Ideal Fermi Gas at Zero Temperature","Degenerate Fermi Gas","At absolute zero a gas of non-interacting fermions fills every single-particle state up to the Fermi energy and leaves the rest empty, a filled Fermi sphere in momentum space. This lesson computes the Fermi momentum, energy, and temperature from the density, the density of states, the total ground-state energy, and the degeneracy pressure that grows as $n^{5\u002F3}$. Numerical Fermi energies for metals set the scale: they are electron-volts, so room temperature is deep in the degenerate regime.\n",{"path":27314,"title":27315,"module":27311,"summary":27316},"\u002Fstatistical-mechanics\u002Ffermi-gas\u002Fsommerfeld-expansion-and-electrons-in-metals","The Sommerfeld Expansion and Electrons in Metals","Turning on a small temperature blurs the Fermi step over a shell of width $k_BT$ around $\\epsilon_F$. The Sommerfeld expansion turns integrals over the Fermi function into a power series in $(k_BT\u002F\\epsilon_F)^2$, giving the shift of the chemical potential and a heat capacity linear in $T$. This resolves the old puzzle of the missing electronic heat capacity, predicts the combined $C=\\gamma T+AT^3$ of a metal, and gives the temperature-independent Pauli paramagnetism of the electron gas.\n",{"path":27318,"title":27319,"module":27311,"summary":27320},"\u002Fstatistical-mechanics\u002Ffermi-gas\u002Fwhite-dwarfs-and-the-chandrasekhar-limit","White Dwarfs and the Chandrasekhar Limit","A white dwarf is held up against its own gravity by the degeneracy pressure of its electrons. Balancing that pressure against gravity gives a mass-radius relation $R\\propto M^{-1\u002F3}$: heavier white dwarfs are smaller and denser. As the density rises the electrons turn relativistic, the pressure softens from $n^{5\u002F3}$ to $n^{4\u002F3}$, and the star can no longer support itself above a critical mass. This lesson derives that Chandrasekhar mass, about $1.4\\,M_\\odot$, and what lies beyond it.\n",{"path":27322,"title":27323,"module":27311,"summary":27324},"\u002Fstatistical-mechanics\u002Ffermi-gas\u002Fneutron-stars-and-nuclear-matter","Neutron Stars and Dense Matter","When a collapsing core passes nuclear density, electron capture converts the matter to neutrons and their degeneracy pressure takes over. The same balance that fixes a white dwarf, rescaled by the neutron mass, gives a neutron star of a few solar masses in a ten-kilometre radius. General relativity is no longer a correction: the Tolman-Oppenheimer-Volkoff equation replaces the Newtonian balance and sets a maximum mass around two solar masses. This lesson rescales the Fermi-gas argument, states where it breaks, and places the compact objects in one stability sequence.\n",{"path":27326,"title":27327,"module":27328,"summary":27329},"\u002Fstatistical-mechanics\u002Finteractions\u002Fthe-cluster-expansion-and-virial-coefficients","The Cluster Expansion and Virial Coefficients","Interacting Gases","A real gas departs from $PV=Nk_BT$ because its molecules interact. The configuration integral factors through the Mayer function $f_{ij}=e^{-\\beta u_{ij}}-1$, and expanding it in powers of density produces the virial expansion $PV\u002FNk_BT = 1 + B_2(T)n + B_3(T)n^2 + \\cdots$. The second virial coefficient $B_2(T)=-\\tfrac12\\int f\\,\\d^3r$ is a single integral over the pair potential; it is positive for a hard core, negative for an attractive well, and vanishes at the Boyle temperature where the two balance.\n",{"path":27331,"title":27332,"module":27328,"summary":27333},"\u002Fstatistical-mechanics\u002Finteractions\u002Fthe-van-der-waals-gas-and-liquid-gas-coexistence","The van der Waals Gas and Liquid-Gas Coexistence","Resumming the second virial coefficient $B_2=b-a\u002Fk_BT$ into an equation of state gives the van der Waals model $(P+a\u002Fv^2)(v-b)=k_BT$, the simplest theory of a fluid that condenses. Below the critical temperature its isotherms develop a mechanically unstable loop; the Maxwell equal-area construction replaces the loop with a coexistence tie line. The critical point sits at $v_c=3b$, $k_BT_c=8a\u002F27b$, $P_c=a\u002F27b^2$, and the model predicts universal but incorrect critical exponents because it ignores fluctuations.\n",{"path":27335,"title":27336,"module":27328,"summary":27337},"\u002Fstatistical-mechanics\u002Finteractions\u002Fquantum-gases-with-interactions-and-exchange","Quantum Gases with Interactions and Statistical Exchange","A quantum gas has a nonzero second virial coefficient even with no forces between the particles: symmetrization alone produces an effective statistical interaction, attractive for bosons and repulsive for fermions, with range the thermal wavelength $\\lambda$. This lesson derives that exchange contribution $B_2=\\mp\\lambda^3\u002F2^{5\u002F2}g$, writes it as a statistical potential $v_s(r)=-k_BT\\ln(1\\pm e^{-2\\pi r^2\u002F\\lambda^2})$, and shows how real interactions add on top through the Beth-Uhlenbeck phase-shift formula, reducing at low temperature to a single scattering length.\n",{"path":27339,"title":27340,"module":27341,"summary":27342},"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fphases-coexistence-and-classification","Phases, Coexistence, and the Classification of Transitions","Phase Transitions","A phase transition is a point where the free energy of a substance loses analyticity, so a small change in temperature or pressure produces a qualitative change of state. This lesson maps the coexistence curves of a pure substance, derives the Clausius-Clapeyron relation between the slope of a coexistence line and its latent heat, and separates first-order transitions (discontinuous entropy and density) from continuous ones (a vanishing order parameter and divergent response). The Ehrenfest scheme, the order parameter, and the triple and critical points fix the vocabulary the rest of the module builds on.\n",{"path":27344,"title":27345,"module":27341,"summary":27346},"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fthe-ising-model-and-exact-solutions","The Ising Model and Exact Results","The Ising model reduces cooperative ordering to spins on a lattice coupled to their neighbors, and the same Hamiltonian describes uniaxial magnets, the liquid-gas critical point through the lattice gas, and binary alloys. This lesson solves the one-dimensional chain exactly with the transfer matrix, shows by a domain-wall argument why one dimension has no ordered phase at any positive temperature, contrasts the survival of order in two dimensions, and quotes Onsager's exact two-dimensional results: the critical temperature, the logarithmically divergent heat capacity, and the magnetization exponent one eighth.\n",{"path":27348,"title":27349,"module":27341,"summary":27350},"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fmean-field-theory-and-the-weiss-model","Mean-Field Theory and Spontaneous Symmetry Breaking","Mean-field theory replaces the neighbors of each spin by their average, turning the interacting Ising model into a single spin in a self-consistent field. The resulting equation m = tanh(beta J z m + beta h) has only the zero solution above a critical temperature and gains a nonzero root below it, giving spontaneous magnetization and a mean-field critical temperature k T_c = J z. The Bragg-Williams free energy turns single-welled above T_c and double-welled below, the picture of spontaneous symmetry breaking. The approximation is exact in high dimension and fails below the upper critical dimension four, quantified by the Ginzburg criterion.\n",{"path":27352,"title":27353,"module":27341,"summary":27354},"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fcritical-exponents-and-landau-theory","Critical Exponents, Scaling, and Landau Theory","Near a continuous transition every singular quantity follows a power law in the reduced temperature, and the exponents alpha, beta, gamma, delta, nu, and eta encode the transition more sharply than T_c itself. Landau theory expands the free energy in the order parameter and delivers the mean-field exponents in a few lines. They disagree with experiment and with the exact two-dimensional Ising values, but the exponents are not independent: the scaling relations of Rushbrooke, Widom, Fisher, and Josephson tie them together, and the correlation length sets the length scale that organizes universality classes.\n",{"path":27356,"title":27357,"module":27341,"summary":27358},"\u002Fstatistical-mechanics\u002Fphase-transitions\u002Fthe-renormalization-group-idea","Scaling and the Renormalization-Group Idea","At a critical point fluctuations exist on every length scale, so the system looks the same after coarse-graining. The renormalization group makes this self-similarity a computation: group spins into blocks, integrate out the short scales, and track how the couplings change. The transformation has fixed points, and the flow near a critical fixed point separates relevant couplings that grow from irrelevant ones that shrink, which is why only dimension and symmetry survive to set the exponents. The one-dimensional Ising decimation carries the whole scheme through in closed form and reproduces the absence of a finite-temperature transition.\n",{"path":27360,"title":27361,"module":27362,"summary":27363},"\u002Fstatistical-mechanics\u002Ffluctuations\u002Fthermodynamic-fluctuations-and-response","Thermodynamic Fluctuations and Response Functions","Fluctuations and Response","Thermodynamic variables are sharp only on average; a macroscopic system in equilibrium fluctuates about its mean values. Einstein inverted Boltzmann's $S=k_B\\ln\\Omega$ into a Gaussian probability for a fluctuation, $w\\propto e^{\\Delta S\u002Fk_B}$, and the second moments it predicts reproduce the response functions: $\\langle\\Delta E^2\\rangle=k_BT^2C_V$, $\\langle\\Delta V^2\\rangle=k_BTV\\kappa_T$, $\\langle\\Delta M^2\\rangle=k_BT\\chi_T$. The variances diverge where the responses diverge, at a critical point, producing critical opalescence and the breakdown of the thermodynamic description.\n",{"path":27365,"title":27366,"module":27362,"summary":27367},"\u002Fstatistical-mechanics\u002Ffluctuations\u002Fbrownian-motion-and-the-langevin-equation","Brownian Motion and the Langevin Equation","A pollen grain in water executes a random walk driven by molecular collisions. Einstein tied its diffusion constant to its mobility, $D=\\mu_{\\mathrm{mob}}k_BT$, turning a visible motion into a measurement of Avogadro's number. The Langevin equation splits the collisions into a systematic drag and a random force whose strength is fixed by the drag through $\\langle\\xi(t)\\xi(t')\\rangle=2\\gamma k_BT\\,\\delta(t-t')$ — the first fluctuation–dissipation relation. The mean-square displacement grows ballistically at short times and linearly, $\\langle r^2\\rangle=2dDt$, at long times, and the Stokes–Einstein relation $D=k_BT\u002F6\\pi\\eta a$ closes the loop to Perrin's experiments.\n",{"path":27369,"title":27370,"module":27362,"summary":27371},"\u002Fstatistical-mechanics\u002Ffluctuations\u002Flinear-response-and-the-fluctuation-dissipation-theorem","Linear Response and the Fluctuation-Dissipation Theorem","A system driven by a weak external field responds through a generalized susceptibility $\\chi(\\omega)$ whose imaginary part measures dissipation. The Wiener–Khinchin theorem makes the power spectrum of equilibrium fluctuations the Fourier transform of their correlation function, and the fluctuation–dissipation theorem ties the two together: $S_x(\\omega)=(2k_BT\u002F\\omega)\\,\\chi''(\\omega)$, so the spectrum of spontaneous fluctuations is fixed by the dissipative response. The Johnson–Nyquist noise of a resistor, $\\langle V^2\\rangle=4k_BTR\\,\\Delta f$, is the canonical example, and Onsager reciprocity closes the subject.\n",{"path":27373,"title":27374,"module":6,"summary":6},"\u002Fstatistical-mechanics","Statistical Mechanics",{"path":27376,"title":27377,"module":27378,"summary":27379},"\u002Fcondensed-matter\u002Fmolecules-and-bonding\u002Fbonding-mechanisms","Bonding Mechanisms","Molecules and Chemical Bonding","A molecule forms when the total energy of two atoms drops below the energy of the separated pair. This lesson works through the four mechanisms that produce that minimum: the ionic bond from charge transfer, the covalent bond from shared electron wave functions, the metallic bond, and the weak dipole-dipole and hydrogen bonds, computing bond lengths and dissociation energies for NaCl, H₂, and H₂⁺.\n",{"path":27381,"title":27382,"module":27378,"summary":27383},"\u002Fcondensed-matter\u002Fmolecules-and-bonding\u002Fmolecular-orbitals-and-h2-plus","The Molecular-Orbital Method and H₂⁺","The hydrogen molecule ion is the two-center problem that fixes the language of chemical bonding. This lesson builds the molecular orbital as a linear combination of atomic orbitals, minimizes the energy through the variational secular equation, and reduces the result to three two-center integrals: the overlap, the Coulomb term, and the exchange (resonance) integral. The bonding and antibonding levels, their potential-energy curves, and the charge piled between the nuclei follow from those integrals.\n",{"path":27385,"title":27386,"module":27378,"summary":27387},"\u002Fcondensed-matter\u002Fmolecules-and-bonding\u002Fhydrogen-molecule-and-exchange","The Hydrogen Molecule, Exchange, and Hybridization","Adding the second electron turns the one-electron ion into the two-electron hydrogen molecule, where electron-electron repulsion and the Pauli principle govern the bond. This lesson contrasts the Heitler-London valence-bond and molecular-orbital wave functions, derives the singlet-triplet splitting as an exchange energy, shows why naive molecular orbitals fail at dissociation, and builds the sp, sp², and sp³ hybrids that fix the directed geometry of covalent bonds.\n",{"path":27389,"title":27390,"module":27378,"summary":27391},"\u002Fcondensed-matter\u002Fmolecules-and-bonding\u002Fvan-der-waals-forces","Van der Waals Forces","The bond of last resort acts between all atoms, even closed-shell noble gases with no permanent moment. This lesson separates the three van der Waals contributions — Keesom orientation, Debye induction, and London dispersion — derives the London 1\u002Fr⁶ attraction from the coupled-oscillator and second-order perturbation pictures, and assembles the Lennard-Jones potential to compute the equilibrium spacing and cohesive energy of the noble-gas crystals, argon in particular.\n",{"path":27393,"title":27394,"module":27395,"summary":27396},"\u002Fcondensed-matter\u002Fmolecular-spectra\u002Frotational-vibrational-spectra","Rotational and Vibrational Spectra of Molecules","Molecular Spectra","A diatomic molecule stores energy in three well-separated ledgers: electronic, vibrational, and rotational. Quantizing the rigid rotor gives levels spaced as ℓ(ℓ+1); quantizing the bond as a harmonic oscillator gives equally spaced vibrational levels. Their combination produces the P and R branches of an infrared absorption band, from which the bond length and force constant are read directly.\n",{"path":27398,"title":27399,"module":27395,"summary":27400},"\u002Fcondensed-matter\u002Fmolecular-spectra\u002Fanharmonicity-and-rovibrational-structure","Anharmonicity and Rovibrational Structure","The rigid rotor and harmonic oscillator are first approximations. A real bond follows the Morse potential, whose levels converge toward dissociation; a real rotor stretches centrifugally; and vibration couples to rotation, so the rotational constant depends on the vibrational level. This lesson works out the anharmonic and centrifugal corrections, the Birge-Sponer route to the dissociation energy, the isotope shift, and the thermal band envelope.\n",{"path":27402,"title":27403,"module":27395,"summary":27404},"\u002Fcondensed-matter\u002Fmolecular-spectra\u002Framan-and-electronic-bands","Raman Scattering and Electronic Bands","Not every vibration absorbs in the infrared. Raman scattering reaches modes that modulate the polarizability, giving Stokes and anti-Stokes lines whose intensity ratio measures temperature, and the mutual-exclusion rule pairs it with infrared absorption. Electronic transitions add the vibronic structure of band spectra, governed by the Franck-Condon principle, and the radiative fates of an excited state are sorted by the Jablonski diagram into fluorescence and phosphorescence.\n",{"path":27406,"title":27407,"module":27395,"summary":27408},"\u002Fcondensed-matter\u002Fmolecular-spectra\u002Flasers-and-masers","Lasers, Masers, and Stimulated Emission","Einstein's three radiative processes — absorption, spontaneous emission, and stimulated emission — and the coefficients that relate them. Stimulated emission produces coherent photons, and inverting the level populations turns it into net amplification. We build the ruby three-level laser and the helium-neon four-level laser, and show why the fourth level makes inversion easy.\n",{"path":27410,"title":27411,"module":27412,"summary":27413},"\u002Fcondensed-matter\u002Fcrystal-structure\u002Fstructure-of-solids","The Structure of Solids","Crystal Structure","A crystal is a unit cell repeated in three dimensions. We classify the common cubic lattices, compute the Coulomb energy of an ionic crystal through the Madelung constant, and show how the divergent naive lattice sum is tamed by cubic shells. The cohesive energy that results predicts melting points and connects the diatomic bond of an earlier lesson to the bulk solid.\n",{"path":27415,"title":27416,"module":27412,"summary":27417},"\u002Fcondensed-matter\u002Fcrystal-structure\u002Fbravais-lattices-and-crystal-systems","Bravais Lattices, Bases, and Crystal Structures","A crystal is a Bravais lattice decorated by a basis. This lesson separates the two, builds primitive and Wigner-Seitz cells, enumerates the seven crystal systems and fourteen Bravais lattices, and fixes the language of point and space groups. Miller indices label planes and directions, and the packing fractions of the close-packed, cubic, and diamond structures follow from the geometry.\n",{"path":27419,"title":27420,"module":27412,"summary":27421},"\u002Fcondensed-matter\u002Fcrystal-structure\u002Freciprocal-lattice-and-brillouin-zones","The Reciprocal Lattice and Brillouin Zones","Every periodic crystal has a dual lattice in wavevector space. This lesson defines the reciprocal lattice through the condition b_i dot a_j equals two pi delta, derives its properties, shows the reciprocal of fcc is bcc, links reciprocal vectors to families of lattice planes, and builds the first Brillouin zone as the Wigner-Seitz cell of the reciprocal lattice, including the higher zones.\n",{"path":27423,"title":27424,"module":27412,"summary":27425},"\u002Fcondensed-matter\u002Fcrystal-structure\u002Fdiffraction-and-structure-factors","X-ray and Neutron Diffraction","A crystal diffracts radiation whose wavelength matches its atomic spacing. This lesson derives the Bragg condition, the equivalent Laue condition 2k dot G equals G squared, and the Ewald-sphere construction, then computes the geometric structure factor that produces systematic absences for bcc and fcc, the atomic form factor, and the powder method. It closes on why neutrons and electrons complement X-rays.\n",{"path":27427,"title":27428,"module":27429,"summary":27430},"\u002Fcondensed-matter\u002Flattice-dynamics\u002Fphonon-dispersion","The Harmonic Crystal and Phonon Dispersion","Lattice Dynamics","Atoms in a crystal vibrate about their equilibrium sites, and expanding the potential to second order turns the whole lattice into a set of coupled harmonic oscillators. This lesson sets up the harmonic approximation and the dynamical matrix, solves the monatomic linear chain for its dispersion omega(k) = 2 sqrt(K\u002FM) times the absolute sine of ka over two, explains why wavevectors outside the first Brillouin zone are redundant, and extends the chain to two atoms per cell to produce acoustic and optical branches with a frequency gap.\n",{"path":27432,"title":27433,"module":27429,"summary":27434},"\u002Fcondensed-matter\u002Flattice-dynamics\u002Fphonons-quantization-and-dos","Phonons, Density of States, and Crystal Momentum","Quantizing the normal modes of a harmonic crystal turns each vibrational mode into a quantum oscillator whose excitations are phonons. This lesson counts phonons with Bose-Einstein statistics, defines crystal momentum and the normal versus Umklapp distinction in momentum conservation, builds the density of states with its van Hove singularities, and shows how inelastic neutron scattering measures a dispersion curve point by point.\n",{"path":27436,"title":27437,"module":27429,"summary":27438},"\u002Fcondensed-matter\u002Flattice-dynamics\u002Fdebye-einstein-heat-capacity","Thermal Properties — Einstein and Debye Models","The lattice heat capacity follows the classical Dulong-Petit value at high temperature but collapses toward zero as T approaches zero, a purely quantum effect. This lesson derives that behavior from the Einstein model of a single frequency, then the Debye model of a linear phonon spectrum with a cutoff, obtaining the Debye T-cubed law at low temperature and the Debye interpolation across all temperatures, and closes with thermal expansion and the Gruneisen parameter.\n",{"path":27440,"title":27441,"module":27429,"summary":27442},"\u002Fcondensed-matter\u002Flattice-dynamics\u002Fanharmonicity-and-thermal-transport","Anharmonicity, Thermal Expansion, and Heat Conduction","A perfectly harmonic crystal neither expands when heated nor resists heat flow. Both effects come from the cubic and higher terms the harmonic approximation discards. This lesson derives thermal expansion from an asymmetric interatomic potential, treats phonon-phonon scattering as the decay channel these terms open, shows why Umklapp processes are what make lattice thermal conductivity finite, and traces the temperature dependence of the conductivity and the phonon mean free path.\n",{"path":27444,"title":27445,"module":27446,"summary":27447},"\u002Fcondensed-matter\u002Ffree-electron-fermi-gas\u002Ffree-electron-gas-and-conduction","Conduction and the Free-Electron Gas","Free-Electron Fermi Gas","Drude's classical free-electron model gets Ohm's law right but the resistivity, its temperature dependence, and the heat capacity wrong. Replacing the Maxwell-Boltzmann distribution with the Fermi-Dirac distribution and treating electron-lattice collisions as wave scattering repairs all three: the Fermi energy, Fermi speed, and a mean free path set by thermal lattice vibrations.\n",{"path":27449,"title":27450,"module":27446,"summary":27451},"\u002Fcondensed-matter\u002Ffree-electron-fermi-gas\u002Fsommerfeld-model-and-heat-capacity","The Sommerfeld Model: Ground State and Heat Capacity","Quantizing the free-electron gas in a box fills a Fermi sphere in k-space. The density of states grows as the square root of energy in three dimensions, and the Fermi energy, temperature, and wavevector follow for real metals. The Sommerfeld expansion shows only a thermal shell of width k_BT near E_F is excited, giving an electronic heat capacity linear in T that sits beneath the phonon T-cubed term.\n",{"path":27453,"title":27454,"module":27446,"summary":27455},"\u002Fcondensed-matter\u002Ffree-electron-fermi-gas\u002Ftransport-and-the-hall-effect","Transport, Wiedemann–Franz, and the Hall Effect","The relaxation-time picture displaces the Fermi sphere under an applied field and gives the electrical conductivity ne-squared-tau over m. The same electrons carry heat, and their ratio yields the Wiedemann–Franz law with the universal Lorenz number. A magnetic field bends the carriers into cyclotron orbits and produces the Hall voltage, whose sign reveals the charge of the carriers.\n",{"path":27457,"title":27458,"module":27446,"summary":27459},"\u002Fcondensed-matter\u002Ffree-electron-fermi-gas\u002Fscreening-and-plasmons","Screening, Plasmons, and the Limits of Free Electrons","A mobile electron gas rearranges to screen any foreign charge, turning the bare Coulomb potential into a short-ranged Yukawa form over the Thomas–Fermi length. Displaced collectively, the gas rings at the plasma frequency, whose quantum is the plasmon and whose value sets the reflectivity edge of metals. A ledger of free-electron successes and failures then motivates band theory.\n",{"path":27461,"title":27462,"module":27463,"summary":27464},"\u002Fcondensed-matter\u002Fband-theory\u002Fblochs-theorem-and-energy-bands","Bloch's Theorem and Energy Bands","Band Theory","An electron in a periodic potential has stationary states that are plane waves modulated by a lattice-periodic envelope. This lesson proves Bloch's theorem two ways, defines crystal momentum and the band index, counts the allowed wavevectors from Born–von Kármán boundary conditions, and sets up the extended, reduced, and repeated-zone descriptions of a band.\n",{"path":27466,"title":27467,"module":27463,"summary":27468},"\u002Fcondensed-matter\u002Fband-theory\u002Fnearly-free-electron-model","The Nearly-Free-Electron Model","A weak periodic potential leaves the free-electron parabola almost intact except near Brillouin-zone boundaries, where two nearly degenerate plane waves mix. This lesson solves the resulting two-by-two secular problem, shows the gap of size twice the potential component opening at each boundary, identifies the two standing waves that pile charge on and between the ions, and works the exactly solvable Kronig–Penney model.\n",{"path":27470,"title":27471,"module":27463,"summary":27472},"\u002Fcondensed-matter\u002Fband-theory\u002Ftight-binding-method","The Tight-Binding Method","The opposite limit to nearly-free electrons builds bands from atomic orbitals. A Bloch sum of one orbital per site gives a dispersion set by the hopping integral between neighbours; the band widens from a sharp atomic level as the atoms approach. This lesson derives the s-band cosine dispersion, extends it to p-bands, and introduces Wannier functions as the localized dual of Bloch states.\n",{"path":27474,"title":27475,"module":27463,"summary":27476},"\u002Fcondensed-matter\u002Fband-theory\u002Ffermi-surfaces-and-semiclassical-dynamics","Fermi Surfaces, Effective Mass, and Metals vs Insulators","Filling the bands settles which crystals conduct. A filled band carries no current, so a crystal with filled bands and a gap is an insulator, while a partly filled band makes a metal. This lesson derives the no-current theorem for a filled band, defines the Fermi surface and Harrison's construction, introduces holes and the effective mass from band curvature, and states the semiclassical equations of motion that lead to Bloch oscillations.\n",{"path":27478,"title":27479,"module":27480,"summary":27481},"\u002Fcondensed-matter\u002Fsemiconductors\u002Fsemiconductor-bands-and-junctions","Band Theory and Semiconductors","Semiconductors","The periodic lattice splits atomic levels into allowed energy bands separated by forbidden gaps. Whether the highest occupied band is full or partly full, and how wide the gap above it is, sorts every solid into conductor, insulator, or semiconductor. Doping adds donor or acceptor levels inside the gap, and a p-n junction built from doped regions gives the diode, the solar cell, the LED, and the transistor.\n",{"path":27483,"title":27484,"module":27480,"summary":27485},"\u002Fcondensed-matter\u002Fsemiconductors\u002Fintrinsic-and-extrinsic-semiconductors","Carrier Statistics: Intrinsic and Extrinsic Semiconductors","The number of mobile electrons and holes in a semiconductor follows from the density of states near each band edge and the Fermi-Dirac tail that reaches into it. This lesson derives the effective densities of states, the intrinsic concentration and its exponential gap dependence, the law of mass action, the temperature march of the Fermi level, and the freeze-out, saturation, and intrinsic regimes of a doped crystal.\n",{"path":27487,"title":27488,"module":27480,"summary":27489},"\u002Fcondensed-matter\u002Fsemiconductors\u002Fcarrier-transport-and-recombination","Carrier Transport and Recombination","Carriers move by drift in a field and by diffusion down a concentration gradient, the two tied together by the Einstein relation. This lesson derives mobility and its scattering-limited temperature dependence, the drift and diffusion currents, the continuity equations, band-to-band and trap-assisted recombination, and the minority-carrier lifetime and diffusion length that set the length scale of every junction device.\n",{"path":27491,"title":27492,"module":27480,"summary":27493},"\u002Fcondensed-matter\u002Fsemiconductors\u002Fthe-pn-junction","The p-n Junction in Depth","Joining p-type and n-type silicon aligns their Fermi levels and leaves a depletion region of fixed charge with a built-in potential. This lesson derives the space-charge field and potential from Poisson's equation in the depletion approximation, the built-in voltage from Fermi-level alignment, the Shockley diode equation from minority-carrier diffusion, junction and diffusion capacitance, and the avalanche and Zener breakdown mechanisms.\n",{"path":27495,"title":27496,"module":27480,"summary":27497},"\u002Fcondensed-matter\u002Fsemiconductors\u002Ftransistors-and-optoelectronics","Transistors and Optoelectronic Devices","Two junctions in series make a bipolar transistor whose thin base gives current gain; a gate over an oxide makes a MOSFET whose inversion channel switches digital logic. Run in reverse, a junction converts photons to current. This lesson derives the transistor current gain and the MOSFET channel current, then treats the LED, the diode laser, and the illuminated solar-cell characteristic.\n",{"path":27499,"title":27500,"module":27501,"summary":27502},"\u002Fcondensed-matter\u002Fdielectrics-and-ferroelectrics\u002Fdielectrics-and-polarization","Dielectrics, Polarization, and the Local Field","Dielectrics and Ferroelectrics","An insulator responds to an electric field by polarizing. This lesson builds the macroscopic polarization and the dielectric constant, sorts the microscopic response into electronic, ionic, and orientational polarizability, and corrects the field an atom actually feels to the Lorentz local field E + P\u002F3 epsilon-0. The Clausius-Mossotti relation links the measured permittivity to the atomic polarizability, and the frequency dependence of each mechanism explains why the static and optical dielectric constants differ.\n",{"path":27504,"title":27505,"module":27501,"summary":27506},"\u002Fcondensed-matter\u002Fdielectrics-and-ferroelectrics\u002Fferroelectrics-and-piezoelectrics","Ferroelectrics, Piezoelectrics, and Structural Transitions","Some crystals carry a polarization with no applied field and switch it under a reversing field, tracing a hysteresis loop. This lesson develops the ferroelectric transition through the perovskite BaTiO3 displacive instability and its soft transverse-optical mode, builds the Landau free-energy theory of first- and second-order polar transitions, derives the Curie-Weiss divergence of the dielectric constant, and closes with piezoelectricity and pyroelectricity and their devices.\n",{"path":27508,"title":27509,"module":27510,"summary":27511},"\u002Fcondensed-matter\u002Fmagnetism\u002Fdiamagnetism-and-paramagnetism","Diamagnetism and Paramagnetism","Magnetism in Solids","Every solid responds to a magnetic field. Filled shells give a small negative diamagnetic susceptibility from induced Larmor currents; localized moments give a positive Curie paramagnetism described by the Brillouin function, with the ground-state moment fixed by Hund's rules. The conduction electrons add a temperature-independent Pauli paramagnetism from the thermal shell near the Fermi surface, partly cancelled by Landau diamagnetism of their orbital motion.\n",{"path":27513,"title":27514,"module":27510,"summary":27515},"\u002Fcondensed-matter\u002Fmagnetism\u002Fexchange-and-ferromagnetism","Exchange and Ferromagnetism","Magnetic ordering at hundreds of kelvin cannot be dipolar; it is an exchange effect, the Coulomb repulsion sorted by the Pauli principle into a spin-dependent energy captured by the Heisenberg Hamiltonian. Weiss molecular-field theory replaces the exchange field by an average proportional to the magnetization, giving a self-consistent equation whose solution is spontaneous magnetization below a Curie temperature and a Curie–Weiss susceptibility above it. Itinerant ferromagnetism follows from the Stoner criterion on the band density of states.\n",{"path":27517,"title":27518,"module":27510,"summary":27519},"\u002Fcondensed-matter\u002Fmagnetism\u002Fantiferromagnetism-and-domains","Antiferromagnetism, Ferrimagnetism, and Domains","A negative exchange coupling orders neighboring spins antiparallel. Two-sublattice molecular-field theory gives a Néel temperature marked by a cusp in the susceptibility, and unequal sublattices leave a net moment — ferrimagnetism, the magnetism of the ferrites. A ferromagnet breaks into domains to reduce its magnetostatic energy, separated by Bloch walls whose width is set by the competition between exchange and magnetocrystalline anisotropy, and the irreversible motion of those walls produces the hysteresis loop.\n",{"path":27521,"title":27522,"module":27510,"summary":27523},"\u002Fcondensed-matter\u002Fmagnetism\u002Fspin-waves-and-magnons","Spin Waves and Magnons","The lowest excitations of a ferromagnet are not single flipped spins but collective precessions in which every moment tips slightly and its phase advances along the crystal. These spin waves have a quadratic dispersion at long wavelength, quantize into magnons obeying Bose statistics, and their thermal population removes magnetization as the Bloch T-to-the-three-halves law. Antiferromagnetic magnons disperse linearly, and inelastic neutron scattering measures both.\n",{"path":27525,"title":27526,"module":27527,"summary":27528},"\u002Fcondensed-matter\u002Fsuperconductivity\u002Fsuperconductivity-phenomenology","Superconductivity: Phenomenology and BCS","Superconductivity","Below a critical temperature some materials lose all resistance and expel magnetic flux — the Meissner effect that defines the state. The isotope effect points to lattice vibrations, and BCS theory binds electrons into Cooper pairs through phonon exchange. The paired condensate opens an energy gap, quantizes magnetic flux, and drives the Josephson effects.\n",{"path":27530,"title":27531,"module":27527,"summary":27532},"\u002Fcondensed-matter\u002Fsuperconductivity\u002Flondon-theory-and-the-meissner-effect","London Theory and the Meissner Effect","A perfect conductor freezes the field it was cooled in; a superconductor expels it. The distinction needs a constitutive law beyond zero resistance — the two London equations — whose solution is exponential flux decay over the penetration depth. The same rigidity follows from a macroscopic condensate wave function, and the thermodynamics of the critical field fixes the condensation energy, the latent heat, and the specific-heat jump.\n",{"path":27534,"title":27535,"module":27527,"summary":27536},"\u002Fcondensed-matter\u002Fsuperconductivity\u002Fginzburg-landau-theory","Ginzburg–Landau Theory, Vortices, and Type-II","A complex order parameter and a free-energy expansion turn the superconducting transition into a Landau theory. Two lengths emerge — the coherence length and the penetration depth — whose ratio kappa sorts superconductors into type I and type II. Type-II materials admit flux as an Abrikosov lattice of vortices, each threading exactly one quantum h\u002F2e, between a lower and an upper critical field.\n",{"path":27538,"title":27539,"module":27527,"summary":27540},"\u002Fcondensed-matter\u002Fsuperconductivity\u002Fbcs-theory","Microscopic BCS Theory","A phonon-mediated attraction, however weak, binds two electrons above the Fermi sea — the Cooper problem shows the sea is unstable. The BCS variational ground state pairs all electrons near the Fermi surface and, through a self-consistent gap equation, opens an energy gap. Weak-coupling solution gives the exponential T_c and the universal ratios 2 Delta(0) = 3.53 k_B T_c and Delta C \u002F C_n = 1.43.\n",{"path":27542,"title":27543,"module":27527,"summary":27544},"\u002Fcondensed-matter\u002Fsuperconductivity\u002Fjosephson-and-high-tc","Josephson Effects and Unconventional Superconductors","Two superconductors joined by a thin barrier carry a supercurrent set by their phase difference — the dc Josephson effect — and oscillate at 2eV\u002Fh under a voltage. A two-junction loop turns flux quantization into a magnetometer of single-quantum sensitivity. The cuprates superconduct in CuO2 planes with a doping-dependent dome, d-wave pairing, and a pseudogap that lie outside the phonon picture.\n",{"path":27546,"title":27547,"module":27548,"summary":27549},"\u002Fcondensed-matter\u002Fnanostructures\u002Fquantum-wells-wires-and-dots","Quantum Wells, Wires, and Dots","Nanostructures","When a crystal is shrunk until one or more of its dimensions approaches the electron wavelength, the continuous bands of the bulk break into discrete subbands. Confining in one direction gives a quantum well with a step-like density of states, in two directions a quantum wire with inverse-square-root singularities, and in all three a quantum dot whose levels are sharp like an atom's. This lesson derives the density of states in each case and applies it to size-tunable dot emission and the Coulomb blockade of a single-electron transistor.\n",{"path":27551,"title":27552,"module":27548,"summary":27553},"\u002Fcondensed-matter\u002Fnanostructures\u002Finteger-quantum-hall-effect","The 2D Electron Gas and the Integer Quantum Hall Effect","A two-dimensional electron gas in a strong perpendicular magnetic field has its continuous density of states collapse into macroscopically degenerate Landau levels. As the field is swept, the Hall resistance locks onto exact plateaus at h over an integer times e squared, while the longitudinal resistance drops to zero. This lesson derives the Landau levels and their degeneracy, explains the plateaus through disorder-localized states and current-carrying edge channels, and states why the von Klitzing constant is now a resistance standard.\n",{"path":27555,"title":27556,"module":27548,"summary":27557},"\u002Fcondensed-matter\u002Fnanostructures\u002Ffractional-quantum-hall-and-topology","The Fractional Quantum Hall Effect and Topological Order","When the lowest Landau level is only partly filled, the non-interacting theory predicts no gap, yet a plateau appears at filling one-third. It is a many-body effect: Coulomb repulsion selects a correlated ground state, the Laughlin wavefunction, whose excitations carry a fraction of the electron charge. This lesson builds the Laughlin state, introduces composite fermions that map the fractional effect onto an integer one, and explains how the quantum Hall effect brought the Chern number and topology into condensed-matter physics.\n",{"path":27559,"title":27560,"module":27548,"summary":27561},"\u002Fcondensed-matter\u002Fnanostructures\u002Fgraphene-and-dirac-materials","Graphene and Dirac Materials","Graphene is one atomic layer of carbon on a honeycomb lattice. A tight-binding calculation on its two-atom basis gives valence and conduction bands that touch at the corners of the Brillouin zone, where the dispersion is linear and the electrons behave as massless two-dimensional Dirac particles with a fixed speed. This lesson derives the Dirac cones, the Berry phase of pi and the sublattice chirality, the anomalous half-integer quantum Hall effect that follows, and how opening a gap in a Dirac cone points toward topological insulators.\n",{"path":27563,"title":27564,"module":6,"summary":6},"\u002Fcondensed-matter","Condensed Matter Physics",{"path":27566,"title":27567,"module":24340,"summary":27568},"\u002Flogic\u002Ffoundations\u002Flogic-as-a-mathematical-model","Logic as a Mathematical Model of Deduction","Symbolic logic models deductive reasoning the way probability theory models chance: it keeps the form of a correct deduction and discards its content. A deduction is valid when its conclusion follows from the form of the premises alone, independent of what the non-logical words mean. Two models carry the subject — coarse sentential logic and fine first-order logic — and four questions organize it: logical consequence, methods of proof, the gap between provable and true, and the link between logic and computability. Tuples, relations, functions, equivalence classes, and cardinality supply the set-theoretic vocabulary every later chapter uses.\n",{"path":27570,"title":27571,"module":27572,"summary":27573},"\u002Flogic\u002Fsentential-logic\u002Fformal-languages-and-well-formed-formulas","Formal Languages and Well-Formed Formulas","Sentential Logic","The language of sentential logic has an alphabet of sentence symbols, five connectives, and two parentheses, with formation rules that pick out the well-formed formulas. The wffs are the least set of expressions closed under the five formula-building operations, and every such generated set carries an induction principle.\n",{"path":27575,"title":27576,"module":27572,"summary":27577},"\u002Flogic\u002Fsentential-logic\u002Ftruth-assignments-and-tautologies","Truth Assignments, Tautologies, and Consequence","A truth assignment fixes the sentence symbols true or false, and a recursion extends it uniquely to every formula. Satisfaction, tautologies, and tautological implication — one formula following semantically from others — rest on that extension, and the truth-table procedure decides implication for finite premise sets.\n",{"path":27579,"title":27580,"module":27572,"summary":27581},"\u002Flogic\u002Fsentential-logic\u002Funique-readability-and-parsing","Unique Readability and a Parsing Algorithm","Parentheses keep a formula from being read two ways. The parenthesis lemmas and a top-down parsing algorithm recover a formula's structure and yield unique readability: every wff has exactly one formation tree, which is what makes the truth recursion well defined.\n",{"path":27583,"title":27584,"module":27572,"summary":27585},"\u002Flogic\u002Fsentential-logic\u002Finduction-and-recursion","Induction and Recursion on Formulas","Two principles govern any set generated from initial elements by operations: prove a property of all its members by covering the initial elements and the closure steps, and define a function on it by recursion on structure. The recursion theorem needs the set to be freely generated, and unique readability supplies that condition for the well-formed formulas.\n",{"path":27587,"title":27588,"module":27572,"summary":27589},"\u002Flogic\u002Fsentential-logic\u002Fexpressive-completeness-and-normal-forms","Sentential Connectives and Normal Forms","Every formula computes a Boolean function of its atoms, and Post's theorem gives the converse: every Boolean function is realized by a wff in disjunctive normal form, so the five connectives are more than enough. Minimal complete sets follow, down to the single connectives NAND and NOR, together with a method for proving a set of connectives incomplete.\n",{"path":27591,"title":27592,"module":27572,"summary":27593},"\u002Flogic\u002Fsentential-logic\u002Fboolean-circuits","Switching Circuits","A memoryless two-valued circuit computes a Boolean function, so every formula names a gate network and every network a formula. Cost and delay are read off the formula by recursion, and tautological equivalence and normal forms design and simplify circuits realizing a given specification.\n",{"path":27595,"title":27596,"module":27572,"summary":27597},"\u002Flogic\u002Fsentential-logic\u002Fcompactness-and-effectiveness","Compactness and Effectiveness","The compactness theorem reduces satisfiability of an infinite set of formulas to its finite subsets, proved by extension to a maximal finitely satisfiable set and applied to color infinite graphs. Effectiveness fixes what \"decidable\" and \"effectively enumerable\" mean and settles the decidability of tautologyhood.\n",{"path":27599,"title":27600,"module":27601,"summary":27602},"\u002Flogic\u002Ffirst-order-languages\u002Ffirst-order-languages","First-Order Languages","First-Order Languages and Structures","Sentential logic cannot see inside a simple statement, so it misses valid arguments that turn on quantifiers and predicates. A first-order language adds a quantifier, variables, and a chosen vocabulary of predicate, function, and constant symbols. Terms and well-formed formulas are built by recursion over this alphabet, and a variable occurs free or bound according to the quantifiers that reach it.\n",{"path":27604,"title":27605,"module":27601,"summary":27606},"\u002Flogic\u002Ffirst-order-languages\u002Fstructures-truth-and-satisfaction","Structures, Truth, and Satisfaction","A structure interprets a language: a nonempty universe plus a meaning for every predicate, function, and constant symbol. Tarski's recursion defines when a structure satisfies a formula under a variable assignment, and hence when a sentence is true. From satisfaction we recover logical implication, validity, and logical equivalence for first-order logic.\n",{"path":27608,"title":27609,"module":27601,"summary":27610},"\u002Flogic\u002Ffirst-order-languages\u002Fdefinability-and-elementary-equivalence","Definability and Elementary Equivalence","Fix a structure and ask which relations a formula can pick out: the definable ones. A set of sentences picks out a class of structures, the elementary classes. Homomorphisms and isomorphisms compare structures, and the homomorphism theorem shows isomorphic structures satisfy the same sentences. Automorphisms bound what first-order logic can distinguish, giving a tool for proving relations undefinable.\n",{"path":27612,"title":27613,"module":27601,"summary":27614},"\u002Flogic\u002Ffirst-order-languages\u002Fterms-substitution-and-parsing","Parsing, Substitution, and Substitutability","Every recursion on first-order syntax rests on unique readability. A parenthesis-counting function proves that terms and formulas decompose in exactly one way, and a parsing algorithm recovers the decomposition. Substituting a term for a free variable can capture it under a quantifier; the substitutability condition rules that out, and the substitution lemma trades syntactic substitution for a change of assignment.\n",{"path":27616,"title":27617,"module":27618,"summary":27619},"\u002Flogic\u002Fdeductive-calculus\u002Fa-deductive-calculus","A Deductive Calculus for First-Order Logic","The Deductive Calculus and Its Metatheorems","A proof must be finite and mechanically checkable. A Hilbert-style calculus meets both demands: six schemas of logical axioms, a single rule of inference (modus ponens), and the syntactic consequence relation they generate. Substitution and substitutability are defined by recursion, and the bridge theorem reduces deducibility to tautological implication from the axioms.\n",{"path":27621,"title":27622,"module":27618,"summary":27623},"\u002Flogic\u002Fdeductive-calculus\u002Fdeduction-theorem-and-derived-rules","The Deduction Theorem and Derived Rules","Raw deductions from axioms are unusable by hand. The generalization theorem, the deduction theorem, contraposition, reductio ad absurdum, and rule T reduce the calculus to the moves of ordinary mathematics, each proved once to license a block of axiom-level steps. Generalization on constants and alphabetic variants handle the quantifier and substitution bookkeeping.\n",{"path":27625,"title":27626,"module":27618,"summary":27627},"\u002Flogic\u002Fdeductive-calculus\u002Fsoundness","The Soundness Theorem","Soundness is the easy half of the match between proof and truth. Whatever the calculus deduces is logically implied, by an induction on deduction length that rests on one lemma: every logical axiom is valid. The only hard case, quantifier instantiation, needs the substitution lemma. The contrapositive corollary states that every satisfiable set is consistent.\n",{"path":27629,"title":27630,"module":27618,"summary":27631},"\u002Flogic\u002Fdeductive-calculus\u002Fcompleteness-and-consistency","The Completeness Theorem","Gödel's completeness theorem is the deep converse of soundness: whatever is logically implied can be deduced. Equivalently, every consistent set has a model. The Henkin proof manufactures that model out of syntax alone: add witnessing constants, extend to a maximal consistent set, and read a term model off the formulas it contains. Compactness and the enumerability theorem drop out.\n",{"path":27633,"title":27634,"module":27635,"summary":27636},"\u002Flogic\u002Fmodels-and-theories\u002Fcompactness-and-lowenheim-skolem","Compactness and the Löwenheim–Skolem Theorems","Models, Compactness, and Theories","A set of first-order sentences has a model whenever each of its finite subsets does. This compactness theorem follows from completeness and yields the finiteness limitation, the downward and upward Löwenheim–Skolem theorems, models of every infinite cardinality, and nonstandard models of arithmetic.\n",{"path":27638,"title":27639,"module":27635,"summary":27640},"\u002Flogic\u002Fmodels-and-theories\u002Ftheories-elementary-classes-and-categoricity","Theories, Elementary Classes, and Categoricity","A theory is a set of sentences closed under logical consequence. Theories correspond to classes of models; a theory may be complete, axiomatizable, or finitely axiomatizable, and completeness together with axiomatizability yields decidability. The Łoś–Vaught test derives completeness from categoricity in a cardinal, applied to dense linear orders and to algebraically closed fields.\n",{"path":27642,"title":27643,"module":27635,"summary":27644},"\u002Flogic\u002Fmodels-and-theories\u002Finterpretations-between-theories","Interpretations Between Theories","An interpretation translates the vocabulary of one theory into formulas of another, relativizing quantifiers to a definable domain and mapping symbols to defining formulas. Defined function symbols meet a noncreativity criterion; the syntactic translation of formulas carries theoremhood forward, and a faithful interpretation transfers decidability and undecidability between theories.\n",{"path":27646,"title":27647,"module":27635,"summary":27648},"\u002Flogic\u002Fmodels-and-theories\u002Fnonstandard-analysis","Nonstandard Analysis","Compactness builds a model of the real ordered field containing infinite elements and nonzero infinitesimals. The transfer principle carries every first-order truth from the reals to this extension, the standard-part map collapses finite hyperreals back onto the reals, and continuity and the derivative are rederived by working with infinitely small quantities directly.\n",{"path":27650,"title":27651,"module":27652,"summary":27653},"\u002Flogic\u002Farithmetic-and-definability\u002Fdefinability-in-arithmetic","The Structure of Arithmetic and Definability","Number Theory and Definability","Number theory is the theory of one fixed structure, the natural numbers under successor, order, addition, multiplication, and exponentiation. Every number is named by a numeral, and a relation is definable when a single formula picks out exactly its tuples. The central gap separates the sentences true in that structure from those any reasonable set of axioms can prove.\n",{"path":27655,"title":27656,"module":27652,"summary":27657},"\u002Flogic\u002Farithmetic-and-definability\u002Fnatural-numbers-with-successor","Natural Numbers with Successor","The weakest reduct keeps only zero and successor. Its models are a standard chain together with disjoint copies of the integers, which makes the theory categorical in every uncountable power, hence complete and decidable. A quantifier-elimination procedure gives a practical decision method and shows a subset is definable if and only if it is finite or cofinite.\n",{"path":27659,"title":27660,"module":27652,"summary":27661},"\u002Flogic\u002Farithmetic-and-definability\u002Fpresburger-and-reducts","Reducts: Order, Addition, and Multiplication","Adding order to the successor reduct keeps decidability and makes the theory finitely axiomatizable; adding addition gives Presburger arithmetic, still decidable by quantifier elimination once congruence predicates are included, with definable sets exactly the eventually periodic ones. Multiplication is the break point: neither addition nor order can define it, and once it joins addition the theory stops being decidable.\n",{"path":27663,"title":27664,"module":27652,"summary":27665},"\u002Flogic\u002Farithmetic-and-definability\u002Fa-subtheory-and-representability","A Subtheory of Number Theory and Representability","A finite set of eleven axioms, the recursion equations for successor, order, addition, multiplication, and exponentiation, already proves every true quantifier-free and existential sentence. Representability asks a theory to prove the right instances of a formula rather than merely make them true, and a relation is defined to be recursive exactly when some consistent finite theory represents it. Church's thesis identifies that with decidability, and closure under composition, minimization, and primitive recursion builds the catalog the incompleteness proofs need.\n",{"path":27667,"title":27668,"module":27669,"summary":27670},"\u002Flogic\u002Fincompleteness\u002Farithmetization-of-syntax","Arithmetization of Syntax","Arithmetization and the Incompleteness Theorems","Gödel numbering assigns a natural number to every symbol, expression, formula, and deduction, turning statements about syntax into statements about numbers. The syntactic operations — substitution, \"is a wff\", \"is an axiom\", \"d codes a deduction of a\" — come out primitive recursive and hence representable in the subtheory, which lets a formula of arithmetic talk about formulas, including itself.\n",{"path":27672,"title":27673,"module":27669,"summary":27674},"\u002Flogic\u002Fincompleteness\u002Fincompleteness-and-undecidability","Incompleteness, Undecidability, and Church's Theorem","The fixed-point lemma manufactures a sentence that talks about its own Gödel number. Pointed at truth it gives Tarski's theorem — arithmetic truth is not arithmetically definable; pointed at provability it gives Gödel's first incompleteness theorem and the undecidability of the theory of the natural numbers, and, applied to validity, Church's theorem that first-order logic is undecidable. The set of theorems of a recursive theory is only recursively enumerable — the gap between provable and true.\n",{"path":27676,"title":27677,"module":27669,"summary":27678},"\u002Flogic\u002Fincompleteness\u002Fsecond-incompleteness-theorem","The Second Incompleteness Theorem","Consistency of a recursively axiomatized theory is itself an arithmetic sentence, built from a provability predicate. When the theory is strong enough to formalize its own reflection and modus ponens — the Hilbert–Bernays–Löb derivability conditions — it cannot prove that sentence unless it is inconsistent. Löb's theorem is the companion result, and set theory is the case that closes Hilbert's program.\n",{"path":27680,"title":27681,"module":27682,"summary":27683},"\u002Flogic\u002Fcomputability-and-representability\u002Frecursive-functions","Recursive Functions and Church's Thesis","Recursive Functions and Representability","The recursive functions are the formal counterpart of the effectively computable ones: built from three initial functions by composition, primitive recursion, and minimization, and equivalently the functions representable in a finitely axiomatized arithmetic. Church's thesis identifies the class with effective calculability; Kleene's normal form theorem and the unsolvable halting problem place the recursive sets strictly inside the recursively enumerable ones.\n",{"path":27685,"title":27686,"module":27682,"summary":27687},"\u002Flogic\u002Fcomputability-and-representability\u002Frepresenting-exponentiation","Representing Exponentiation and the β-Function","Coding finite sequences by prime-power exponents already assumes exponentiation, so representing exponentiation from addition and multiplication alone needs a different encoder. Gödel's β-function, built from a pairing function and the Chinese remainder theorem, reads back arbitrary finite sequences using only plus and times. This represents exponentiation in the addition-multiplication arithmetic and closes the last gap in the representability of every recursive syntactic operation.\n",{"path":27689,"title":27690,"module":27691,"summary":27692},"\u002Flogic\u002Fsecond-order-logic\u002Fsecond-order-languages","Second-Order Languages","Second-Order Logic and Beyond","Second-order logic quantifies over relations and functions, not just individuals. Second-order Peano arithmetic and the second-order theory of the reals become categorical, and finiteness is definable by a single sentence. Compactness, completeness, and the Löwenheim–Skolem theorems all fail for the standard semantics.\n",{"path":27694,"title":27695,"module":27691,"summary":27696},"\u002Flogic\u002Fsecond-order-logic\u002Fskolem-functions-and-many-sorted-logic","Skolem Functions and Many-Sorted Logic","Skolem functions replace existential quantifiers with named witnesses, putting any first-order formula into a prenex form with all existentials — now over functions — pulled to the front. The Skolemized formula is equisatisfiable with the original, which reduces satisfiability to universal sentences and, through Herbrand expansions, to sentential logic. Many-sorted logic then adds several universes at once and reduces cleanly to ordinary one-sorted logic.\n",{"path":27698,"title":27699,"module":27691,"summary":27700},"\u002Flogic\u002Fsecond-order-logic\u002Fgeneral-structures","General (Henkin) Structures","General semantics reinterprets second-order logic by letting the predicate and function quantifiers range over a designated collection of relations and functions rather than all of them. Recast as many-sorted first-order logic with comprehension axioms, general second-order logic recovers a sound and complete calculus together with compactness and Löwenheim–Skolem, giving up the categoricity of the standard semantics. The ω-models of analysis show the trade.\n",{"path":27702,"title":27703,"module":6,"summary":6},"\u002Flogic","Logic",{"path":27705,"title":27706,"module":24340,"summary":27707},"\u002Freinforcement-learning\u002Ffoundations\u002Fwhat-is-reinforcement-learning","What Is Reinforcement Learning?","Reinforcement learning is learning what to do — how to map situations to actions — so as to maximize a numerical reward signal, discovered by trial and error rather than told. We set up the agent–environment loop, separate it from supervised and unsupervised learning, name the four elements (policy, reward, value, and an optional model), and train a tic-tac-toe player with a temporal-difference value update.\n",{"path":27709,"title":27710,"module":24340,"summary":27711},"\u002Freinforcement-learning\u002Ffoundations\u002Fa-brief-history-of-rl","A Brief History of Reinforcement Learning","The origins of reinforcement learning. Three threads — trial-and-error learning from animal psychology, optimal control and dynamic programming, and temporal-difference learning — ran independently for decades and merged around 1989 into the modern field. Replacing the lookup table with a neural network then produced deep reinforcement learning: DQN, AlphaGo, AlphaZero, MuZero, and RLHF.\n",{"path":27713,"title":27714,"module":24340,"summary":27715},"\u002Freinforcement-learning\u002Ffoundations\u002Fmulti-armed-bandits","Multi-Armed Bandits","A bandit is reinforcement learning stripped to a single decision, repeated: no state, no consequences, only the tension between exploiting the arm that looks best and exploring the ones that might be better. We build up the whole toolkit — sample-average value estimates, the incremental update rule, ε-greedy, optimistic initialization, UCB, and gradient bandits — and use it to study exploration in isolation, the one problem that carries over to the full setting.\n",{"path":27717,"title":27718,"module":24340,"summary":27719},"\u002Freinforcement-learning\u002Ffoundations\u002Fbandit-exploration-algorithms","Bandit Exploration Algorithms","Better ways to explore than picking at random. Upper-confidence-bound selection explores by optimism about what it hasn't measured; gradient bandits learn action preferences by stochastic gradient ascent on reward. We then add context to get the contextual bandit, the bridge to full RL, and measure everything by regret — where UCB1 and Thompson sampling reach the logarithmic optimum that fixed-ε greedy cannot.\n",{"path":27721,"title":27722,"module":24340,"summary":27723},"\u002Freinforcement-learning\u002Ffoundations\u002Fmarkov-decision-processes","Markov Decision Processes","A Markov decision process is the formal interface between an agent and its environment: at each step the agent reads a state, chooses an action, and receives a reward and a next state. We fix that loop, the dynamics function that governs it, and the Markov property that makes the state sufficient; then turn goals into a scalar reward and rewards into a discounted return, with one notation that covers both episodic and continuing tasks.\n",{"path":27725,"title":27726,"module":24340,"summary":27727},"\u002Freinforcement-learning\u002Ffoundations\u002Fvalue-functions-and-optimality","Value Functions and Optimality","A value function scores how good a state (or state–action pair) is under a policy: the expected return from there onward. Its defining property is the Bellman equation, a self-consistency condition linking a state's value to its successors' values, which we derive from the return and the dynamics. Pushing the same idea to the best-achievable value gives the Bellman optimality equations, whose solution yields an optimal policy — and whose intractability is what the rest of the course is about.\n",{"path":27729,"title":24559,"module":27730,"summary":27731},"\u002Freinforcement-learning\u002Ftabular-methods\u002Fdynamic-programming","Tabular Solution Methods","Dynamic programming computes optimal policies when a perfect model of the MDP is given, by turning the Bellman equations into assignment statements. We build up iterative policy evaluation (the expected update), the policy improvement theorem, and the two classic algorithms that alternate them — policy iteration and value iteration — worked on the gridworld, a two-state MDP, Jack's car rental, and the gambler's problem.\n",{"path":27733,"title":27734,"module":27730,"summary":27735},"\u002Freinforcement-learning\u002Ftabular-methods\u002Fdp-async-and-gpi","Dynamic Programming: Asynchronous DP and Generalized Policy Iteration","Policy and value iteration both sweep the entire state set on every pass, which is impossible once the state space is huge. This lesson loosens the schedule: asynchronous DP updates states in any order, generalized policy iteration names the alternation of evaluation and improvement that underlies nearly every RL method, and a look at efficiency and the curse of dimensionality places DP among the alternatives. We close past Sutton & Barto with prioritized sweeping, neuro-dynamic programming, value-iteration networks, and MuZero.\n",{"path":27737,"title":27738,"module":27730,"summary":27739},"\u002Freinforcement-learning\u002Ftabular-methods\u002Fmonte-carlo-methods","Monte Carlo Methods","Monte Carlo methods learn value functions and optimal policies from complete sampled episodes, with no model of the environment: they simply average the returns that actually followed each state. We build prediction (first-visit and every-visit averaging), see why estimating action values forces the exploration question, and answer it two ways on-policy — exploring starts and epsilon-soft control. Throughout, Monte Carlo samples one whole trajectory to termination and never bootstraps.\n",{"path":27741,"title":27742,"module":27730,"summary":27743},"\u002Freinforcement-learning\u002Ftabular-methods\u002Fmonte-carlo-off-policy","Monte Carlo Methods: Off-Policy Learning","On-policy Monte Carlo can only reach the best exploring policy, not the true optimum. Off-policy methods remove that ceiling by learning about a greedy target policy from data generated by a soft behavior policy, corrected with importance sampling. We derive the importance-sampling ratio, weigh ordinary against weighted estimators on real numbers, give the incremental off-policy algorithm, sharpen it with discounting-aware sampling, and close by placing Monte Carlo on the model\u002Fbootstrap map beside DP and temporal-difference learning.\n",{"path":27745,"title":27746,"module":27730,"summary":27747},"\u002Freinforcement-learning\u002Ftabular-methods\u002Ftemporal-difference-learning","Temporal-Difference Learning","Temporal-difference learning is the one idea most central to reinforcement learning: learn a value directly from experience, like Monte Carlo, but update each guess toward the next guess before the episode ends, like dynamic programming. We derive the TD(0) prediction rule and its reward-prediction error, contrast its one-step backup with MC and DP, work the driving-home and random-walk examples, and show the batch-updating optimality that makes TD approximate the certainty-equivalence estimate.\n",{"path":27749,"title":27750,"module":27730,"summary":27751},"\u002Freinforcement-learning\u002Ftabular-methods\u002Ftd-control-sarsa-and-q-learning","TD Control: Sarsa, Q-learning, and Double Learning","With TD prediction in hand, control follows the generalized-policy-iteration pattern with TD as the evaluation step. We build Sarsa (on-policy), Q-learning (off-policy, targeting the optimal policy), and Expected Sarsa that spans the two, then confront the maximization bias every max-based method inherits and fix it with Double Q-learning. We close past Sutton & Barto, following each one-step tabular update into its deep-RL descendant — DQN, Double DQN, and Rainbow.\n",{"path":27753,"title":27754,"module":27730,"summary":27755},"\u002Freinforcement-learning\u002Ftabular-methods\u002Fn-step-bootstrapping","n-Step Bootstrapping","Monte Carlo waits for the full return; one-step TD bootstraps after a single reward. Between them lies a whole spectrum, indexed by one integer n: look ahead n real rewards, then bootstrap from the value n steps out. The n-step return unifies the previous two lessons, and — on the random walk — an intermediate n beats both extremes. We build the n-step return, the n-step TD update, the backup-diagram spectrum, and n-step Sarsa for control.\n",{"path":27757,"title":27758,"module":27730,"summary":27759},"\u002Freinforcement-learning\u002Ftabular-methods\u002Fn-step-off-policy-methods","n-Step Bootstrapping: Off-Policy Methods","Taking the n-step family off-policy raises the same importance-sampling questions Monte Carlo did, now over a window of exactly n actions. We reweight n-step returns by the policy ratio, watch the ratio product inflate variance on real numbers, then build the tree-backup algorithm that learns off-policy with no ratios at all — and finally n-step Q(sigma), one algorithm whose per-step switch recovers Sarsa, tree backup, and Expected Sarsa as special cases.\n",{"path":27761,"title":27762,"module":27730,"summary":27763},"\u002Freinforcement-learning\u002Ftabular-methods\u002Fplanning-and-learning","Planning and Learning","Planning and learning are the same operation run on two kinds of experience. A model turns states and actions into simulated transitions; planning backs up values over that simulated experience exactly as learning backs them up over real experience. We build the Dyna architecture that interleaves acting, model-learning, direct RL, and planning in one loop, trace a single Dyna-Q step by hand, and patch the architecture for when the model goes stale.\n",{"path":27765,"title":27766,"module":27730,"summary":27767},"\u002Freinforcement-learning\u002Ftabular-methods\u002Fplanning-focusing-and-decision-time","Planning: Focusing Updates and Decision-Time Search","Dyna plans by replaying remembered transitions, but sampling them uniformly wastes most of the effort. This lesson sharpens planning: prioritized sweeping works backward from states whose value just changed, expected versus sample updates weigh thoroughness against cost, and trajectory sampling and real-time DP focus updates on the states the policy actually visits. We trace Dyna forward to model-based deep RL, then turn to decision-time planning — heuristic search, rollouts, and Monte Carlo Tree Search.\n",{"path":27769,"title":27770,"module":27730,"summary":27771},"\u002Freinforcement-learning\u002Ftabular-methods\u002Fdecision-time-planning","Decision-Time Planning","Planning need not build a global policy. Decision-time planning runs a fresh lookahead every time a state arrives and returns just one action, then throws the work away. We start from real-time dynamic programming — asynchronous value iteration on the states the agent actually visits — then move through heuristic search and rollout algorithms, each a one-step policy improvement applied on the fly to the current state.\n",{"path":27773,"title":27774,"module":27730,"summary":27775},"\u002Freinforcement-learning\u002Ftabular-methods\u002Fmonte-carlo-tree-search","Monte Carlo Tree Search","Monte Carlo Tree Search is a rollout algorithm with memory: it accumulates value estimates across simulations and steers later ones toward promising branches. We work through the four steps — selection, expansion, simulation, backup — the UCT selection rule computed on real numbers, the asymmetric growing tree, and the full pseudocode. We close past Sutton & Barto with the lineage from UCT to AlphaGo, AlphaZero, and MuZero, where a learned network stands in for the leaf value and the rollout.\n",{"path":27777,"title":27778,"module":27779,"summary":27780},"\u002Freinforcement-learning\u002Fapproximation\u002Fon-policy-prediction","On-Policy Prediction with Approximation","Approximate Solution Methods","Every tabular method so far stored one number per state, which fails once the state space is large or continuous. We replace the table with a parameterized value function $\\hat v(s,\\mathbf{w})$, define the mean squared value error it should minimize under the on-policy distribution, and derive stochastic- and semi-gradient learning rules — the semi-gradient TD(0) update that bootstraps and so is not a true gradient. Linear methods make the analysis clean and give the TD fixed point; feature construction (polynomials, Fourier basis, coarse and tile coding, RBFs) supplies the vectors $\\mathbf{x}(s)$, and neural networks are the nonlinear bridge to deep RL.\n",{"path":27782,"title":27783,"module":27779,"summary":27784},"\u002Freinforcement-learning\u002Fapproximation\u002Ffeature-construction-and-nonlinear","Feature Construction and Nonlinear Approximation","Linear methods are only as good as the feature vectors $\\mathbf{x}(s)$ fed to them, and this lesson builds those vectors. Polynomials and the Fourier basis turn a state's coordinates into smooth global features; coarse coding, tile coding, and radial basis functions cover a continuous space with overlapping local receptive fields whose size sets the reach of generalization. Then we stop designing features by hand: a neural network learns the representation itself by gradient descent, trading the convergence guarantees of the linear case for expressiveness — the bridge to deep reinforcement learning.\n",{"path":27786,"title":27787,"module":27779,"summary":27788},"\u002Freinforcement-learning\u002Fapproximation\u002Fon-policy-control","On-Policy Control with Approximation","Prediction learned a value function from features; control learns to act. We carry semi-gradient methods over to action values $\\hat q(s,a,\\mathbf{w})$, giving episodic semi-gradient Sarsa and its n-step form, and solve Mountain Car by descending a cost-to-go surface. In the continuing case, function approximation makes discounting unable to affect which policy is best, so we replace it with the average-reward setting — the differential return, differential value functions, and differential semi-gradient Sarsa.\n",{"path":27790,"title":27791,"module":27779,"summary":27792},"\u002Freinforcement-learning\u002Fapproximation\u002Faverage-reward-control","Average-Reward Control for Continuing Tasks","With function approximation, discounting has no effect on a continuing task: averaged over the on-policy distribution, the discounted objective equals the average reward times a policy-independent constant, so $\\gamma$ cannot change which policy is best. This lesson replaces discounting with the average-reward setting — the long-run reward rate $r(\\pi)$, the differential return that measures each state's transient advantage over that rate, differential value functions and TD error, and differential semi-gradient Sarsa, the control method for continuing tasks that never invokes a discount factor.\n",{"path":27794,"title":27795,"module":27779,"summary":27796},"\u002Freinforcement-learning\u002Fapproximation\u002Foff-policy-and-the-deadly-triad","Off-Policy Methods and the Deadly Triad","Off-policy learning with function approximation is where the convergence guarantees of reinforcement learning fail. We extend the tabular off-policy updates to semi-gradient form with per-step importance sampling, show Baird's counterexample driving the weights to infinity, and identify the cause: the deadly triad of function approximation, bootstrapping, and off-policy training — any two are safe, all three can diverge. The divergence is not caused by sampling noise: a fully synchronous dynamic-programming update blows up just the same, which is what makes the triad a structural hazard rather than a fluke.\n",{"path":27798,"title":27799,"module":27779,"summary":27800},"\u002Freinforcement-learning\u002Fapproximation\u002Fbellman-error-and-gradient-td","Value-Function Geometry and Gradient-TD Methods","Why does the deadly triad diverge, and how do you stop it? This lesson develops the geometry that explains the failure: value functions as vectors, the projection operator onto the representable subspace, and the split between the Bellman error, the value error, and the projected Bellman error: the three objectives have different minimizers. The projected Bellman error is the learnable one, and Gradient-TD methods (GTD2, TDC) do true stochastic gradient descent on it, staying stable even off-policy at $O(d)$ cost. Emphatic TD reweights states instead, and a survey of variance-reduction techniques closes the gap between stability and usable learning.\n",{"path":27802,"title":27803,"module":27779,"summary":27804},"\u002Freinforcement-learning\u002Fapproximation\u002Feligibility-traces","Eligibility Traces","n-step methods unify TD and Monte Carlo by storing the last n feature vectors; eligibility traces do the same job with a single short-term memory vector. The λ-return averages every n-step return under a geometric weighting; the forward view looks ahead to that average, and the backward view produces nearly the same updates online through a decaying trace vector. We build the λ-return, TD(λ) with its trace, the two ways λ recovers TD(0) and Monte Carlo, a note on the exact equivalence of true online TD(λ), and Sarsa(λ) for control.\n",{"path":27806,"title":27807,"module":27779,"summary":27808},"\u002Freinforcement-learning\u002Fapproximation\u002Ftrue-online-and-sarsa-lambda","True Online TD(λ) and Sarsa(λ)","Plain TD(λ) makes the forward and backward views nearly agree; this lesson closes the gap. True online TD(λ) uses a dutch trace and a small correction term to produce exactly the same weight sequence as the online λ-return algorithm, at the same memory and only a constant factor more compute — the sharpest statement of the forward\u002Fbackward duality. The whole apparatus then lifts to control unchanged: Sarsa(λ) threads a single delayed reward back along an entire trajectory in one sweep, and the λ-weighting reappears in modern deep RL as generalized advantage estimation.\n",{"path":27810,"title":27811,"module":27779,"summary":27812},"\u002Freinforcement-learning\u002Fapproximation\u002Fpolicy-gradient-methods","Policy Gradient Methods","Every method so far learned values and read a policy off them. Policy gradient methods drop the intermediary: parameterize the policy directly and climb the performance gradient. We build the softmax-in-preferences parameterization, prove the policy gradient theorem that makes the gradient computable without the unknown state distribution, and derive REINFORCE and its variance-cutting state-value baseline — the launch point for the bootstrapping actor-critic that follows.\n",{"path":27814,"title":27815,"module":27779,"summary":27816},"\u002Freinforcement-learning\u002Fapproximation\u002Factor-critic-and-continuous-actions","Actor-Critic Methods and Continuous Actions","REINFORCE with a baseline learns a value function but never bootstraps; this lesson adds the bootstrapping critic that completes the actor-critic architecture. The critic scores each transition into a single TD error that steers both the actor's policy step and its own value step, trading a little bias for much lower variance and fully online, continuing-task learning. The policy gradient theorem carries over unchanged to the average-reward setting, a Gaussian policy handles real-valued actions with self-tuning exploration, and the natural policy gradient leads straight to TRPO, PPO, and the deep actor-critic methods that train today's agents.\n",{"path":27818,"title":27819,"module":27779,"summary":27820},"\u002Freinforcement-learning\u002Fapproximation\u002Fleast-squares-and-memory-based-methods","Least-Squares TD","Semi-gradient TD spends one cheap step per example and needs many examples; this lesson makes the opposite tradeoff. Least-Squares TD (LSTD) accumulates the matrices $\\mathbf{A}$ and $\\mathbf{b}$ and solves the TD fixed point $\\mathbf{w} = \\mathbf{A}^{-1}\\mathbf{b}$ directly, using the Sherman-Morrison identity to maintain the inverse in $O(d^2)$ — the most data-efficient linear TD method, at a quadratic cost. We work a solve by hand, weigh the quadratic cost against semi-gradient TD's cheap steps, and note that LSTD never forgets — a problem in control, where least-squares policy iteration is the natural extension.\n",{"path":27822,"title":27823,"module":27779,"summary":27824},"\u002Freinforcement-learning\u002Fapproximation\u002Fmemory-and-kernel-methods","Memory-Based and Kernel Methods","Least-squares TD spent more compute to extract more from each example; this lesson drops the parametric form entirely. Memory-based methods store training examples untouched and answer a query locally at retrieval time — nearest neighbor, weighted average, locally weighted regression — so accuracy grows with the data and effort concentrates where the agent actually goes. Kernel-based methods weight stored examples by a similarity kernel $k(s,s')$, and every linear method turns out to be a kernel method. Interest and emphasis, finally, make the on-policy weighting itself a design choice, aiming scarce approximation capacity at the states that matter.\n",{"path":27826,"title":27827,"module":27779,"summary":27828},"\u002Freinforcement-learning\u002Fapproximation\u002Foff-policy-eligibility-traces","Off-Policy Eligibility Traces","Eligibility traces meet off-policy learning and function approximation — the corner where stability gets hard. We first let the bootstrapping and discounting parameters vary with state, so a single generalized return covers episodic and continuing tasks and folds termination into the discount. Then we fold the per-decision importance ratio into the trace with a control-variate correction, and build Watkins's Q(λ) and its importance-sampling-free successor Tree-Backup(λ) — all correct in expectation, but still semi-gradient, so the deadly triad and its fixes wait for the next lesson.\n",{"path":27830,"title":27831,"module":27779,"summary":27832},"\u002Freinforcement-learning\u002Fapproximation\u002Fstable-off-policy-traces","Stable Off-Policy Methods with Traces","Off-policy traces get the expected target right, but with $\\lambda \u003C 1$ they bootstrap, so off-policy plus bootstrapping plus function approximation is the deadly triad and the weights can diverge. This lesson carries the two one-step fixes to traces: GTD(λ) and GQ(λ) add a second weight vector and a gradient correction for true gradient descent on the projected Bellman error, while Emphatic TD(λ) reweights updates through a followon trace and interest to recover the on-policy stability. It closes with the implementation reality that traces are cheap because they are sparse, and with Retrace and V-trace — the clipped-ratio descendants that make off-policy traces work at deep-RL scale.\n",{"path":27834,"title":27153,"module":27835,"summary":27836},"\u002Freinforcement-learning\u002Fdeep-rl\u002Fdeep-q-networks","Deep Reinforcement Learning","Deep Q-networks replace the linear value function with a neural network $Q(s,a;\\mathbf{w})$ and confront the fact that a nonlinear approximator, off-policy bootstrapping, and correlated online data — the deadly triad — make naive Q-learning diverge. DQN counters this empirically with two stabilizers: an experience replay buffer that decorrelates and reuses samples, and a periodically-frozen target network that fixes the bootstrap target. We derive the DQN loss and gradient, walk through the Atari convolutional architecture and its results, and then add the three refinements that define modern value-based deep RL — Double DQN, dueling networks, and prioritized experience replay.\n",{"path":27838,"title":27839,"module":27835,"summary":27840},"\u002Freinforcement-learning\u002Fdeep-rl\u002Fdqn-improvements","DQN Improvements: Double, Dueling, and Prioritized Replay","Three refinements that turn plain DQN into the standard modern value-based agent, each touching a different part of the system. Double DQN fixes the maximization bias in the target by splitting action selection from evaluation; dueling networks restructure the network around a state value and per-action advantages; prioritized replay changes which transitions are learned from. We close with Rainbow, which combines them, and the distributional view that predicts the whole return distribution rather than its mean.\n",{"path":27842,"title":27843,"module":27835,"summary":27844},"\u002Freinforcement-learning\u002Fdeep-rl\u002Factor-critic-and-ppo","Actor–Critic and GAE","Make the actor and the critic deep networks and the policy-gradient architecture becomes modern deep RL. We build the neural actor-critic, the advantage estimate that replaces the raw return, and Generalized Advantage Estimation as a λ-blend of n-step advantages, then the parallel-worker methods A3C and A2C that decorrelate on-policy data. The step-size constraints — trust regions, PPO, and the continuous-control family — follow in the next lesson.\n",{"path":27846,"title":27847,"module":27835,"summary":27848},"\u002Freinforcement-learning\u002Fdeep-rl\u002Fppo-and-continuous-control","PPO and Continuous Control","Keeping the policy-gradient step from destroying the policy, and the algorithms that result. Trust-region optimization bounds each update by a KL constraint; PPO keeps that goal but replaces the second-order machinery with a first-order clip on the probability ratio, which is why it is the modern default and the optimizer inside RLHF. We then tour the off-policy continuous-control family — DDPG, TD3, and SAC — and where actor-critic went at scale, from OpenAI Five to language-model alignment.\n",{"path":27850,"title":27851,"module":27835,"summary":27852},"\u002Freinforcement-learning\u002Fdeep-rl\u002Fcase-studies","Case Studies: Learning to Play","The game-playing systems that turned reinforcement learning from a theory into a track record: Samuel's checkers player, TD-Gammon, Watson's Daily-Double wagering, a reinforcement-learning memory controller, DQN, and AlphaGo through AlphaGo Zero. Read as a set they draw one line — a value function, learned by self-play or interaction, refined by search, carried by a deep network — that runs from a 1959 checkers program to superhuman Go.\n",{"path":27854,"title":27855,"module":27835,"summary":27856},"\u002Freinforcement-learning\u002Fdeep-rl\u002Frl-beyond-games","Reinforcement Learning Beyond Games","The same value-and-reward machinery, pointed at problems with no opponent. Web personalization as a contextual bandit and then a full MDP for life-time value; thermal soaring, where a glider learns to climb on turbulent air and reward design does most of the work; and the industrial-scale systems that carried the same design past Sutton & Barto — AlphaStar, OpenAI Five, GT Sophy, and RLHF, where the reward itself is learned from human preference.\n",{"path":27858,"title":27859,"module":27835,"summary":27860},"\u002Freinforcement-learning\u002Fdeep-rl\u002Ffrontiers","Frontiers: Beyond the Standard MDP","The standard MDP fixes three things — state, reward, and single-step actions — and this lesson loosens two of them. We generalize the value function into a general value function that predicts any signal, and use those predictions as auxiliary tasks that shape representations; we extend actions in time with the options framework; and we treat state as a construction the agent builds from a stream of observations. Reward design and the open problems follow in the next lesson.\n",{"path":27862,"title":27863,"module":27835,"summary":27864},"\u002Freinforcement-learning\u002Fdeep-rl\u002Freward-design-and-open-problems","Reward Design and Open Problems","How to design a reward signal that encodes the intended goal — sparse reward, shaping, and reward hacking — and the problems the whole tabular, approximate, and deep arc leaves unsolved. We close with how the frontiers were pushed after Sutton & Barto: auxiliary tasks, learned options, intrinsic-motivation bonuses, learned world models, and offline RL, then the two concerns of reward hacking and safety that any real-world agent must address.\n",{"path":27866,"title":27867,"module":27868,"summary":27869},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fdistributional-and-rainbow","Sharpening DQN: Improvements and the Distributional Idea","Modern Deep Reinforcement Learning","In the years after the 2015 DQN paper, a stream of focused improvements each fixed one weakness of the baseline without disturbing its frame. This lesson recaps five that keep the scalar $Q$-value — Double DQN, multi-step returns, dueling networks, prioritized replay, and NoisyNets, each changing a different slot of the same Q-learning loop — then develops the sixth, distributional RL, which changes the objective itself: learn the whole return distribution $Z(s,a)$. We build the distributional Bellman equation and the C51 categorical algorithm, projection step and all, worked end to end on real numbers. A companion lesson takes up QR-DQN, Rainbow, and the modern distributional line.\n",{"path":27871,"title":27872,"module":27868,"summary":27873},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fdistributional-and-rainbow-part-2","Distributional RL and Rainbow","A companion to the DQN improvements lesson. C51 fixed the return atoms and learned their probabilities; QR-DQN does the reverse — fix the probabilities, learn the values — which removes the projection and trains with a quantile loss. We cover why the distribution helps even when you act on the mean, then assemble Rainbow: all six improvements in one Q-learning loop, with the component ablation that shows each one's real weight. The distributional line then runs on through IQN, FQF, and Agent57, the first agent to beat the human baseline on all 57 Atari games.\n",{"path":27875,"title":27876,"module":27868,"summary":27877},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fcontinuous-control","Continuous Control: DDPG and TD3","When actions are real-valued, the $\\arg\\max_a Q(s,a)$ in Q-learning becomes an optimization problem on every step. This lesson builds the off-policy actor-critic family that sidesteps it: the deterministic policy gradient and DDPG, which replaces the max with a learned actor, and the three fixes of TD3 that counter the value overestimation DDPG inherits. A companion lesson takes up SAC's maximum-entropy objective and the methods built on this template.\n",{"path":27879,"title":27880,"module":27868,"summary":27881},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fcontinuous-control-part-2","Continuous Control: SAC and Beyond","A companion to the DDPG and TD3 lesson. Where those actors are deterministic and explore with bolted-on noise, soft actor-critic (SAC) changes the objective itself: maximize return plus the entropy of the policy, so exploration becomes intrinsic and the agent stays robust. We develop the maximum-entropy objective, the reparameterized squashed-Gaussian actor, and automatic temperature tuning, then survey the methods built on this off-policy template — distributional critics (D4PG), critic ensembles (REDQ), and control from pixels (DrQ, RAD).\n",{"path":27883,"title":27884,"module":27868,"summary":27885},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmodel-based-rl","Model-Based Deep RL: Sample Efficiency and PETS","A model turns experience into imagined planning. This lesson makes the sample-efficiency case for learning a dynamics model, works through why a learned model's errors compound over the planning horizon, and builds the most direct model-based method: PETS plans online with a probabilistic ensemble under model-predictive control, distrusting the model exactly where its members disagree. A companion lesson takes up latent world models (Dreamer) and MuZero.\n",{"path":27887,"title":27888,"module":27868,"summary":27889},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmodel-based-rl-part-2","Model-Based Deep RL: World Models, Dreamer, and MuZero","A companion to the PETS lesson. PETS plans in the environment's native state space; these methods change what the model represents. World Models and Dreamer learn a compact latent state and do almost all their learning by imagining inside it, with value gradients flowing through the differentiable dynamics. MuZero predicts neither states nor pixels — only the reward, value, and policy that MCTS reads — and plans with search against that learned model, AlphaZero without the rules. We close with MBPO, TD-MPC, and EfficientZero.\n",{"path":27891,"title":27892,"module":27868,"summary":27893},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fexploration","Exploration in Deep RL: Novelty as Reward","When the state space is enormous and reward is rare, ε-greedy amounts to a random walk that almost never reaches the first reward. This lesson scales the bandit's exploration ideas up to deep RL through the dominant approach — manufacture a reward for novelty and let the agent chase it: optimism and pseudo-counts from density models, and intrinsic motivation and curiosity (the Intrinsic Curiosity Module and Random Network Distillation). A companion lesson takes up posterior sampling, Go-Explore, and the modern methods.\n",{"path":27895,"title":27896,"module":27868,"summary":27897},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fexploration-part-2","Exploration in Deep RL: Posterior Sampling and Go-Explore","A companion to the novelty-as-reward lesson. Pseudo-counts and curiosity reward the unfamiliar after the agent stumbles into it; this lesson covers two ideas that go further. Bootstrapped DQN keeps an ensemble that approximates a posterior over value functions and explores by committing to one sampled hypothesis per episode — the deep, directed exploration ε-greedy cannot manage. Go-Explore remembers and returns to the frontier, defeating detachment and derailment to solve Montezuma's Revenge. We close with episodic memory (Never Give Up), Agent57, and model-based exploration.\n",{"path":27899,"title":27900,"module":27868,"summary":27901},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Foffline-rl","Offline RL: The Problem and Value-Based Fixes","Offline reinforcement learning learns a policy from a fixed logged dataset with no further environment interaction — off-policy learning pushed to the extreme, and it breaks for the extreme version of the same reason. Bootstrapping queries the value function at out-of-distribution actions the data never covers, those errors are optimistic, and with no online feedback to correct them they compound through the Bellman backup. This lesson sets up the failure and off-policy evaluation, then builds the first two families of pessimistic fixes: policy constraint (BCQ) and conservative value estimation (CQL). A companion lesson takes up implicit methods, model-based offline RL, and Decision Transformer.\n",{"path":27903,"title":27904,"module":27868,"summary":27905},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Foffline-rl-part-2","Offline RL: Implicit Methods, Sequence Models, and Beyond","A companion to the offline-RL problem lesson. Policy constraint and conservative value estimation both still query a learned value function; implicit methods (IQL) avoid querying it off the data at all, using an in-sample expectile backup. We then build pessimism into a learned model (MOPO, COMBO) and drop bootstrapping entirely with Decision Transformer's return-conditioned sequence modeling, closing with offline-to-online fine-tuning, diffusion planners, and the offline view of RLHF. The one rule throughout: without online correction, be pessimistic about what you cannot verify.\n",{"path":27907,"title":27908,"module":27868,"summary":27909},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fimitation-and-inverse-rl","Imitation Learning: Cloning, DAgger, and Inverse RL","When a reward is hard to specify but an expert is easy to watch, learn from demonstrations instead. Behavioral cloning treats control as supervised learning of the expert's state-to-action map, and fails through compounding error: small mistakes carry the agent off the expert's distribution, where it was never trained. DAgger fixes the mismatch by querying the expert on the learner's own states. Inverse RL instead recovers the reward the expert seems to optimize — an ill-posed problem that maximum-entropy IRL disambiguates. A companion lesson casts imitation as adversarial occupancy matching (GAIL, AIRL).\n",{"path":27911,"title":27912,"module":27868,"summary":27913},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fimitation-and-inverse-rl-part-2","Imitation as Adversarial Matching: GAIL and AIRL","A companion to the imitation-learning lesson. If the point of recovering a reward is only to re-run RL and match the expert, you can skip the reward and match the behavior directly. GAIL casts imitation as a GAN — a discriminator separating expert from learner state-action pairs supplies the reward a policy-gradient method optimizes — matching occupancy measures without ever naming a reward. AIRL reads a transferable reward back out of the discriminator. We compare all four methods and close with reward models in RLHF, scaled cloning, and diffusion policies.\n",{"path":27915,"title":27916,"module":27868,"summary":27917},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmulti-agent-rl","Multi-Agent RL: Markov Games and Centralized Training","With more than one learning agent in an environment, each agent's world becomes non-stationary because the others are changing too. This lesson builds the Markov-game generalization of the MDP, diagnoses non-stationarity as the central obstacle, shows why the naive baselines fail, and develops the dominant fix — centralized training with decentralized execution (MADDPG, VDN, QMIX). A companion lesson takes up self-play, the landmark game-playing systems, and the equilibrium concepts that define what \"solved\" means.\n",{"path":27919,"title":27920,"module":27868,"summary":27921},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmulti-agent-rl-part-2","Multi-Agent RL: Self-Play and Solution Concepts","A companion to the Markov-games lesson. In the purely competitive setting, an agent can generate its own training curriculum by playing against copies of itself — self-play, the method behind AlphaGo, OpenAI Five, and AlphaStar. We develop why self-play produces an ever-improving opponent, the systems it built, and then the equilibrium solution concepts (Nash, correlated, coarse-correlated) that define what \"solved\" means once there is an opponent, closing with PSRO, MAPPO, and the language-model-agent frontier.\n",{"path":27923,"title":27924,"module":27868,"summary":27925},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fhierarchical-rl","Hierarchical RL: Options and the Option-Critic","Flat RL cannot explore a long horizon: reaching reward through hundreds of primitive actions is exponentially unlikely, and every credit-assignment update crawls one step at a time. Hierarchy breaks one hard long-horizon problem into many short ones. This lesson develops temporal abstraction — the options framework and its semi-Markov view, and learning options end to end with the option-critic. A companion lesson takes up goal-conditioned manager\u002Fworker hierarchies (FeUdal Networks and HIRO), hindsight relabeling, and unsupervised skill discovery.\n",{"path":27927,"title":27928,"module":27868,"summary":27929},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fhierarchical-rl-part-2","Hierarchical RL: Goal-Conditioned Hierarchies and Skills","A companion to the options lesson. Options package a behavior; goal-conditioned hierarchies instead give the top level an explicit language of goals — a manager proposes a target state or a latent direction, and a worker is rewarded for reaching it (FeUdal Networks, HIRO). We develop that architecture, the hindsight relabeling that lets it learn from sparse reward, and unsupervised skill discovery (DIAYN) that learns a repertoire of behaviors with no reward at all. The shared idea throughout: shorten the horizon by inserting a level that decides less often.\n",{"path":27931,"title":27932,"module":27868,"summary":27933},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Frlhf-and-language-models","RLHF and Language Models","A language model trained to predict the next token is fluent but not helpful, honest, or harmless — the objective it was optimized for is not the objective we want. RLHF closes that gap by turning the one thing humans do reliably, comparing two outputs, into a reward. We build the three-stage pipeline: supervised fine-tuning, a Bradley-Terry reward model fit to preference pairs, then PPO against that reward with a KL penalty keeping it near the reference policy. We then cover reward hacking and why the KL penalty matters, Direct Preference Optimization, which folds the reward model into a single classification loss, and the RLAIF and verifiable-reward variants. This pipeline is what makes the largest models usable as assistants.\n",{"path":27935,"title":27936,"module":27868,"summary":27937},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fpartial-observability-pomdps","Partial Observability: POMDPs and the Belief State","Drop the assumption that the agent sees the state. It sees an observation, a partial and noisy function of a hidden state, and one observation is no longer a Markov signal. This lesson builds the POMDP tuple, shows that the belief state — the posterior over hidden states — is a sufficient statistic that turns a POMDP back into an MDP over beliefs, and works the Bayes-filter belief update step by step. A companion lesson explains why exact planning is intractable and develops the deep-RL answer of recurrent, history-based policies.\n",{"path":27939,"title":27940,"module":27868,"summary":27941},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fpartial-observability-pomdps-part-2","Partial Observability: Planning and Recurrent Policies","A companion to the belief-state lesson. In principle a POMDP reduces to an MDP over beliefs; in practice two obstacles block that. Exact planning over the belief simplex is intractable — the value function is piecewise-linear-and-convex with a number of pieces that can explode — and computing the belief needs a model the agent rarely has. This lesson develops the intractability, the point-based approximations that address it, and the deep-RL answer: make the policy a function of history with a recurrent network (DRQN, R2D2), with frame-stacking, attention, and world-model latents as learned beliefs.\n",{"path":27943,"title":27944,"module":27868,"summary":27945},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fsafe-and-constrained-rl","Safe and Constrained RL: The CMDP and Policy Methods","Maximizing a scalar reward is not the same as behaving well: a capable optimizer will find and exploit any gap between the reward and what its designer actually meant, a failure called specification gaming or reward hacking. The remedy is to add explicit cost constraints — the constrained MDP — maximizing return subject to an expected-cost budget. This lesson builds the core toolkit: the CMDP itself, Lagrangian primal-dual methods that learn a multiplier on the constraint (RCPO), and constrained policy optimization (CPO) with its trust-region cost bound. A companion lesson covers risk-sensitivity, safe exploration, and the alignment framing.\n",{"path":27947,"title":27948,"module":27868,"summary":27949},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fsafe-and-constrained-rl-part-2","Safe RL: Risk, Safe Exploration, and Alignment","A companion to the constrained-MDP lesson. Constraining the mean cost is not enough: a policy safe on average can be catastrophic in the tail, and a policy safe at convergence can violate its limits wildly while learning. This lesson optimizes the tail with risk-sensitive objectives (CVaR), then makes exploration itself safe with shields, Lyapunov methods, and safety layers that project unsafe actions onto the feasible set — closing with benchmarks, safe RLHF, robustness, and the alignment framing that ties safety back to the problem of incompletely specified reward.\n",{"path":27951,"title":27952,"module":27868,"summary":27953},"\u002Freinforcement-learning\u002Fmodern-deep-rl\u002Fmeta-rl-and-generalization","Meta-RL and Generalization","An agent that masters one task often fails on the next; it has overfit to a single environment. This lesson treats fast adaptation as a meta-problem over a distribution of tasks: meta-train so that a few episodes at meta-test time suffice. We cover the two families — optimization-based (MAML learns an initialization) and context-based (RL-squared and PEARL infer a latent task) — the exploration cost of adaptation, and the parallel problem of generalization: why deep RL memorizes environments and what fixes it (domain randomization, procedural generation, augmentation, regularization). It closes on foundation models and sequence-model agents as the generalist endpoint.\n",{"path":27955,"title":27956,"module":27957,"summary":27958},"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fpsychology-of-reinforcement","The Psychology of Reinforcement","Reinforcement Learning in Minds and Brains","Reinforcement learning is both an engineering method and a theory of how animals learn. The prediction\u002Fcontrol split of the algorithms mirrors the psychologist's split between classical and instrumental conditioning. We trace the correspondence: the Rescorla–Wagner model as a prediction-error rule that explains blocking, its real-time TD extension, Thorndike's Law of Effect behind trial-and-error control, and the habitual\u002Fgoal-directed distinction that maps onto model-free versus model-based learning.\n",{"path":27960,"title":27961,"module":27957,"summary":27962},"\u002Freinforcement-learning\u002Fminds-and-brains\u002Finstrumental-conditioning-and-control","The Psychology of Reinforcement: Instrumental Control","Classical conditioning was prediction; instrumental conditioning is control. Thorndike's Law of Effect is trial-and-error control — selection plus association, search plus memory — and Skinner's shaping and schedules are reward engineering. The habitual\u002Fgoal-directed distinction maps onto model-free versus model-based control, dissociated by outcome devaluation and arbitrated by uncertainty. Delayed reinforcement is the credit-assignment problem, and the stimulus traces and secondary reinforcers of animal-learning theory are eligibility traces and value functions.\n",{"path":27964,"title":27965,"module":27957,"summary":27966},"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fdopamine-and-td-error","Dopamine and the TD Error","The TD error was invented as an algorithm; a decade later it turned out to closely describe the firing of the brain's dopamine neurons. We follow Schultz's experiments — dopamine fires at an unpredicted reward, shifts to the earliest predictive cue, and dips below baseline when a predicted reward is withheld — and match each result to the TD error term by term. We then read the basal ganglia as a neural actor–critic with dopamine as its shared training signal, and close on addiction as a hijacking of that signal.\n",{"path":27968,"title":27969,"module":27957,"summary":27970},"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fdopamine-in-the-brain","Dopamine in the Brain: The Neural Actor–Critic","If phasic dopamine is a TD error, where does it go and what does it change? We follow the axons into the basal ganglia, read the corticostriatal synapse as the place where state, action, and error meet, and map the ventral and dorsal striatum onto the critic and the actor of an actor–critic. Addiction becomes a broken cancellation in the same learning signal, and distributional dopamine extends the scalar RPE into a population code.\n",{"path":27972,"title":27973,"module":27957,"summary":27974},"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fanimal-learning-and-cognition","Animal Learning and Cognition","Three classic associative phenomena turn out to be reinforcement-learning mechanisms seen in behavior. Blocking says learning is driven by prediction error, not co-occurrence, and reduces to least-squares regression fitting a collinear feature. Higher-order conditioning and conditioned reinforcement make a value estimate a secondary reinforcer — bootstrapping in an animal. Delayed reinforcement is the credit-assignment problem, and the stimulus traces and goal gradients of Pavlov and Hull are eligibility traces and TD-learned value functions.\n",{"path":27976,"title":27977,"module":27957,"summary":27978},"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fcognitive-maps-and-planning","Cognitive Maps and Model-Based Learning","Tolman's rats learned the layout of a maze with no reward, then used it the moment food appeared — latent learning, a cognitive map, and the behavioral face of model-based reinforcement learning. The map is learned by system identification (stimulus–stimulus associations), which fills in whether or not reward is present, and queried by planning, which re-solves a route from a single changed reward. The successor representation sits between cache and model, and hippocampal predictive maps and scaled-up world models carry the same idea into brain and machine.\n",{"path":27980,"title":27981,"module":27957,"summary":27982},"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fneuroscience-of-reinforcement","The Neuroscience of Reinforcement","The dopamine story is one contact point between reinforcement learning and the brain; this lesson fills in the surrounding neuroscience so the mapping stands on its own. We build a working primer of neurons, synapses, and neuromodulation; separate four signals that casual usage conflates — reward, reinforcement, value, and prediction error; and read the actor and critic as corticostriatal synapses updated by two- and three-factor rules, grounded in spike-timing-dependent and reward-modulated plasticity.\n",{"path":27984,"title":27985,"module":27957,"summary":27986},"\u002Freinforcement-learning\u002Fminds-and-brains\u002Fseveral-learning-systems","The Brain's Several Learning Systems","The actor's three-factor rule has an ancestor in Klopf's hedonistic neuron — a single cell as a reinforcement-seeking agent — and a bacterium's run-and-twiddle shows the Law of Effect with no synapses at all. Teams of such neurons implement policy gradient collectively, the broadcast reward replacing backpropagation. And the brain is not only model-free: outcome devaluation, prefrontal value coding, and hippocampal forward sweeps localize a model-based system. The recurring conclusion is that the brain is several interacting learning systems, not one algorithm.\n",{"path":27988,"title":27145,"module":6,"summary":6},"\u002Freinforcement-learning",{"path":27990,"title":27991,"module":24340,"summary":27992},"\u002Fartificial-intelligence\u002Ffoundations\u002Fwhat-is-ai","What Is Artificial Intelligence?","Eight definitions of AI fall into a two-by-two grid: think versus act, and measure success against human performance versus an ideal standard of rationality. We work through all four schools — the Turing test, cognitive modelling, the laws of thought, and the rational agent — and adopt the last as the frame for the whole course: AI is the study and design of rational agents.\n",{"path":27994,"title":27995,"module":24340,"summary":27996},"\u002Fartificial-intelligence\u002Ffoundations\u002Ffoundations-of-ai","The Foundations of AI","Where the rational-agent idea came from and what surrounds it. AI inherited its core tools from eight older disciplines — philosophy, mathematics, economics, neuroscience, psychology, computer engineering, control theory, and linguistics. Its history runs in cycles of boom and winter, from the 1956 Dartmouth workshop through expert systems to the statistical turn. And the deep-learning era — AlexNet, the Transformer, GPT-3, AlphaGo — is a new way of computing the agent function at scale, not a new definition of AI.\n",{"path":27998,"title":27999,"module":24340,"summary":28000},"\u002Fartificial-intelligence\u002Ffoundations\u002Fintelligent-agents","Intelligent Agents","An agent perceives an environment through sensors and acts on it through actuators; its behavior is an agent function mapping percept sequences to actions. A rational agent chooses, for each percept sequence, the action that maximizes its expected performance measure given its knowledge. We build the first half of the vocabulary the whole course rests on — the agent function, rationality, PEAS task specifications, and the six axes along which task environments vary.\n",{"path":28002,"title":28003,"module":24340,"summary":28004},"\u002Fartificial-intelligence\u002Ffoundations\u002Fagent-architectures","Agent Architectures","How to build a program that computes a good agent function without storing an astronomically large lookup table. Four skeleton architectures in order of increasing power — simple reflex, model-based, goal-based, and utility-based — plus the learning agent that improves any of them, the scale of world representations (atomic, factored, structured) they rest on, and how a modern language-model agent fits the same frame.\n",{"path":28006,"title":28007,"module":28008,"summary":28009},"\u002Fartificial-intelligence\u002Fsearch\u002Funinformed-search","Uninformed Search","Search","A goal-based agent that cannot see which action is best turns the problem into a state space — an initial state, a set of actions, a transition model, a goal test, and a path cost — and searches for a sequence of actions reaching the goal. We build the state-space formulation on the 8-puzzle and route-finding, give the one TREE-SEARCH \u002F GRAPH-SEARCH skeleton every algorithm specializes, and measure strategies by completeness, optimality, and complexity. This lesson develops the first two frontier disciplines — breadth-first and uniform-cost search; the rest follow in the next lesson.\n",{"path":28011,"title":28012,"module":28008,"summary":28013},"\u002Fartificial-intelligence\u002Fsearch\u002Fsearch-strategies-compared","Search Strategies Compared","Breadth-first and uniform-cost search pay for optimality in memory. This lesson develops the strategies that trade memory for depth: depth-first search, which keeps only the current path; depth-limited and iterative-deepening search, which fix DFS's failure on infinite paths; and bidirectional search, which meets in the middle for a square-root saving. It closes by lining up all six uninformed strategies against completeness, optimality, and complexity, and tracing where the algorithms came from and where they went.\n",{"path":28015,"title":28016,"module":28008,"summary":28017},"\u002Fartificial-intelligence\u002Fsearch\u002Finformed-search","Informed Search and A*","An informed search uses a heuristic $h(n)$, an estimate of the cost from a node to the goal, to decide what to expand next. Greedy best-first search follows the heuristic blindly and gives up optimality; A* corrects it by ranking nodes on $f(n) = g(n) + h(n)$, and is optimal when the heuristic is admissible (tree search) or consistent (graph search). This lesson defines the heuristic, builds best-first search, and proves why A* is optimal, with the contour picture that explains its pruning. Where good heuristics come from is the next lesson.\n",{"path":28019,"title":28020,"module":28008,"summary":28021},"\u002Fartificial-intelligence\u002Fsearch\u002Fheuristic-functions","Heuristic Functions and Memory-Bounded Search","A* is only as good as its heuristic, so this lesson answers where good heuristics come from: relaxed problems, whose exact solution cost is an admissible heuristic, and pattern databases, which precompute subproblem costs. It measures heuristic quality with dominance and the effective branching factor, then tackles A*'s memory problem with IDA*, RBFS, and SMA*. It closes with modern heuristic search — weighted A*, learned and disjoint pattern-database heuristics, and bidirectional A*.\n",{"path":28023,"title":28024,"module":28008,"summary":28025},"\u002Fartificial-intelligence\u002Fsearch\u002Flocal-search","Local Search and Optimization","When the path to a goal is irrelevant and only the final state matters, we can discard the search tree entirely and keep just the current state, moving to a better neighbor at each step. This lesson builds the state-space landscape metaphor, works through hill climbing and the three obstacles that defeat it (local maxima, ridges, plateaus), then develops the first escapes: random restarts and simulated annealing with its temperature schedule. The population-based methods and continuous-space calculus follow in the next lesson.\n",{"path":28027,"title":28028,"module":28008,"summary":28029},"\u002Fartificial-intelligence\u002Fsearch\u002Fpopulation-and-continuous-search","Population and Continuous Search","Single-state local search escapes a trap by restarting or tolerating downhill moves. This lesson develops the alternatives that keep several states at once — local beam search, which shares successors across parallel threads, and genetic algorithms, which recombine two parents through crossover and mutation — then crosses into continuous spaces, where calculus replaces the finite neighbor set: gradient ascent, line search, and Newton's method. It closes with the industrial descendants of these methods and the loop they all share.\n",{"path":28031,"title":28032,"module":28008,"summary":28033},"\u002Fartificial-intelligence\u002Fsearch\u002Fadversarial-search","Adversarial Search and Games","When another agent plans against you, search becomes a game. We formalize two-player, zero-sum, perfect-information games as search problems, define the minimax value that optimal play backs up through the game tree, and give the MINIMAX algorithm that computes it. Alpha–beta pruning then cuts the cost of that search roughly in half in the exponent without changing the answer, and a heuristic evaluation function plus a cutoff test turns the exact algorithm into a real-time player that copes with the horizon effect.\n",{"path":28035,"title":28036,"module":28008,"summary":28037},"\u002Fartificial-intelligence\u002Fsearch\u002Fgames-of-chance-and-imperfect-information","Games of Chance and Imperfect Information","Minimax and alpha–beta assume a deterministic game both players can see in full. Drop either assumption and search must change. This lesson adds chance nodes and the expectiminimax value for games with dice, then belief-state reasoning for partially observable games — Kriegspiel and card games — where averaging over clairvoyance both helps and misleads. It closes with the line from Deep Blue's alpha–beta to AlphaGo's learned evaluation and Monte Carlo tree search, and the provable-pruning and self-play research around each end of that story.\n",{"path":28039,"title":28040,"module":28008,"summary":28041},"\u002Fartificial-intelligence\u002Fsearch\u002Fconstraint-satisfaction","Constraint Satisfaction Problems","A constraint satisfaction problem replaces the black-box state with a factored one: variables, domains, and constraints. That structure supports inference before any search runs. This lesson defines the CSP on map coloring, Sudoku, and scheduling, then develops constraint propagation: node and arc consistency, the AC-3 algorithm that makes a whole network arc-consistent, and the way one deleted value cascades across the graph to prune impossible options ahead of search.\n",{"path":28043,"title":28044,"module":28008,"summary":28045},"\u002Fartificial-intelligence\u002Fsearch\u002Fcsp-search-and-structure","CSP Search and Structure","Propagation prunes a CSP but rarely finishes it, so we search. This lesson builds backtracking search over partial assignments and the general-purpose heuristics that make it fast — MRV, degree, least-constraining-value, forward checking, MAC, and intelligent backtracking. It then shows how the shape of the constraint graph controls difficulty: tree-structured problems fall in linear time, cutset conditioning handles the rest, and min-conflicts local search solves a million queens in a constant number of steps.\n",{"path":28047,"title":28048,"module":28008,"summary":28049},"\u002Fartificial-intelligence\u002Fsearch\u002Fsearch-under-uncertainty","Search Under Uncertainty","Classical search assumes the agent knows the state it is in and exactly what each action does. Drop the second assumption and a plan can no longer be a fixed sequence of actions. This lesson develops the first response: AND-OR search over nondeterministic actions, which returns a branching contingency plan rather than a straight line. We build it on the erratic vacuum world, show how OR nodes (the agent's choices) alternate with AND nodes (nature's outcomes), trace the recursion that finds a plan, and handle the case where the only solution is a cyclic \"try, try again.\"\n",{"path":28051,"title":28052,"module":28008,"summary":28053},"\u002Fartificial-intelligence\u002Fsearch\u002Fbelief-state-and-online-search","Belief-State and Online Search","When the agent cannot see the full state, a plan can no longer test where it actually is — it must reason over the set of states it might be in. This lesson develops belief-state search, from sensorless (conformant) planning that coerces an unknown world into a goal, through the predict-observe-update cycle of contingent planning with percepts, to online search in unknown environments, where the agent must act in order to learn. It closes with LRTA*, which refines its own heuristic as it explores, one step from reinforcement learning.\n",{"path":28055,"title":28056,"module":28057,"summary":28058},"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fpropositional-logic","Logical Agents and Propositional Logic","Logic and Planning","A knowledge-based agent keeps a store of sentences and acts by asking it what to do. To make \"asking\" mean something we need entailment — the relation $KB \\models \\alpha$ that holds when every model of the knowledge base is a model of the query. Propositional logic gives a syntax and a truth-table semantics for which entailment is decidable. This first part builds the foundations: the agent loop, the Wumpus World, models and entailment, the connectives and truth tables, theorem proving by refutation, and the resolution rule with its CNF conversion — a single complete inference procedure for all of propositional logic.\n",{"path":28060,"title":28061,"module":28057,"summary":28062},"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fpropositional-inference","Propositional Inference and Logical Agents","Model checking and resolution decide entailment, but both can blow up. This part turns propositional logic into a practical engine and a working agent. Horn clauses give linear-time forward and backward chaining — the basis of logic programming. DPLL and WalkSAT make satisfiability testing fast in the common case. Then we make the agent situated: time-indexed fluents, the frame problem and its solution by successor-state axioms, a hybrid agent that deduces a safe map and plans a route through it, and SATPlan, which finds a plan by asking a SAT solver for a satisfying model.\n",{"path":28064,"title":28065,"module":28057,"summary":28066},"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Ffirst-order-logic","First-Order Logic","Propositional logic can only say that facts hold; it cannot talk about the objects a fact is about, or state a rule once and have it cover every object. First-order logic fixes this by committing to a world of objects, relations, and functions. This first part builds the language from the ground up: the ontology it commits to, the model that gives a sentence a truth value, the syntax of terms and sentences, the two quantifiers with their standard mistakes, and equality.\n",{"path":28068,"title":28069,"module":28057,"summary":28070},"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Ffirst-order-logic-in-use","First-Order Logic in Use","With the language of first-order logic in hand, this part is about using it well. Database semantics trades expressive power for the convenience of a single intended model; higher-order logic shows what first-order logic gives up for decidability. Then we put the language to work: the Tell\u002FAsk interface, the kinship domain axiomatized from scratch, and the seven-step knowledge-engineering process applied to a digital circuit.\n",{"path":28072,"title":28073,"module":28057,"summary":28074},"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Finference-and-resolution","Inference in First-Order Logic","Propositional inference lifts to first-order logic once we can make terms match. Unification is that machinery: the algorithm that finds the substitution making two expressions identical, and the basis of generalized modus ponens. This first part builds the lifted inference rules and the two chaining algorithms they drive — forward chaining, the data-driven procedure behind production systems and Datalog, and backward chaining, the goal-driven procedure behind Prolog.\n",{"path":28076,"title":28077,"module":28057,"summary":28078},"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Ffirst-order-resolution","First-Order Resolution","Chaining is complete only for Horn knowledge bases. General first-order sentences — with disjunctive conclusions and negations — need a single sound and complete rule: resolution. This part converts arbitrary sentences to CNF by skolemizing away the existentials, lifts the resolution rule with unification, and proves entailment by refuting the negated goal. The result is the proof procedure Gödel's completeness theorem guarantees will find any entailment, together with the search strategies that make it usable.\n",{"path":28080,"title":28081,"module":28057,"summary":28082},"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fclassical-planning","Classical Planning","Classical planning represents a problem in a factored language, PDDL: states are sets of ground fluents, and actions are lifted schemas with a precondition and an effect. That structure turns planning into search — forward through states or backward through goals — and lets a program read heuristics straight off the schemas by relaxing the problem. This first part develops the representation, the two search directions, and the domain-independent heuristics that come from ignoring preconditions or delete lists.\n",{"path":28084,"title":28085,"module":28057,"summary":28086},"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fplanning-graphs-and-graphplan","Planning Heuristics and GraphPlan","Every relaxation heuristic can be inaccurate, and none can tell how far apart subgoals sit. The planning graph is a polynomial-size structure that does better: leveled off the problem, it yields admissible distance estimates and a record of which actions and fluents cannot coexist. This part builds the graph, reads heuristics from it, extracts plans with GraphPlan, and closes with the other classical approaches — SATPlan and partial-order planning — and the representational trade that makes all of it work.\n",{"path":28088,"title":28089,"module":28057,"summary":28090},"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fplanning-in-the-real-world","Planning and Acting in the Real World","Classical planning's clean theory rests on four assumptions: time is ignored, actions are atomic, the world is deterministic and fully observable, and the agent is alone. This first part drops the first two. We add durations and resource constraints — turning a plan into a schedule, solved by the critical-path method and, once resources contend, by NP-hard job-shop scheduling — and let a planner reason at multiple levels of abstraction through high-level actions and their angelic reachable sets.\n",{"path":28092,"title":28093,"module":28057,"summary":28094},"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fplanning-under-uncertainty","Planning Under Uncertainty","Classical planning assumed the world was deterministic, fully observable, and the agent alone. This part drops the last two assumptions. When the agent cannot see or predict the world, planning moves into belief-state space: sensorless plans that coerce the world into the goal without sensing, contingent plans that branch on what is sensed, and online agents that monitor and replan when execution diverges. Then we add other agents — joint plans, the coordination problem, and the conventions that let a team act without constant negotiation.\n",{"path":28096,"title":28097,"module":28057,"summary":28098},"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Fknowledge-representation","Knowledge Representation","First-order logic gives you the language; this lesson is about what to say in it. This first part builds the content: a general upper ontology from the top down, categories as first-class objects with taxonomies and inheritance, physical composition and the count-noun\u002Fmass-noun split, events and time reified through the event calculus, and belief modeled with modal logic — the machinery for representing the world an agent reasons about.\n",{"path":28100,"title":28101,"module":28057,"summary":28102},"\u002Fartificial-intelligence\u002Flogic-and-planning\u002Freasoning-systems-and-defaults","Reasoning Systems and Default Logic","Having represented the world, this part is about reasoning with it at scale. Semantic networks give a graphical notation with fast inheritance; description logics keep subsumption and classification tractable by design. Then we confront the fact that most useful rules hold only by default: circumscription and default logic give a logical account of nonmonotonic reasoning, and truth maintenance systems retract conclusions cleanly when the beliefs beneath them change.\n",{"path":28104,"title":28105,"module":28106,"summary":28107},"\u002Fartificial-intelligence\u002Funcertainty\u002Fprobability-and-bayes","Quantifying Uncertainty","Uncertainty","Logic breaks down in any domain where the rules have exceptions you cannot enumerate — the qualification problem. Probability replaces truth values with degrees of belief that obey Kolmogorov's axioms, and the full joint distribution becomes a knowledge base from which any query is answered by summing entries: marginalization, conditioning, and normalization. Independence factors that joint into smaller pieces — the first step toward a calculus of rational belief that an agent can actually compute with.\n",{"path":28109,"title":28110,"module":28106,"summary":28111},"\u002Fartificial-intelligence\u002Funcertainty\u002Fbayes-rule-and-naive-bayes","Bayes' Rule and Naive Bayes","Bayes' rule inverts a causal model into a diagnostic one, turning \"how a cause produces its symptoms\" into \"which cause explains what I observed.\" Ignoring the prior is the base-rate fallacy behind overconfident test results. Conditional independence then lets several pieces of evidence combine by multiplying likelihood ratios instead of building an exponential joint, giving the naive Bayes model and pointing directly at Bayesian networks.\n",{"path":28113,"title":28114,"module":28106,"summary":28115},"\u002Fartificial-intelligence\u002Funcertainty\u002Fbayesian-networks","Bayesian Networks","A Bayesian network is a directed acyclic graph of random variables in which each node carries a conditional probability table for itself given its parents. That structure factors the full joint distribution into a product of local terms, turning an exponential table into a linear one, and it makes the conditional independences of the domain explicit. We build the canonical burglary–alarm network, read compactness and d-separation off the graph, run exact inference by variable elimination, and, where that is intractable, estimate answers by sampling.\n",{"path":28117,"title":28118,"module":28106,"summary":28119},"\u002Fartificial-intelligence\u002Funcertainty\u002Finference-in-bayesian-networks","Bayesian Networks: Inference and Relational Models","When exact inference is intractable, sampling estimates the posterior instead: prior and rejection sampling, likelihood weighting, and Gibbs\u002FMCMC, whose error shrinks as one over the square root of the sample count. The same graphical idea then lifts from a fixed set of variables to whole populations — relational and open-universe probability models write dependencies once and unroll them over objects — and we close by placing probability against the rule-based, Dempster–Shafer, and fuzzy alternatives it displaced.\n",{"path":28121,"title":28122,"module":28106,"summary":28123},"\u002Fartificial-intelligence\u002Funcertainty\u002Freasoning-over-time","Probabilistic Reasoning over Time","A world that changes needs a state variable at every point in time. The Markov assumption cuts the dependence on history down to the previous slice, leaving a transition model and a sensor model that define a temporal Bayesian network. Four recursive tasks fall out — filtering, prediction, smoothing, and the most likely explanation — each a message passed along the sequence. We ground them in hidden Markov models and their matrix form, sketch the Kalman filter for continuous state, and reach dynamic Bayesian networks with particle filtering as the general approximate method.\n",{"path":28125,"title":28126,"module":28106,"summary":28127},"\u002Fartificial-intelligence\u002Funcertainty\u002Ftracking-and-data-association","Reasoning over Time: Tracking and Data Association","Dynamic Bayesian networks generalize HMMs and Kalman filters to arbitrarily many state variables per slice, and when exact inference blows up, particle filtering approximates the belief state with a population of weighted samples that propagate, reweight, and resample. Tracking several objects at once adds the data-association problem — which observation came from which object — whose combinatorics defeat any exact filter, so particle filters and MCMC keep many hypotheses alive. We close with SLAM and learned state-space models.\n",{"path":28129,"title":28130,"module":28106,"summary":28131},"\u002Fartificial-intelligence\u002Funcertainty\u002Fmaking-decisions","Making Decisions: Utility Theory","A rational agent chooses the action that maximizes expected utility — the probability of each outcome weighted by how much the agent wants it. We derive the utility function from six axioms on preferences, so maximizing expected utility is forced by consistency rather than assumed; look at risk aversion in the utility-of-money curve; package one-shot choices into decision networks; and quantify what an observation is worth with the value of information.\n",{"path":28133,"title":27722,"module":28106,"summary":28134},"\u002Fartificial-intelligence\u002Funcertainty\u002Fmarkov-decision-processes","When an agent must act repeatedly in a stochastic world, a fixed plan is useless — it needs a policy, an action for every state. The Markov decision process makes this precise with a transition model, a reward, and a discount factor; the Bellman equation characterizes the optimal state utilities, and value iteration and policy iteration solve it. Partial observability lifts the problem to belief states, and bandits, Monte-Carlo tree search, and scalable POMDP solvers extend it — this is the model-known half of reinforcement learning.\n",{"path":28136,"title":28137,"module":28106,"summary":28138},"\u002Fartificial-intelligence\u002Funcertainty\u002Fdecision-networks-and-game-theory","Decision Analysis: Multi-Attribute Utility and Decision Networks","Decision analysis takes the single-agent utility framework and makes it practical: utility over several attributes, dominance and additive value functions, influence diagrams that fold Bayesian networks together with decision and utility nodes, and the value of information that tells an agent which questions are worth asking. Structure in an agent's preferences — dominance, preferential and utility independence — collapses an exponential utility table into a few one-dimensional functions, the same move that made Bayesian networks compact.\n",{"path":28140,"title":28141,"module":28106,"summary":28142},"\u002Fartificial-intelligence\u002Funcertainty\u002Fgame-theory-and-mechanism-design","Game Theory and Mechanism Design","When outcomes depend on other rational agents, single-agent utility maximization no longer suffices. Game theory studies decisions among agents — normal-form games, dominant strategies, Nash and maximin equilibria, and repeated games — and mechanism design runs the logic backwards, engineering rules (auctions, VCG) so that self-interested play produces a good collective outcome. Algorithmic game theory then asks whether equilibria can be computed, what selfishness costs society, and how the mechanisms deployed at internet scale actually behave.\n",{"path":28144,"title":28145,"module":28146,"summary":28147},"\u002Fartificial-intelligence\u002Flearning\u002Flearning-from-examples","Learning from Examples","Learning","An agent that improves with experience does not need its designer to anticipate every situation. Inductive learning takes that ambition and narrows it to one tractable problem: from labelled input-output pairs, recover a function that predicts the output for inputs never seen. This first part builds the foundation around a single organizing question — generalization — through decision trees and information gain, and the training\u002Fvalidation\u002Ftest discipline for evaluating and choosing hypotheses. A second part takes up the theory of learning and the main model families.\n",{"path":28149,"title":28150,"module":28146,"summary":28151},"\u002Fartificial-intelligence\u002Flearning\u002Ftheory-and-model-families","The Theory of Learning and Model Families","Cross-validation measures generalization but does not explain it. This part supplies the theory — PAC learning, sample complexity, and the VC dimension — that says when a hypothesis consistent with enough data is probably approximately correct, and why an unrestricted hypothesis space can never generalize. It then surveys the model families a practitioner reaches for: linear regression and gradient descent, the perceptron and logistic regression, support vector machines and the kernel trick, and ensembles by bagging and boosting — closing with what deep learning changed about the classical picture.\n",{"path":28153,"title":28154,"module":28146,"summary":28155},"\u002Fartificial-intelligence\u002Flearning\u002Fprobabilistic-learning","Learning Probabilistic Models","A [Bayesian network](\u002Fartificial-intelligence\u002Funcertainty\u002Fbayesian-networks) is useless until its numbers are filled in, and those numbers come from data. This first part casts learning itself as probabilistic inference: hypotheses carry a prior, data update it to a posterior, and predictions average over what remains. From that frame fall the standard estimators — maximum likelihood by counting, MAP with a conjugate prior, full Bayesian updating — for the case where every variable is observed. A second part takes up the harder case of hidden variables and the EM algorithm.\n",{"path":28157,"title":28158,"module":28146,"summary":28159},"\u002Fartificial-intelligence\u002Flearning\u002Fexpectation-maximization","Learning with Hidden Variables: The EM Algorithm","Complete data can be learned by counting; real data usually hide some variables — the disease behind the symptoms, the cluster behind the points. This part develops the expectation-maximization algorithm, which learns those models by alternating an expected completion of the missing data with a re-estimation of the parameters. It works the idea through mixtures of Gaussians, Bayesian networks, and hidden Markov models, proves the monotone-likelihood guarantee from the evidence lower bound, and traces the line from EM to variational inference and the variational autoencoder.\n",{"path":28161,"title":27145,"module":28146,"summary":28162},"\u002Fartificial-intelligence\u002Flearning\u002Freinforcement-learning","Reinforcement learning is an MDP with the model unknown: the agent knows neither how its actions move the world nor which states are rewarded, and must recover good behaviour from experienced transitions and rewards alone. This first part builds the classical tabular theory — passive learning (fix a policy, learn its value, by direct estimation, adaptive dynamic programming, and temporal differences) and active learning (choose actions, trade exploration against exploitation, and learn control with Q-learning and SARSA). A second part lifts it off the lookup table with function approximation and policy search.\n",{"path":28164,"title":28165,"module":28146,"summary":28166},"\u002Fartificial-intelligence\u002Flearning\u002Fgeneralization-and-policy-search","Reinforcement Learning: Generalization and Policy Search","Tabular reinforcement learning stores one number per state, which is hopeless for backgammon or chess. This part lifts RL off the lookup table with function approximation, so that updating one state generalizes to related ones, then turns to policy search — representing and optimizing the policy directly, up to the REINFORCE policy gradient and correlated sampling. It closes with the bridge to deep reinforcement learning (deep Q-networks, actor-critic, PPO), the classic applications, and the hand-off to the dedicated RL subject.\n",{"path":28168,"title":28169,"module":28146,"summary":28170},"\u002Fartificial-intelligence\u002Flearning\u002Fknowledge-in-learning","Knowledge in Learning","Pure induction learns a function from labelled examples while knowing almost nothing to begin with. This first part brings prior knowledge into the loop by recasting learning as logical inference — hypotheses, examples, and classifications as sentences. It develops current-best-hypothesis search, the version space and its general\u002Fspecific boundary maintained by candidate elimination, and states the three entailment constraints that fix how background knowledge enters. A second part builds the three knowledge-based methods those constraints define.\n",{"path":28172,"title":28173,"module":28146,"summary":28174},"\u002Fartificial-intelligence\u002Flearning\u002Fknowledge-based-learning-methods","Knowledge-Based Learning: EBL, Relevance, and ILP","Once learning is cast as logical inference, three methods follow from the three ways prior knowledge can enter. Explanation-based learning generalizes a single example by explaining it with the domain theory, gaining speed but nothing new. Relevance-based learning uses determinations to shrink the hypothesis space and converge from fewer examples. Inductive logic programming learns genuinely new first-order rules — top-down with FOIL, bottom-up by inverting resolution, even inventing new predicates — and connects to modern statistical relational and neuro-symbolic learning.\n",{"path":28176,"title":28177,"module":28178,"summary":28179},"\u002Fartificial-intelligence\u002Ffrontiers\u002Fvision-and-perception","Vision and Perception","Frontiers","Perception connects an agent to the physical world. We follow one modality — vision — from the physics of image formation (the pinhole camera, perspective projection, lenses, shading, color) through the early operations that turn a pixel array into edges, texture, and motion, and into recognition by appearance. The recurring problem is inversion: a camera collapses a 3-D world onto a 2-D grid, and an agent that wants to act must build the scene back up. Rebuilding the scene is the subject of the companion lesson.\n",{"path":28181,"title":28182,"module":28178,"summary":28183},"\u002Fartificial-intelligence\u002Ffrontiers\u002Freconstructing-the-3d-world","Vision: Reconstructing the 3D World","A camera collapses a three-dimensional world onto a flat grid; this lesson inverts that collapse. We build the camera projection matrix (intrinsics and extrinsics), triangulate a point from two views, then work through the toolbox of depth cues — motion parallax, binocular stereopsis, multiple views, texture, shading, and contour — that turn an ambiguous image back into a scene. We add structural recognition (pictorial-structure \"cardboard people\"), the task-driven use of vision in cars and robots, and the shift from hand-built pipelines to learned deep-vision networks.\n",{"path":28185,"title":28186,"module":28178,"summary":28187},"\u002Fartificial-intelligence\u002Ffrontiers\u002Frobotics","Robotics","A robot is an agent with a body: sensors that read the physical world and effectors that push back on it. This lesson grounds the abstract AI machinery in that body. We build up the hardware (range finders, proprioception, degrees of freedom), then cast perception as probabilistic filtering — the kinematic motion and sensor models, Monte Carlo localization, the extended Kalman filter, and simultaneous localization and mapping (SLAM). The companion lesson takes the estimated pose forward into planning and control.\n",{"path":28189,"title":28190,"module":28178,"summary":28191},"\u002Fartificial-intelligence\u002Ffrontiers\u002Frobot-planning-and-control","Robotics: Planning and Control","A robot that knows where it is still has to decide how to move, and then make a slipping, sensing-imperfect body actually go there. This lesson takes the pose estimate forward: planning motion in configuration space with cell decomposition and sampling-based roadmaps (PRMs and RRTs), planning under uncertainty with most-likely-state and online replanning, closing the loop with P\u002FPD\u002FPID control and potential fields, and finally the software architectures — subsumption, three-layer, and pipeline — that assemble it all, plus the learning-based turn in modern robotics.\n",{"path":28193,"title":28194,"module":28178,"summary":28195},"\u002Fartificial-intelligence\u002Ffrontiers\u002Fnatural-language-in-ai","Natural Language for AI Agents","Language is how agents acquire the knowledge already written down and how they communicate with the humans they serve. This lesson gives the classical AI account of language as a source of information: n-gram language models and the information-seeking tasks built on them — text classification, information retrieval (BM25, the inverted index, PageRank), and information extraction with finite-state templates and hidden Markov models. Throughout, we point to the dedicated NLP subject for the modern deep-learning treatment; the companion lesson takes up grammar, translation, and speech.\n",{"path":28197,"title":28198,"module":28178,"summary":28199},"\u002Fartificial-intelligence\u002Ffrontiers\u002Fnlp-grammar-translation-and-speech","Language for AI Agents: Grammar, Translation, and Speech","N-gram models see only a local window; they cannot say why \"black dog\" is well-formed English and \"dog black\" is not, because that is a fact about structure. This lesson takes up structure: phrase-structure and probabilistic context-free grammars, syntactic analysis by chart parsing and CYK, augmented grammars and compositional semantics, then the two major statistical successes — machine translation and speech recognition — cast as noisy-channel problems. It closes with the bridge from n-grams to transformers and where the classical account sits relative to modern NLP.\n",{"path":28201,"title":28202,"module":28178,"summary":28203},"\u002Fartificial-intelligence\u002Ffrontiers\u002Fphilosophy-and-future","Philosophy, Ethics, and the Future of AI","Two questions have shadowed the field since its founding: can machines act intelligently (weak AI), and can they really think (strong AI)? We work through Turing's objections and their rebuttals — the arguments from disability, mathematics, and informality — then the strong-AI debate: the mind-body problem, functionalism and the brain prosthesis, Searle's Chinese Room and the systems reply, and consciousness and qualia. The companion lesson turns from what AI can do to what it should, and closes the course.\n",{"path":28205,"title":28206,"module":28178,"summary":28207},"\u002Fartificial-intelligence\u002Ffrontiers\u002Fai-ethics-and-future","The Ethics and Future of AI","Having asked whether machines can act intelligently and really think, we turn to whether we should build them at all. This lesson works through the six ethical risks — lost jobs, autonomous weapons, surveillance and privacy, biased decisions, the safety of superintelligence, and the erosion of accountability — then the value-alignment problem in the LLM era, and where the classical agent components could go next. It closes the course by tying search, logic, probability, and learning into a single picture of intelligence as rational agency.\n",{"path":28209,"title":28210,"module":6,"summary":6},"\u002Fartificial-intelligence","Artificial Intelligence",{"path":28212,"title":28213,"module":28214,"summary":28215},"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fnuclear-constituents-nuclide-chart","Nuclear Composition and Ground-State Properties","Nuclear Properties","The nucleus is a bound assembly of Z protons and N neutrons packed to a radius R = R0 A^(1\u002F3) at a nearly constant density of about 10^17 kg\u002Fm^3. We fix the vocabulary of nuclides, derive nuclear size from mirror-nuclide and electron-scattering data, read the binding-energy-per-nucleon curve, and model it with the liquid-drop semiempirical mass formula.\n",{"path":28217,"title":28218,"module":28214,"summary":28219},"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fnuclear-size-charge-distributions","Nuclear Size, Shape, and Charge Distributions","Elastic electron scattering resolves the nucleus by its de Broglie wavelength. The measured cross section is the Mott point-charge cross section modulated by a form factor, and that form factor is the Fourier transform of the charge density. Diffraction minima fix the radius, the small-angle slope fixes the mean-square radius, and the fitted Woods-Saxon profile gives a central density and a skin thickness. Mirror-nucleus Coulomb energies, muonic-atom X-rays, and optical isotope shifts give independent radii that all track R = R0 A^(1\u002F3).\n",{"path":28221,"title":28222,"module":28214,"summary":28223},"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fnuclear-masses-binding-energy","Nuclear Masses, Mass Excess, and Separation Energies","The atomic mass unit fixes the scale, and the mass excess collects the small binding-driven deviation from the integer mass number. Penning-trap cyclotron frequencies now measure masses to parts in a billion, and every decay and reaction Q-value is a difference of these masses. One- and two-nucleon separation energies read the binding difference between neighbouring nuclides directly, showing the even-odd pairing stagger and the sharp drops at magic numbers, and their vanishing marks the neutron and proton drip lines that bound the chart of the nuclides.\n",{"path":28225,"title":28226,"module":28214,"summary":28227},"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fsemi-empirical-mass-formula","The Semi-Empirical Mass Formula and the Valley of Stability","Five physical terms reproduce nuclear binding across the chart: a volume term from saturation, a surface term from the deficit of edge neighbours, a Coulomb term from the electrostatic self-energy of a charged sphere, an asymmetry term from the Pauli cost of unequal proton and neutron filling, and a pairing term. The formula is quadratic in Z at fixed A, so isobars lie on a mass parabola whose minimum sets the most stable charge and whose slope dictates the direction of beta decay. The same competition between surface and Coulomb energy defines the fissility parameter and the onset of fission.\n",{"path":28229,"title":28230,"module":28214,"summary":28231},"\u002Fnuclear-physics\u002Fnuclear-properties\u002Fnuclear-moments-multipoles","Nuclear Spin, Magnetic Dipole, and Electric Quadrupole Moments","The ground state of a nucleus carries a definite spin and parity, a magnetic dipole moment of order the nuclear magneton, and, when its spin exceeds one-half, an electric quadrupole moment that measures its shape. The single-particle Schmidt lines predict the magnetic moment of an odd-A nucleus from the last unpaired nucleon, and the measured moments fall between them. The quadrupole moment distinguishes prolate from oblate deformation, and hyperfine structure is the experimental handle that fixes the spin and the moments from an atomic spectrum.\n",{"path":28233,"title":28234,"module":28235,"summary":28236},"\u002Fnuclear-physics\u002Fnuclear-force-deuteron\u002Fnuclear-force-shell-overview","The Nuclear Force and the Shell Model","The Nuclear Force","The strong force between nucleons is short-range, charge-independent, saturated, and repulsive at its core, about a hundred times stronger than Coulomb. Yukawa explained it as an exchange of massive mesons, tying the force's range to the meson mass through the uncertainty principle. Layered on top, an independent-particle shell model with strong spin-orbit coupling reproduces the magic numbers 2, 8, 20, 28, 50, 82, 126.\n",{"path":28238,"title":28239,"module":28235,"summary":28240},"\u002Fnuclear-physics\u002Fnuclear-force-deuteron\u002Fthe-deuteron","The Deuteron and the Tensor Force","The deuteron is the only bound two-nucleon state: one shallow level at 2.22 MeV, no excited states. A square-well fit fixes a depth near 35 MeV over a 2 fm range, yet the wavefunction leaks so far past the edge that most of the probability lies outside the force. Its spin-1 ground state, magnetic moment close to the sum of the free-nucleon moments, and small but nonzero electric quadrupole moment together force a D-state admixture and a non-central tensor force.\n",{"path":28242,"title":28243,"module":28235,"summary":28244},"\u002Fnuclear-physics\u002Fnuclear-force-deuteron\u002Fnucleon-nucleon-scattering","Nucleon-Nucleon Scattering and the Interaction's Structure","Scattering probes the nuclear force above threshold. Partial-wave analysis reduces low-energy data to a single s-wave phase shift, and the effective-range expansion packages that into a scattering length and an effective range. The triplet channel binds (the deuteron) while the singlet is only virtual, which together explain the anomalously large free neutron-proton cross section. Comparing pp, nn, and np results establishes charge symmetry and charge independence, and polarization experiments expose the spin-orbit and tensor pieces.\n",{"path":28246,"title":28247,"module":28235,"summary":28248},"\u002Fnuclear-physics\u002Fnuclear-force-deuteron\u002Fmeson-theory-isospin","Meson Exchange, the Yukawa Potential, and Isospin","Yukawa's massive-field propagator turns the range of the nuclear force into a meson mass: the exchanged quantum's Compton wavelength is the range. One-pion exchange fixes the long-range tail, complete with the tensor structure the deuteron demanded, while heavier mesons build the intermediate attraction and the repulsive core. Charge independence becomes an isospin symmetry, the force is diagonalized by the total isospin through a tau-dot-tau interaction, and the whole picture sits inside QCD as a residual color force between color-neutral nucleons.\n",{"path":28250,"title":28251,"module":28252,"summary":28253},"\u002Fnuclear-physics\u002Fnuclear-models\u002Ffermi-gas-model","The Fermi Gas Model","Nuclear Models","Treating the nucleus as two degenerate Fermi gases of protons and neutrons confined in a common well fixes the Fermi momentum near 250 MeV\u002Fc and the Fermi energy near 33 MeV from the nuclear density alone. The average kinetic energy per nucleon is about 20 MeV, the well depth is the Fermi energy plus the separation energy, and unequal proton and neutron Fermi levels reproduce the asymmetry term of the mass formula.\n",{"path":28255,"title":28256,"module":28252,"summary":28257},"\u002Fnuclear-physics\u002Fnuclear-models\u002Fliquid-drop-collective-coordinates","The Liquid-Drop Model and Collective Deformation","Deforming a charged liquid drop into a spheroid raises its surface energy and lowers its Coulomb energy; the two effects compete through the deformation parameter to set a stability minimum and a fission barrier. The ratio of Coulomb to twice the surface energy is the fissility Z-squared over A, which crosses one near 49 and marks the point where the sphere is unstable. The same surface tension that restores small deformations quantizes into collective vibrations, carrying the static mass formula into dynamic collective motion.\n",{"path":28259,"title":28260,"module":28252,"summary":28261},"\u002Fnuclear-physics\u002Fnuclear-models\u002Fshell-model-single-particle","The Shell Model: Single-Particle States and Spin-Orbit Coupling","A harmonic-oscillator well reproduces the first three magic numbers but fails above twenty; adding a strong inverted spin-orbit term that drives the stretched j equals l plus one-half level down closes the gaps at 28, 50, 82, and 126. The filled shells couple to zero, so the last unpaired nucleon fixes the ground-state spin and parity, and its single-particle magnetic moment falls on the Schmidt lines. Configuration mixing sets the limits of the extreme single-particle model.\n",{"path":28263,"title":28264,"module":28252,"summary":28265},"\u002Fnuclear-physics\u002Fnuclear-models\u002Fcollective-model-rotations-vibrations","The Collective Model: Rotations, Vibrations, and Deformed Nuclei","Deformed nuclei rotate with energies proportional to I times I plus one, giving the ground-state band its characteristic level ratios, while near-spherical nuclei vibrate in quantized surface phonons that build one- and two-phonon multiplets. The Nilsson model tracks single-particle levels as the well deforms, moments of inertia fall between the rigid and irrotational limits, backbending marks the sudden alignment of a broken pair, and giant resonances are the bulk dipole and quadrupole modes of the whole nucleus.\n",{"path":28267,"title":28268,"module":28269,"summary":28270},"\u002Fnuclear-physics\u002Fradioactive-decay\u002Fdecay-law-modes","Radioactivity and Decay Modes","Radioactive Decay","Unstable nuclei decay at a rate proportional to how many remain, giving the exponential law N(t) = N0 e^(-lambda t) with half-life t = 0.693\u002Flambda. We work through the three common modes: alpha decay as Coulomb-barrier tunneling with the Geiger-Nuttall rule, beta decay whose continuous spectrum demands the neutrino, and gamma de-excitation, and follow a decay chain across the chart of nuclides.\n",{"path":28272,"title":28273,"module":28269,"summary":28274},"\u002Fnuclear-physics\u002Fradioactive-decay\u002Fdecay-kinetics-equilibrium","Serial Decay, the Bateman Equations, and Radioactive Equilibrium","A radioactive parent that decays into a radioactive daughter obeys a coupled pair of rate equations whose solution is the Bateman formula. Depending on the half-life ordering the chain settles into secular equilibrium (equal activities), transient equilibrium (a fixed activity ratio), or no equilibrium. Constant production under irradiation drives the activity toward a saturation value equal to the production rate, competing decay modes split the total decay constant into partial constants, and the natural decay series in secular equilibrium underpin radiometric dating.\n",{"path":28276,"title":28277,"module":28278,"summary":28279},"\u002Fnuclear-physics\u002Falpha-decay\u002Falpha-decay-gamow-theory","Alpha Decay and the Gamow Theory of Tunneling","Alpha Decay","The alpha Q-value turns positive above mass number 150 because the emitted helium-4 is exceptionally tightly bound. Emission proceeds by quantum tunneling through the Coulomb barrier: a WKB integral from the nuclear surface to the outer turning point gives the Gamow factor, and multiplying its penetrability by the assault frequency yields half-lives spanning more than twenty orders of magnitude. The leading term reproduces the Geiger-Nuttall relation, log t½ proportional to the daughter charge over the square root of Q.\n",{"path":28281,"title":28282,"module":28278,"summary":28283},"\u002Fnuclear-physics\u002Falpha-decay\u002Falpha-fine-structure-hindrance","Fine Structure, Angular Momentum, and Hindrance Factors","A single parent emits several alpha groups of slightly different energy, each feeding a distinct level of the daughter, so the alpha spectrum maps the daughter's low-lying states. Emission with orbital angular momentum L raises the barrier by a centrifugal term and is allowed only when angular-momentum and parity selection rules permit. Comparing the measured partial half-life to the Gamow estimate defines a hindrance factor near unity for even-even ground-state transitions and large for odd-A decays that must rearrange the unpaired nucleon.\n",{"path":28285,"title":28286,"module":28287,"summary":28288},"\u002Fnuclear-physics\u002Fbeta-decay\u002Fbeta-decay-energetics-neutrino","Beta Decay Energetics and the Neutrino","Beta Decay and the Weak Interaction","Beta decay converts a neutron into a proton or the reverse, adjusting Z at fixed A along an isobaric mass parabola. We write the three processes (beta-minus, beta-plus, electron capture), reduce every Q-value to a difference of neutral atomic masses, and read the continuous electron spectrum as the fingerprint of a third, nearly massless particle. Pauli's neutrino, its detection by Reines and Cowan, and the endpoint bound on its mass close the lesson.\n",{"path":28290,"title":28291,"module":28287,"summary":28292},"\u002Fnuclear-physics\u002Fbeta-decay\u002Ffermi-theory-beta-decay","Fermi's Theory: Kurie Plots and ft Values","Fermi treated beta decay as a point-contact weak transition and read its rate from the golden rule. The electron spectrum then follows from phase space and the Coulomb Fermi function; the Kurie plot straightens it to a line whose intercept is the endpoint. Integrating the spectrum gives the comparative half-life ft, whose logarithm sorts transitions into superallowed, allowed, and forbidden classes governed by the Fermi and Gamow-Teller selection rules.\n",{"path":28294,"title":28295,"module":28287,"summary":28296},"\u002Fnuclear-physics\u002Fbeta-decay\u002Fweak-interaction-parity-violation","The Weak Interaction and Parity Violation","Beta decay violates mirror symmetry. The Wu experiment on polarized cobalt-60 showed electrons emitted preferentially against the nuclear spin, a pseudoscalar correlation forbidden if parity were conserved. The result fixes the weak charged current as left-handed V minus A, forces neutrinos to be left-handed and antineutrinos right-handed (measured by Goldhaber), and places beta decay within the electroweak theory as W-boson exchange turning a down quark into an up quark.\n",{"path":28298,"title":28299,"module":28287,"summary":28300},"\u002Fnuclear-physics\u002Fbeta-decay\u002Fdouble-beta-decay-neutrino-mass","Double Beta Decay and Neutrino Mass","For even-A isobars the pairing term splits the mass parabola into two curves, and a handful of even-even nuclides sit below their odd-odd neighbor yet above the next even-even one: single beta decay is forbidden but second-order double beta decay is allowed. The two-neutrino mode is a standard-model process with the longest measured lifetimes in nature; the neutrinoless mode would require the neutrino to be its own antiparticle and its rate measures the effective Majorana mass, the sharpest probe of the absolute neutrino mass scale.\n",{"path":28302,"title":28303,"module":28304,"summary":28305},"\u002Fnuclear-physics\u002Fgamma-decay\u002Fgamma-multipole-radiation","Multipole Radiation and Selection Rules","Gamma Decay","Gamma decay carries a nucleus from an excited state to a lower one by emitting a photon of definite angular momentum and parity. We correct the photon energy for nuclear recoil, expand the radiation field into electric and magnetic multipoles, and read off how the transition rate collapses with each increase in multipole order. The Weisskopf single-particle estimates set the scale, and angular-momentum and parity conservation fix which multipole dominates.\n",{"path":28307,"title":28308,"module":28304,"summary":28309},"\u002Fnuclear-physics\u002Fgamma-decay\u002Finternal-conversion-isomers","Internal Conversion and Isomers","A nucleus can shed excitation energy without emitting a photon by handing it directly to an atomic electron. We define the internal-conversion coefficient, trace its growth with atomic number, multipole order, and decreasing energy, and treat the electron-only E0 transitions and internal pair formation. When the lowest allowed multipole is high and the energy low, the gamma rate falls so far that the excited state survives as a metastable isomer.\n",{"path":28311,"title":28312,"module":28304,"summary":28313},"\u002Fnuclear-physics\u002Fgamma-decay\u002Fangular-correlations-mossbauer","Angular Correlations and the Mössbauer Effect","Two gammas emitted in cascade are not independent in direction: detecting the first selects magnetic substates of the intermediate level and makes the second anisotropic, so the correlation function fixes the intermediate spin. The same nuclear resonance that recoil normally destroys is recovered when the emitter is locked in a lattice, giving the Mössbauer effect and its part-in-a-trillion resolution of isomer shifts and hyperfine fields.\n",{"path":28315,"title":28316,"module":28317,"summary":28318},"\u002Fnuclear-physics\u002Fnuclear-reactions\u002Freaction-kinematics-cross-sections","Nuclear Reactions, Fission, and Fusion","Nuclear Reactions","A nuclear reaction X(x, y)Y is governed by its Q value and its cross section, the effective target area for a given process. Splitting the curve of binding energy near iron in either direction releases energy: fission of heavy nuclei by neutron capture and a chain reaction, and fusion of light nuclei that powers the Sun and needs Lawson's density-confinement criterion to be practical.\n",{"path":28320,"title":28321,"module":28317,"summary":28322},"\u002Fnuclear-physics\u002Fnuclear-reactions\u002Fcompound-nucleus-resonances","The Compound Nucleus and Resonance Reactions","Low-energy reactions proceed through a long-lived intermediate state whose decay forgets how it formed. Bohr's independence hypothesis factorizes the cross section into a formation step and a branching ratio, an isolated level gives the single-level Breit-Wigner line shape with total width Γ tied to the lifetime by Γτ = ħ, and at high excitation overlapping levels merge into a statistical continuum described by evaporation spectra and the Hauser-Feshbach average.\n",{"path":28324,"title":28325,"module":28317,"summary":28326},"\u002Fnuclear-physics\u002Fnuclear-reactions\u002Fdirect-reactions-optical-model","Direct Reactions and the Optical Model","A complex optical potential replaces the many-body target by a single particle moving in an average field whose imaginary part removes flux into non-elastic channels, reproducing the diffraction pattern of elastic scattering. Direct reactions bypass the compound nucleus, transferring a nucleon in one step: stripping and pickup deposit or remove a single nucleon, the angle of the first peak in the distorted-wave angular distribution fixes the transferred orbital angular momentum, and its magnitude gives the spectroscopic factor.\n",{"path":28328,"title":28329,"module":28330,"summary":28331},"\u002Fnuclear-physics\u002Ffission\u002Ffission-barrier-dynamics","The Fission Barrier and Fragment Energetics","Nuclear Fission","Fission is the large-amplitude collective deformation of a heavy nucleus into two fragments. The liquid-drop model sets a barrier from the competition between rising surface energy and falling Coulomb energy under quadrupole deformation, with the fissility parameter Z²\u002FA measuring how close a nucleus is to instability. Bohr-Wheeler theory separates spontaneous from neutron-induced fission, the fragment mass yield is double-humped and asymmetric, about 200 MeV is released per event, and shell corrections add a second minimum that produces fission isomers.\n",{"path":28333,"title":28334,"module":28330,"summary":28335},"\u002Fnuclear-physics\u002Ffission\u002Fchain-reactions-reactor-physics","Chain Reactions and Reactor Physics","A self-sustaining chain reaction is a fixed point of neutron bookkeeping: the multiplication factor k counts the neutrons in one generation per neutron in the last, and criticality is k = 1. The four-factor formula tracks a neutron through fast fission, resonance escape, thermal utilization, and reproduction; moderation slows fission neutrons to the thermal energies where the fission cross section is largest; and the small delayed-neutron fraction sets the timescale that makes a reactor controllable. Breeding converts fertile U-238 and Th-232 into new fissile fuel.\n",{"path":28337,"title":28338,"module":28339,"summary":28340},"\u002Fnuclear-physics\u002Ffusion-nucleosynthesis\u002Ffusion-reactions-confinement","Fusion Reactions and Confinement","Fusion and Nucleosynthesis","Light nuclei release energy when they fuse because binding per nucleon rises steeply toward the iron peak, but the Coulomb barrier suppresses the rate at reactor temperatures. The thermonuclear rate is a convolution of the Maxwell distribution with the tunneling probability, sharply peaked at the Gamow energy. The deuterium-tritium reaction has the lowest barrier and largest cross section; sustained energy gain requires the Lawson triple product of density, temperature, and confinement time, reached by magnetic or inertial confinement.\n",{"path":28342,"title":28343,"module":28339,"summary":28344},"\u002Fnuclear-physics\u002Ffusion-nucleosynthesis\u002Fstellar-nucleosynthesis","Stellar Nucleosynthesis","Main-sequence stars burn hydrogen to helium through the proton-proton chain and the CNO cycle, both releasing 26.7 MeV per helium nucleus. Helium burning bridges the mass-5 and mass-8 gaps by the triple-alpha process through the Beryllium-8 and Hoyle resonances, and successive carbon-to-silicon burning stages climb to the iron peak, where fusion stops. The elements beyond iron are built by slow and rapid neutron capture, and the solar neutrino flux confirms the reactions directly.\n",{"path":28346,"title":28347,"module":28339,"summary":28348},"\u002Fnuclear-physics\u002Ffusion-nucleosynthesis\u002Fbig-bang-nucleosynthesis","Big-Bang Nucleosynthesis","In the first three minutes the expanding universe forged the light elements. The weak interaction froze the neutron-to-proton ratio near one in six when the reaction rate fell below the expansion rate, and free-neutron decay lowered it to about one in seven before the deuterium bottleneck broke. Almost every surviving neutron ended in helium-4, fixing the primordial helium mass fraction near 0.25, with trace deuterium, helium-3, and lithium-7. The deuterium abundance measures the cosmic baryon density.\n",{"path":28350,"title":28351,"module":28352,"summary":28353},"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fcharged-particle-stopping-power","Stopping Power and the Range of Charged Particles","Radiation and Applications","A heavy charged particle loses energy in a dense sequence of small Coulomb collisions with atomic electrons, at a rate the Bethe-Bloch formula fixes from the particle's charge and speed and the medium's electron density and mean excitation energy. The rate scales as the inverse square of the speed, so most energy is deposited at the end of the track in the Bragg peak, and integrating the reciprocal rate gives a sharp range. Electrons differ: they also radiate, and above a critical energy bremsstrahlung dominates. Fast particles above the phase velocity of light in the medium emit Cherenkov radiation.\n",{"path":28355,"title":28356,"module":28352,"summary":28357},"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fphoton-neutron-interactions","Interactions of Photons and Neutrons","Photons are removed from a beam in single events, so their intensity falls exponentially with a linear attenuation coefficient built from three processes: the photoelectric effect at low energy, Compton scattering at intermediate energy, and pair production above twice the electron rest energy, each with its own atomic-number and energy dependence. Neutrons carry no charge and interact only with nuclei, moderating by elastic scattering and being captured with a cross section that rises as one over speed away from resonances.\n",{"path":28359,"title":28360,"module":28352,"summary":28361},"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fradiation-detectors","Radiation Detectors and Nuclear Spectroscopy","Every detector converts the energy a radiation deposits into a measurable electrical signal. Gas counters read the ionization directly, in three operating regions set by the applied voltage; scintillators convert the energy to light read out by a photomultiplier; semiconductor detectors collect electron-hole pairs and give the best energy resolution because so many carriers are made per event. The resolution is governed by the number of independent charge carriers, and the pulse-height spectrum of a gamma line shows a full-energy photopeak, a Compton continuum with its edge, and escape peaks.\n",{"path":28363,"title":28364,"module":28352,"summary":28365},"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fdosimetry-radiation-biology","Dosimetry, Radiation Biology, and Protection","Absorbed dose is the energy deposited per unit mass, measured in gray. Equal absorbed doses do unequal biological damage because densely ionizing radiation deposits its energy along short tracks: weighting the dose by a radiation factor gives the equivalent dose, and weighting by tissue sensitivity gives the effective dose, both in sieverts. Deterministic effects have a threshold and a severity that grows with dose; stochastic effects are assumed to follow a linear-no-threshold probability. Natural background dominates the dose to the population, and protection rests on time, distance, and shielding.\n",{"path":28367,"title":28368,"module":28352,"summary":28369},"\u002Fnuclear-physics\u002Fradiation-matter-applications\u002Fnuclear-applications-dating-medicine","Applications — Dating, Analysis, and Nuclear Medicine","Charged particles lose energy continuously and stop at a well-defined range with a Bragg peak, while gamma rays are attenuated exponentially. These interactions define radiation detectors and dosimetry (gray and sievert) and drive the applications: neutron activation analysis, magnetic resonance imaging, PET, and radiometric dating with carbon-14 and long-lived rock clocks.\n",{"path":28371,"title":28372,"module":6,"summary":6},"\u002Fnuclear-physics","Nuclear Physics",{"path":28374,"title":28375,"module":24340,"summary":28376},"\u002Fnatural-language-processing\u002Ffoundations\u002Fwhat-is-nlp","What Is Natural Language Processing?","Natural language processing is the computational treatment of human language: reading it, representing it, and generating it. We set up why the problem is hard — ambiguity at every level, from sound to intent — trace the field from ELIZA's pattern-matching through statistical methods to today's neural models, lay out the linguistic levels and task families the course covers, and fix the vocabulary of tokens, types, and corpora the rest of the notes rely on.\n",{"path":28378,"title":28379,"module":24340,"summary":28380},"\u002Fnatural-language-processing\u002Ffoundations\u002Fregex-and-text-normalization","Regular Expressions and Text Normalization","Before any model touches text, the text has to be found and cleaned. Regular expressions give an algebra for describing string patterns; tokenization, case folding, and stemming turn raw characters into the units a model counts; and byte-pair encoding builds a subword vocabulary that spells out any word. Measuring how far apart two strings are — minimum edit distance — is the next lesson.\n",{"path":28382,"title":28383,"module":24340,"summary":28384},"\u002Fnatural-language-processing\u002Ffoundations\u002Fminimum-edit-distance","Minimum Edit Distance","Much of language processing needs to measure how similar two strings are — a speller ranking corrections, a diff tool, a coreference resolver. Minimum edit distance counts the insertions, deletions, and substitutions that turn one string into another, computed by a dynamic-programming table. We fill the table for intention to execution, backtrace to recover the alignment, and see how the same machinery generalizes to weighted edits, Viterbi, and biological sequence alignment.\n",{"path":28386,"title":28387,"module":24340,"summary":28388},"\u002Fnatural-language-processing\u002Ffoundations\u002Fn-gram-language-models","N-Gram Language Models","A language model assigns a probability to a sequence of words and, equivalently, predicts the next word from its history. The n-gram model makes this tractable by truncating the history to the last few words, estimates the resulting conditional probabilities by counting, and is scored by perplexity. We build the model from the chain rule, work a bigram example on a small corpus, and read perplexity as a branching factor. The next lesson covers the zero counts that break this model and the smoothing that repairs them.\n",{"path":28390,"title":28391,"module":24340,"summary":28392},"\u002Fnatural-language-processing\u002Ffoundations\u002Fsmoothing-and-backoff","Smoothing and Backoff","Every finite corpus is missing good word sequences it simply never saw, so a raw n-gram model assigns them probability zero and breaks. Smoothing repairs the zeros: add-one and add-k shave mass off seen events, backoff and interpolation fall back on shorter contexts, and Kneser-Ney — worked here by hand — replaces raw frequency with how many contexts a word completes. We close on web-scale stupid backoff and the neural models that dissolve the zero problem rather than patch it.\n",{"path":28394,"title":28395,"module":28396,"summary":28397},"\u002Fnatural-language-processing\u002Fclassification\u002Fnaive-bayes-and-sentiment","Naive Bayes and Sentiment Classification","Text Classification","Text classification assigns a category to a document — positive or negative, spam or not, one topic among many. Naive Bayes is a generative solution: apply Bayes' rule, assume the words are conditionally independent given the class, and the winning class is the one maximizing the product of a prior and per-word likelihoods. We train it by counting with add-one smoothing, work a full sentiment example by hand, sharpen it for sentiment (binary counts, negation, lexicons), and place it among the transformer classifiers that came after.\n",{"path":28399,"title":28400,"module":28396,"summary":28401},"\u002Fnatural-language-processing\u002Fclassification\u002Fevaluating-classifiers","Evaluating Classifiers","A trained classifier is only useful once we can measure how good it is. We build the confusion matrix, see why accuracy misleads on unbalanced data, and define precision, recall, and the F-measure that balances them. Multi-class tasks need macro- versus micro-averaging; reliable estimates need cross-validation. We close on statistical significance — the paired bootstrap test for whether one system's lead over another is significant.\n",{"path":28403,"title":28404,"module":28396,"summary":28405},"\u002Fnatural-language-processing\u002Fclassification\u002Flogistic-regression","Logistic Regression","Logistic regression is the discriminative counterpart to naive Bayes: instead of modelling how a document is generated, it learns weights that directly separate the classes. We build it from the sigmoid, derive the cross-entropy loss from maximum likelihood, learn the weights by stochastic gradient descent, regularize to curb overfitting, and generalize to many classes with the softmax. The two-class model is already a one-neuron network, so this is the bridge to neural language models.\n",{"path":28407,"title":28408,"module":28396,"summary":28409},"\u002Fnatural-language-processing\u002Fclassification\u002Fsentiment-and-affect-lexicons","Sentiment and Affect Lexicons","A sentiment lexicon is a list of words annotated with the affective meaning they carry — positive or negative, or scores along valence, arousal, and dominance. We fix what \"emotion\" means (basic-emotion versus dimensional models), survey the standard lexicons, and then build lexicons three ways: by human labeling with best-worst scaling, by semi-supervised induction from seed words over an embedding space, and by supervised learning from starred reviews. We close on connotation frames, which record the sentiment a verb implies about each of its arguments.\n",{"path":28411,"title":28412,"module":28413,"summary":28414},"\u002Fnatural-language-processing\u002Fsemantics\u002Fvector-semantics-and-embeddings","Vector Semantics and Embeddings","Semantics","Vector semantics represents a word's meaning as a point in space, derived from the company the word keeps. This first part builds the count-based side: the distributional hypothesis, co-occurrence matrices in their term-document and word-word forms, cosine as the similarity measure, and the two weightings — tf-idf and PPMI — that fix what raw counts get wrong. The result is a sparse, interpretable vector for every word, and the setup for the dense embeddings of the next lesson.\n",{"path":28416,"title":28417,"module":28413,"summary":28418},"\u002Fnatural-language-processing\u002Fsemantics\u002Fstatic-word-embeddings","Static Word Embeddings: word2vec and After","Count-based vectors are long and sparse; embeddings are the short, dense alternative. This lesson builds them with word2vec's skip-gram and negative sampling — a classifier whose learned weights are the vectors — derives its gradient, and works one update by hand. It then reads relations off the analogy parallelogram, surveys the papers that framed the static-embedding era (word2vec, GloVe, the SGNS-as-PPMI equivalence, fastText, ELMo), and closes on the biases embeddings inherit and the single-vector-per-word ceiling that contextual models break.\n",{"path":28420,"title":28421,"module":28413,"summary":28422},"\u002Fnatural-language-processing\u002Fsemantics\u002Fneural-language-models","Neural Networks and Neural Language Models","A neural network is a stack of units, each a weighted sum passed through a non-linearity — a single unit on its own is logistic regression. We build the network up from that unit: the activation functions that give it power, the XOR problem that forces a hidden layer, the feedforward forward pass in matrix form, and the Bengio-style feedforward neural language model that concatenates word embeddings and predicts the next word with a softmax. Training is cross-entropy minimized by gradient descent, with backpropagation supplying the gradient. Embeddings let the model share statistical strength across similar words, avoiding the sparsity that limits n-gram models.\n",{"path":28424,"title":28425,"module":24806,"summary":28426},"\u002Fnatural-language-processing\u002Fsequences\u002Fsequence-labeling","Sequence Labeling: POS and NER","Sequence labeling assigns one tag to every token in a sentence. This first part sets up the task through its two canonical cases — part-of-speech tagging over the Penn Treebank tagset, and named-entity recognition reframed as token labeling with the BIO scheme — then builds the hidden Markov model, the classic probabilistic tagger. The HMM tags by Bayesian inference: transition and emission probabilities under two Markov assumptions, reducing tagging to an argmax over tag sequences. That argmax is exponential to enumerate, which sets up the Viterbi decoder, the CRF, and neural taggers of the next lesson.\n",{"path":28428,"title":28429,"module":24806,"summary":28430},"\u002Fnatural-language-processing\u002Fsequences\u002Fcrfs-and-neural-taggers","Viterbi Decoding, CRFs, and Neural Taggers","The HMM reduced tagging to an argmax over exponentially many tag sequences. This lesson builds the decoder that makes it tractable — the Viterbi dynamic program, worked through a full numeric trace on real WSJ probabilities — then keeps that same decoder while replacing the HMM's rigid tables. The linear-chain conditional random field is a discriminative log-linear model whose global feature functions can inspect any part of the input, which is why CRFs win for NER. Finally it traces the shift to neural taggers (biLSTM-CRF, character-aware NER, ELMo), where hand-built features become learned representations while the Viterbi decoder carries over unchanged.\n",{"path":28432,"title":28433,"module":24806,"summary":28434},"\u002Fnatural-language-processing\u002Fsequences\u002Frnns-and-lstms","RNNs and LSTMs","A feedforward neural language model sees a fixed window of words and can look no further back. The recurrent neural network removes that limit: it carries a hidden state across time, so each word is read in the context of everything before it. We build the RNN from its one recurrent equation, use it as a language model, train it by backpropagation through time, and diagnose the vanishing-gradient problem that makes plain RNNs forget. The LSTM fixes the forgetting with a cell state and three gates, and the encoder-decoder stacks two RNNs into a sequence-to-sequence model — and its single-vector bottleneck is the problem attention was invented to remove.\n",{"path":28436,"title":28437,"module":25288,"summary":28438},"\u002Fnatural-language-processing\u002Ftransformers\u002Ftransformers-and-attention","Transformers and Self-Attention","Recurrence forced language models to read one word at a time and to squeeze every dependency through a chain of hidden states. Self-attention removes the recurrence: at every layer each position compares itself to every other and reads a weighted mixture of them, in a single parallel step. This first part builds the attention operation from the ground up — the soft lookup, queries and keys and values, the scaled dot-product, the numeric trace, the matrix form, and the causal mask — and sets up the full transformer architecture that follows.\n",{"path":28440,"title":26991,"module":25288,"summary":28441},"\u002Fnatural-language-processing\u002Ftransformers\u002Fthe-transformer-architecture","This part takes the scaled dot-product attention of the previous lesson and assembles the full transformer architecture around it: multi-head attention so several relations can be read at once, the transformer block of residual connections and layer norm that makes deep stacks trainable, positional embeddings that restore word order, the decoder-only language model, and the encoder, decoder, and encoder-decoder shapes — closing with the 2017 paper and the pre-norm, FlashAttention, and RoPE refinements that scaled it up.\n",{"path":28443,"title":27099,"module":25288,"summary":28444},"\u002Fnatural-language-processing\u002Ftransformers\u002Flarge-language-models","A large language model is a decoder-only transformer trained on one objective — predict the next token. This first part assembles the inference side: the language-modeling head that turns a hidden state into a distribution over the vocabulary, autoregressive generation, and the decoding strategies — greedy, beam, and sampling with temperature, top-k, and nucleus — that read text back out of that distribution. Training the distribution at web scale comes next.\n",{"path":28446,"title":28447,"module":25288,"summary":28448},"\u002Fnatural-language-processing\u002Ftransformers\u002Fllm-pretraining-and-scaling","Large Language Models: Pretraining and Scaling","A language model's next-token distribution is only as good as the parameters behind it. This part is where those parameters come from: self-supervised pretraining on web-scale text with teacher forcing and cross-entropy, the scaling laws that make test loss a predictable power law in parameters, data, and compute, the KV cache that keeps long-context inference affordable, and how a finished model is evaluated by perplexity and benchmarks — closing with the Kaplan, Chinchilla, GPT-3, and emergence papers behind the scaling story.\n",{"path":28450,"title":28451,"module":25288,"summary":28452},"\u002Fnatural-language-processing\u002Ftransformers\u002Ffine-tuning-and-prompting","Fine-Tuning and Prompting","A pretrained transformer is a general-purpose knowledge source; a task is what you do with it. There are two ways to adapt one, and this first part covers the one that updates the weights: fine-tuning. A bidirectional encoder like BERT is pretrained by masked language modeling, then a small task head is bolted on and the whole thing is trained on labelled data for classification, sequence labeling, or span-based question answering — with parameter-efficient variants (adapters, LoRA) that touch only a sliver of the weights. Prompting, the family that leaves the weights frozen, comes next.\n",{"path":28454,"title":28455,"module":25288,"summary":28456},"\u002Fnatural-language-processing\u002Ftransformers\u002Fprompting-and-alignment","Prompting and Alignment","Fine-tuning adapts a model by changing its weights. The second family of adaptation changes nothing: a large frozen model performs a task from an instruction and a few examples placed in its context. This part covers prompting and in-context learning, chain-of-thought that elicits reasoning, and the two training stages — instruction tuning and RLHF — that turn a fluent base predictor into an aligned assistant, closing with the BERT, LoRA, chain-of-thought, InstructGPT, and retrieval-augmentation papers behind the modern adaptation pipeline.\n",{"path":28458,"title":28459,"module":28460,"summary":28461},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fconstituency-parsing","Constituency Parsing","Linguistic Structure","A constituency parse groups a sentence into nested phrases described by a context-free grammar. We build the CFG formalism, read the phrase structure of English off a treebank, confront the structural ambiguity that makes parsing hard, convert to Chomsky normal form, and then solve it with CKY — the dynamic-programming chart that fills a triangular table bottom-up. Probabilistic and neural span parsers, evaluation, and shallow parsing follow in the companion lesson.\n",{"path":28463,"title":28464,"module":28460,"summary":28465},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fcky-scoring-and-evaluation","CKY Scoring, Evaluation, and Shallow Parsing","The CKY chart returns every parse but does not say which is correct. Disambiguation needs a score on trees. This lesson attaches probabilities to a grammar (the PCFG and lexicalization), replaces the grammar with a neural span scorer over a pretrained encoder, states the self-attentive results that made it the state of the art, evaluates parsers against a treebank with PARSEVAL, and closes with chunking and shallow parsing for tasks that need only the flat phrases.\n",{"path":28467,"title":28468,"module":28460,"summary":28469},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fdependency-parsing","Dependency Parsing","A dependency parse throws away phrases and keeps only directed, labeled arcs from heads to their dependents, so the subject and object of a verb hang off the verb directly. We fix the formalism (rooted trees, typed Universal-Dependency relations, projectivity), then build the first parser family: transition-based arc-standard and arc-eager parsing, a greedy stack-and-buffer machine trained from an oracle. Graph-based and neural dependency parsing follow in the companion lesson.\n",{"path":28471,"title":28472,"module":28460,"summary":28473},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fgraph-based-and-neural-dependency-parsing","Graph-Based and Neural Dependency Parsing","Greedy transition parsing commits locally; the graph-based family scores whole trees instead. This lesson scores every candidate head-dependent edge and extracts the maximum spanning tree with Chu-Liu\u002FEdmonds, develops the biaffine neural scorer that made graph-based parsing the accuracy leader, evaluates parsers with the unlabeled and labeled attachment scores (UAS and LAS), and closes on where the two parser families sit and what they feed downstream.\n",{"path":28475,"title":28476,"module":28460,"summary":28477},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fword-senses-and-wsd","Word Senses and Disambiguation","A word is not an atom of meaning: \"bass\" names a fish, a voice, and an instrument, and one static embedding blurs them into a single point. This lesson pulls those senses apart. We define polysemy and the relations that organize senses — synonymy, antonymy, hyponymy, meronymy — build them into WordNet's synset graph, measure similarity along that graph, and then solve the core of word sense disambiguation: the most-frequent-sense baseline, the Lesk gloss-overlap algorithm, feature-based classifiers, and the nearest-neighbor method over BERT embeddings. WSD variants, embeddings, and evaluation follow in the companion lesson.\n",{"path":28479,"title":28480,"module":28460,"summary":28481},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fwsd-in-practice-and-induction","WSD in Practice and Word Sense Induction","Beyond core word sense disambiguation lie the variants and loose ends: the sense-inventory-free Word-in-Context task, retrofitting static embeddings to a thesaurus, discovering senses without a fixed inventory (word sense induction), the gloss-aware and bi-encoder neural systems that hold the state of the art, and how WSD and its cousins are evaluated. Together they connect one-vector-per-word embeddings to sense-aware contextual representations.\n",{"path":28483,"title":28484,"module":28460,"summary":28485},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fsemantic-roles-and-information-extraction","Semantic Roles and Information Extraction","Semantic roles answer \"who did what to whom\" for a single event, abstracting away the syntax that expresses it. We show why syntax alone is not enough, generalize over diathesis alternations with thematic roles, number a predicate's arguments with PropBank and group predicates into frames with FrameNet, tag each argument automatically with semantic role labeling, and factor predicates into primitives. Information extraction scales the idea to a corpus in the companion lesson.\n",{"path":28487,"title":28488,"module":28460,"summary":28489},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Frelations-events-and-templates","Relations, Events, and Templates","Semantic roles answer \"who did what\" for one predicate; information extraction scales the idea to a whole corpus. This lesson turns unstructured text into structured data: relation extraction pulls entity-relation-entity triples out of sentences by patterns, supervision, and distant supervision; event and temporal extraction place those facts on a timeline; and template filling and knowledge-base population assemble them into a database a downstream system can query.\n",{"path":28491,"title":28492,"module":28460,"summary":28493},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fcoreference-and-discourse","Coreference and Discourse","A text is more than a bag of sentences: entities recur under different names. Coreference resolution links every mention to the discourse entity it evokes — the linguistic background of pronouns, definite NPs, and names; mention detection; the mention-pair, mention-ranking, and entity-based architectures; a neural end-to-end span model that scores candidate antecedents; features, evaluation by the CoNLL F1, gender bias, and the neural coreference lineage. Discourse coherence follows in the companion lesson.\n",{"path":28495,"title":28496,"module":28460,"summary":28497},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fcoherence-and-discourse-structure","Coherence and Discourse Structure","Coherence is what makes a run of sentences a discourse rather than an arbitrary collection. This lesson develops coherence relations and Rhetorical Structure Theory trees, discourse-structure parsing, Centering and the entity grid for entity-based coherence, and representation-learning models of local coherence, measured in part over the coreference chains recovered in the companion lesson.\n",{"path":28499,"title":28500,"module":28460,"summary":28501},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Flogical-semantics","Logical Representations of Meaning","A meaning representation turns a sentence into a formal structure a machine can check against a world and reason over. We set the desiderata a good representation must meet, ground truth in a model, build up first-order logic for sentences with its connectives, quantifiers, and inference, and reify events with the neo-Davidsonian event variable to escape fixed predicate arity. The compositional lambda calculus, quantifier scope, and description logics follow in the companion lesson.\n",{"path":28503,"title":28504,"module":28460,"summary":28505},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fcompositional-semantics-and-description-logics","Compositional Semantics and Description Logics","How do you compute a logical form from a sentence automatically? This lesson builds the compositional machinery: the lambda calculus that assembles a formula from a parse tree one beta-reduction at a time, the quantifier-scope ambiguity a single syntax tree leaves open, and the decidable description logics — TBox, ABox, subsumption, role restrictions — behind the Web Ontology Language, closing with how the map from string to logical form can be learned.\n",{"path":28507,"title":28508,"module":28460,"summary":28509},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fsemantic-parsing","Semantic Parsing","Turning a sentence into a structured, executable meaning, the grammar-based way. We take the logical forms defined earlier and build them compositionally: a rule-based parser that walks a syntax tree applying lambda terms, then Combinatory Categorial Grammar (CCG), which fuses syntax and semantics so one lexicalized derivation produces both — including supertagging and A* parsing. Learned and neural semantic parsers follow in the companion lesson.\n",{"path":28511,"title":28512,"module":28460,"summary":28513},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fneural-semantic-parsing","Learned and Neural Semantic Parsing","Hand-writing a lexicon of lambda terms does not scale, so this lesson learns the parser instead. We cover the two supervision regimes (from logical forms and from denotations), Abstract Meaning Representation as a rooted concept graph, neural sequence-to-sequence parsing with constrained decoding and copy mechanisms, executable text-to-SQL and knowledge-based question answering, the practical systems that made learned parsers accurate, and how the task is evaluated.\n",{"path":28515,"title":28516,"module":28460,"summary":28517},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Finformation-extraction","Information Extraction","Information extraction turns free text into a database, and the first step is relation extraction: pulling entity-relation-entity triples out of sentences. We cover all five families — hand-built patterns, supervised classifiers, semi-supervised bootstrapping, distant supervision, and unsupervised Open IE — with worked bootstrapping and distant-supervision traces, then the neural and LLM systems that extended them. Times, events, and templates follow in the companion lesson.\n",{"path":28519,"title":28520,"module":28460,"summary":28521},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Ftimes-events-and-templates","Extracting Times, Events, and Templates","Once relation extraction has produced typed triples, the information-extraction pipeline still has to place facts in time and assemble them into records. This lesson detects and normalizes temporal expressions to ISO 8601 values, detects events and orders them on a timeline with the 13 Allen relations, and fills slot-and-filler templates — flat and hierarchical — for stereotyped situations, closing the loop from text to a queryable database.\n",{"path":28523,"title":28524,"module":28460,"summary":28525},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fdiscourse-coherence","Discourse Coherence","A text is more than a set of sentences. What binds a run of sentences into a discourse is coherence, and one of its sources is structured relations between clauses. This lesson develops relational coherence — RST and the PDTB models of coherence relations — and discourse-structure parsing: EDU segmentation and shift-reduce RST parsing, then PDTB relation classification. Entity-based and global coherence follow in the companion lesson.\n",{"path":28527,"title":28528,"module":28460,"summary":28529},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fentity-based-and-global-coherence","Entity-Based and Global Coherence","A text coheres not only through relations between clauses but by staying about the same entities and the same topic, and by obeying the macro-structure of its genre. This lesson develops Centering Theory and the entity grid for entity-based coherence, representation-learning models of local coherence, and global coherence — topic segmentation, narrative and argumentation structure, and scientific discourse — then the neural models that learn each.\n",{"path":28531,"title":28532,"module":28460,"summary":28533},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Fconstituency-grammars","Constituency Grammars","A constituency grammar is the declarative theory of sentence structure that a parser operates on. We build the context-free grammar formalism from its four parts, show how derivations become parse trees, and work through the phrase structure of English — noun phrases, verb phrases and their subcategorization frames, agreement, coordination, and long-distance dependencies. The treebank, normal-form, and lexicalized views follow in the companion lesson.\n",{"path":28535,"title":28536,"module":28460,"summary":28537},"\u002Fnatural-language-processing\u002Flinguistic-structure\u002Ftreebanks-and-lexicalized-grammars","Treebanks and Lexicalized Grammars","Where does a grammar come from, and how is it prepared for a parser? We read a context-free grammar off the Penn Treebank, normalize it to Chomsky Normal Form for the CKY chart, then invert the phrase-structure emphasis with lexicalized grammars — Combinatory Categorial Grammar and its slash categories — and close with the grammar's fate in the neural era: span scoring, self-attention, and grammar induction.\n",{"path":28539,"title":28540,"module":27087,"summary":28541},"\u002Fnatural-language-processing\u002Fapplications\u002Fmachine-translation","Machine Translation","Machine translation is the task that built the modern toolkit: the encoder-decoder was invented for it, attention was invented to fix its fixed-context bottleneck, and both were later folded into the general transformer. We work through why translation is hard (word order, morphology, lexical and structural divergences), the sequence-to-sequence model and its attention mechanism, transformer-based NMT with cross-attention, subword tokenization with a shared vocabulary, beam-search decoding, and evaluation by BLEU and its successors chrF, BERTScore, and COMET — closing on multilingual and low-resource translation and backtranslation.\n",{"path":28543,"title":28544,"module":27087,"summary":28545},"\u002Fnatural-language-processing\u002Fapplications\u002Fmachine-translation-decoding-and-evaluation","Machine Translation: Decoding, Evaluation, and Scale","Having built the transformer translation model, we now decode from it and measure the output. Beam search turns the decoder's per-step distributions into a single output string; length normalization keeps it from favoring short translations. We then score translations automatically — BLEU with its n-gram precision, clipping, and brevity penalty, worked through by hand, then its successors chrF, BERTScore, and COMET — and close on the parts of MT that scale beyond one language pair: multilingual and low-resource translation, backtranslation, gender bias, and the lineage from the Transformer to massively multilingual models like NLLB-200.\n",{"path":28547,"title":28548,"module":27087,"summary":28549},"\u002Fnatural-language-processing\u002Fapplications\u002Fquestion-answering","Question Answering","A question-answering system takes a natural-language question and returns an answer, not a ranked list of documents. Almost every modern system is built on one pattern: retrieve then read. We start with the information-retrieval machinery that finds candidate text — tf-idf and BM25 term weighting, a worked ranking example, the inverted index, and dense embedding retrieval — then build the retriever-reader pipeline that extracts an answer span with BERT and trace a full retrieve-and-read example end to end.\n",{"path":28551,"title":28552,"module":27087,"summary":28553},"\u002Fnatural-language-processing\u002Fapplications\u002Fquestion-answering-knowledge-and-llms","Question Answering: Knowledge Bases and Language Models","The retrieve-and-read pipeline extracts an answer span from prose, but not all knowledge lives in prose. This part covers the rest of the QA stack: entity linking (Wikification) that grounds a question's entities to a knowledge base, knowledge-based QA by semantic parsing a question into an executable query, and the modern default — closed-book QA and retrieval-augmented generation with a large language model — closing on the DPR\u002FRAG\u002Ffusion-in-decoder lineage and how factoid answers are scored by exact match and F1.\n",{"path":28555,"title":28556,"module":27087,"summary":28557},"\u002Fnatural-language-processing\u002Fapplications\u002Fdialogue-and-chatbots","Dialogue and Chatbots","Conversation is the most natural interface to a machine and one of the hardest to build. We set up what makes human dialogue work — turns, speech acts, grounding, and the local structure of adjacency pairs — then trace the two traditions that answer it: chatbots built to chat (ELIZA's pattern-matching, corpus retrieval, and seq2seq generation with its blandness problem) and task-oriented systems built to get something done (the GUS frame-and-slot architecture and the modern NLU \u002F state-tracker \u002F policy \u002F NLG pipeline that accumulates a frame across turns).\n",{"path":28559,"title":28560,"module":27087,"summary":28561},"\u002Fnatural-language-processing\u002Fapplications\u002Fdialogue-systems-and-assistants","Dialogue Systems: LLM Assistants, Evaluation, and Design","Two dialogue traditions — chatbots built to chat and task-oriented frame systems built to get something done — met in the aligned LLM assistant. Instruction tuning plus RLHF fold chit-chat and task dialogue into one model; the LaMDA \u002F InstructGPT \u002F ChatGPT lineage fills in how. The lesson then turns to evaluation (human ratings and acute-eval for chatbots, task success and slot error rate for task systems), user-centered design with Wizard-of-Oz prototyping, and the ethical stakes of building agents people talk to.\n",{"path":28563,"title":28564,"module":27087,"summary":28565},"\u002Fnatural-language-processing\u002Fapplications\u002Ftext-summarization","Text Summarization","Summarization compresses a document to its essential meaning, by either selecting sentences to keep (extractive) or writing new ones (abstractive). This part fixes the task and its flavors — single vs. multi-document, generic vs. query-focused, extractive vs. abstractive — then works through extractive summarization in full: scoring by position and centrality, the TextRank\u002FLexRank graph algorithm run as PageRank over a sentence-similarity graph with a worked iteration, and supervised sentence selection.\n",{"path":28567,"title":28568,"module":27087,"summary":28569},"\u002Fnatural-language-processing\u002Fapplications\u002Fabstractive-summarization-and-evaluation","Abstractive Summarization and Evaluation","Extractive methods can only reuse the source's own sentences; to compress within a sentence or paraphrase, a summarizer has to generate. This part covers abstractive summarization: the sequence-to-sequence approach, the pointer-generator's copy switch and coverage mechanism, pretrained summarizers (BART, PEGASUS) and zero-shot LLM prompting, the long-document and factuality problems, and ROUGE evaluation with a worked example and its limits — closing on the abstractive lineage from See 2017 through faithfulness metrics.\n",{"path":28571,"title":28572,"module":28573,"summary":28574},"\u002Fnatural-language-processing\u002Fspeech\u002Fphonetics","Phonetics","Speech","Before a recognizer can read speech it has to know what speech is. This first part covers the linguistic substrate: phones and their transcription in the IPA and ARPAbet; articulatory phonetics — how the vocal tract shapes airflow into consonants and vowels; and prosody — stress, tune, and the F0 contour. The acoustic side — the waveform, its spectrum, formants, and the spectrogram — is the second part.\n",{"path":28576,"title":28577,"module":28573,"summary":28578},"\u002Fnatural-language-processing\u002Fspeech\u002Facoustic-phonetics","Acoustic Phonetics","Articulation is the cause; the acoustic signal is the effect, and the effect is all a microphone ever gets. This part follows the sound out of the mouth: waves, sampling and the Nyquist limit, F0 and the pitch track, the mel scale, the spectrum and Fourier analysis, the source-filter model that explains why each vowel carries its own formants, and the spectrogram the log-mel front end of every ASR system sits directly on top of — closing with neural TTS, wav2vec, HuBERT, and Whisper, where phonetics went in neural speech.\n",{"path":28580,"title":28581,"module":28573,"summary":28582},"\u002Fnatural-language-processing\u002Fspeech\u002Fautomatic-speech-recognition","Automatic Speech Recognition","Speech recognition maps an acoustic waveform to a string of words, and once the waveform is turned into a sequence of log-mel spectrogram frames the problem is the same sequence-to-sequence transduction the rest of the course already solved. This first part builds the feature front end (framing, the DFT, the mel filterbank, the log), then the modern architectures: the attention-based encoder-decoder, the CTC alignment trick that collapses repeated and blank frames, and RNN-T for streaming. Training-data advances, evaluation, TTS, and the other speech tasks come next.\n",{"path":28584,"title":28585,"module":28573,"summary":28586},"\u002Fnatural-language-processing\u002Fspeech\u002Fasr-evaluation-and-applications","ASR Evaluation and Speech Applications","A recognizer turns a waveform into text; this part scores that text and puts the same machinery to other uses. It opens with the self-supervised and weakly- supervised systems (wav2vec 2.0, HuBERT, Whisper) that made ASR error rates fall. Word error rate reuses the edit distance from the first module, run over words. Text-to-speech runs the whole pipeline in reverse — text to mel spectrogram to waveform. And a family of smaller tasks — wake-word detection, speaker recognition and diarization, language identification — reuse the same log-mel front end without the decoder.\n",{"path":28588,"title":28589,"module":6,"summary":6},"\u002Fnatural-language-processing","Natural Language Processing",{"path":28591,"title":28592,"module":24340,"summary":28593},"\u002Fparticle-physics\u002Ffoundations\u002Fhistorical-overview-particle-zoo","From the Electron to the Particle Zoo","A timeline of the subject, from J. J. Thomson's electron in 1897 to the Higgs boson in 2012. The electron, photon, nucleus, proton, and neutron gave a tidy picture that Yukawa's meson prediction and the muon–pion confusion complicated; strange particles in cosmic rays and the accelerator-era flood of hadrons then produced a \"particle zoo\" that only the quark model organized.\n",{"path":28595,"title":28596,"module":24340,"summary":28597},"\u002Fparticle-physics\u002Ffoundations\u002Fparticle-physics-basic-concepts","Basic Concepts and Particle Classification","Every particle has an antiparticle of equal mass and opposite charge, a consequence of the Dirac equation confirmed by the positron. Feynman diagrams track interactions in spacetime; the material particles sort into leptons and the composite hadrons built from quarks, with baryons carrying three quarks and mesons a quark-antiquark pair.\n",{"path":28599,"title":28600,"module":24340,"summary":28601},"\u002Fparticle-physics\u002Ffoundations\u002Ffundamental-interactions-force-carriers","Fundamental Interactions and Force Carriers","Four interactions account for every force in nature: strong, electromagnetic, weak, and gravitational, in decreasing strength. Each is carried by a boson — the gluon, photon, W and Z, and the graviton — with a range fixed by the carrier's mass through the Yukawa relation, and a coupling constant that itself varies with distance.\n",{"path":28603,"title":28604,"module":28605,"summary":28606},"\u002Fparticle-physics\u002Funits-kinematics\u002Fnatural-units-and-scales","Natural Units and Scales","Units and Kinematics","Setting $\\hbar = c = 1$ collapses mass, momentum, and energy into a single unit, the GeV, and turns lengths and times into inverse energies through the conversion $\\hbar c = 197.3$ MeV·fm. This lesson fixes the natural-unit conventions used for the rest of the course, converts cross sections between barns and GeV$^{-2}$, and shows how to restore factors of $\\hbar$ and $c$ by dimensional analysis.\n",{"path":28608,"title":28609,"module":28605,"summary":28610},"\u002Fparticle-physics\u002Funits-kinematics\u002Ffour-vectors-invariant-mass","Four-Vectors and Invariant Mass","The energy and momentum of a particle form a four-vector whose square is the frame-independent quantity $p^2 = m^2$. This lesson develops the metric and four-vector products, the invariant mass of a multiparticle system, the center-of-momentum and laboratory frames, and the description of collinear boosts by rapidity, whose additivity replaces the awkward velocity-addition law.\n",{"path":28612,"title":28613,"module":28605,"summary":28614},"\u002Fparticle-physics\u002Funits-kinematics\u002Fdecay-scattering-kinematics-mandelstam","Decay, Scattering, and Mandelstam Variables","Two-body decay in the rest frame fixes the daughter momenta from the three masses alone; production thresholds follow from the minimum invariant mass. This lesson works both, then introduces the Mandelstam invariants $s$, $t$, $u$ for $2\\to2$ scattering, proves the identity $s+t+u=\\sum m_i^2$, and maps the physical regions and the crossing that relates channels.\n",{"path":28616,"title":28617,"module":28605,"summary":28618},"\u002Fparticle-physics\u002Funits-kinematics\u002Fcross-sections-golden-rule","Cross Sections and the Golden Rule","The cross section measures how often a scattering happens and the decay width how fast a particle disintegrates. This lesson defines both, relates event rate to luminosity through $R=\\mathcal L\\,\\sigma$ and lifetime to width through $\\tau=\\hbar\u002F\\Gamma$, and states Fermi's golden rule with Lorentz-invariant phase space, giving the master formulas that turn an amplitude $\\mathcal M$ into a measurable rate for $1\\to2$ decay and $2\\to2$ scattering.\n",{"path":28620,"title":28621,"module":28622,"summary":28623},"\u002Fparticle-physics\u002Fsymmetries\u002Fconservation-laws-symmetries","Conservation Laws and Symmetries","Symmetries and Conservation Laws","Which decays occur is decided by conservation laws, each tied by Noether's theorem to a symmetry of physical law. Energy, charge, baryon number, and lepton number are conserved universally; strangeness, isospin, and parity hold in the strong and electromagnetic interactions but break in the weak one, whose parity and CP violation distinguish matter from antimatter.\n",{"path":28625,"title":28626,"module":28622,"summary":28627},"\u002Fparticle-physics\u002Fsymmetries\u002Fdiscrete-symmetries-cpt","Discrete Symmetries — C, P, T, and CPT","Parity reflects space, charge conjugation swaps particle for antiparticle, and time reversal runs the clock backward. Each assigns multiplicative quantum numbers that act as selection rules — intrinsic parities, the photon's C = −1, the C-parity argument fixing the pion's two-photon decay. Their product CPT is a theorem of any local relativistic field theory, forcing particle and antiparticle to share mass and lifetime.\n",{"path":28629,"title":28630,"module":28622,"summary":28631},"\u002Fparticle-physics\u002Fsymmetries\u002Fparity-violation-weak","Parity Violation and the Weak Force","The tau–theta puzzle forced a choice: two particles with identical mass but opposite parity, or one particle whose decay violates parity. Lee and Yang proposed the latter, Wu's polarized cobalt-60 confirmed it, and the violation proved maximal. The charged weak force couples only to left-handed chirality — the Goldhaber experiment showed the neutrino is left-handed — which is why the mirror image of a weak decay is something nature never produces.\n",{"path":28633,"title":28634,"module":28622,"summary":28635},"\u002Fparticle-physics\u002Fsymmetries\u002Fsu2-su3-flavor-symmetry","Isospin, SU(2), and Flavor SU(3)","The near-equal masses of the proton and neutron, and of the three pions, signal a continuous internal symmetry of the strong force: isospin, an SU(2) whose ladder operators move between the members of a multiplet. Adding strangeness enlarges it to an approximate SU(3) flavor symmetry, and the Gell-Mann–Nishijima relation Q = I3 + Y\u002F2 places every hadron on a weight diagram in the isospin–hypercharge plane — the language in which the quark model is written.\n",{"path":28637,"title":28638,"module":28639,"summary":28640},"\u002Fparticle-physics\u002Fquark-model\u002Feightfold-way-su3","The Eightfold Way and SU(3) Flavor","The Quark Model","Gell-Mann and Ne'eman's classification of the hadrons into geometric multiplets, read as representations of an approximate flavor SU(3). The fundamental triplet (u, d, s) and its antitriplet combine into the meson nonet from 3⊗3̄ = 8⊕1 and the baryon octet and decuplet from 3⊗3⊗3, and the empty corner of the decuplet forecast the Ω⁻.\n",{"path":28642,"title":28643,"module":28639,"summary":28644},"\u002Fparticle-physics\u002Fquark-model\u002Fmeson-spectroscopy","Meson Multiplets and Quantum Numbers","Mesons as quark–antiquark bound states. The spin singlet and triplet, orbital excitations, and the assignment of J^PC from the quark spins and orbital angular momentum, giving the pseudoscalar and vector nonets. The η–η' and ω–φ mixing problems, and the charmonium and bottomonium spectra read as heavy-quark positronium.\n",{"path":28646,"title":28647,"module":28639,"summary":28648},"\u002Fparticle-physics\u002Fquark-model\u002Fbaryon-spectroscopy","Baryon Multiplets, Spin, and the Color Puzzle","Baryons as three-quark states, with a wavefunction factored into space, spin, flavor, and color. The spin-3\u002F2 Δ⁺⁺ = uuu forces a totally symmetric state that the Pauli principle forbids, and the resolution is an antisymmetric color factor — the first evidence for color. The octet and decuplet spin content, and baryon magnetic moments as a quantitative test of the model.\n",{"path":28650,"title":28651,"module":28639,"summary":28652},"\u002Fparticle-physics\u002Fquark-model\u002Fcolor-confinement-exotics","Color, Confinement, and Exotic Hadrons","Color as the gauged SU(3) charge, and the requirement that every physical hadron be a color singlet — which selects q-qbar mesons and qqq baryons as the simplest states. The R-ratio of e⁺e⁻ annihilation measures three colors directly. Beyond the simplest singlets lie glueballs, tetraquarks, and pentaquarks, and the recent XYZ states, read as either compact multiquarks or loose hadronic molecules.\n",{"path":28654,"title":28655,"module":28656,"summary":28657},"\u002Fparticle-physics\u002Frelativistic-wave-equations\u002Fklein-gordon-equation","The Klein-Gordon Equation","Relativistic Wave Equations","Quantizing the relativistic energy relation $E^2 = p^2 + m^2$ produces the Klein-Gordon equation for a scalar field. Its plane-wave solutions come in positive- and negative-energy branches, and the conserved density it supplies is not positive-definite — the two difficulties that first drove physicists to seek a first-order equation. The static Klein-Gordon equation with a point source gives the Yukawa potential, and the free equation gives the scalar propagator that later modules attach to exchanged lines.\n",{"path":28659,"title":28660,"module":28656,"summary":28661},"\u002Fparticle-physics\u002Frelativistic-wave-equations\u002Fdirac-equation-spinors","The Dirac Equation and Spinors","Dirac demanded a wave equation first order in time to fix the Klein-Gordon density problem. Factorizing $E^2 = p^2 + m^2$ into a linear form forces the coefficients to be anticommuting matrices — the gamma matrices of the Clifford algebra — so the wavefunction becomes a four-component spinor. The plane-wave solutions split into two particle and two antiparticle states, spin appears automatically with the correct $g = 2$ magnetic moment, and the chirality projectors that the weak interaction later needs fall straight out of the fifth gamma matrix.\n",{"path":28663,"title":28664,"module":28656,"summary":28665},"\u002Fparticle-physics\u002Frelativistic-wave-equations\u002Fantiparticles-hole-theory","Antiparticles and Hole Theory","The negative-energy solutions of the Dirac equation refuse to go away, so they must mean something. Dirac read them as a filled sea of occupied negative-energy states whose holes are positive-energy antiparticles, predicting the positron before its discovery. The picture works for fermions but not bosons, and the Feynman-Stückelberg interpretation replaces it: an antiparticle is a negative-energy solution propagating backward in time, equivalent to a positive-energy antiparticle going forward. Crossing symmetry ties incoming particles to outgoing antiparticles in a single amplitude.\n",{"path":28667,"title":28668,"module":28669,"summary":28670},"\u002Fparticle-physics\u002Fqed\u002Ffeynman-rules-qed","Feynman Rules for QED","Quantum Electrodynamics","Quantum electrodynamics computes a process by summing diagrams, each a term in a power series in the coupling. Every diagram translates into an amplitude by a fixed dictionary: spinors and polarization vectors for external lines, propagators for internal lines, and the vertex factor $ie\\gamma^\\mu$ for each photon-fermion junction. Squaring the amplitude and feeding it to Fermi's golden rule produces a cross section or decay rate, with each extra vertex costing one power of $\\alpha$.\n",{"path":28672,"title":28673,"module":28669,"summary":28674},"\u002Fparticle-physics\u002Fqed\u002Fqed-tree-processes","Tree-Level QED Processes","The Feynman rules become numbers on the reference reactions of QED. Muon pair production $e^+e^-\\to\\mu^+\\mu^-$ sets the scale with its $1+\\cos^2\\theta$ distribution and $4\\pi\\alpha^2\u002F3s$ total cross section, and its ratio to hadron production counts colors. Compton scattering gives the Klein-Nishina formula and the Thomson limit; Bhabha scattering shows $s$- and $t$-channel interference. Casimir's trick turns every spin-averaged square into a trace of gamma matrices.\n",{"path":28676,"title":28677,"module":28669,"summary":28678},"\u002Fparticle-physics\u002Fqed\u002Frenormalization-running-coupling","Renormalization and the Running Coupling","Beyond tree level, QED loops diverge. The three primitive one-loop diagrams — vacuum polarization, electron self-energy, and vertex correction — carry ultraviolet divergences that regularization exposes as logarithms of a cutoff. Renormalization absorbs them into the measured mass, charge, and field normalization, leaving finite predictions. The surviving physical content is that the coupling runs: vacuum polarization screens charge, so $\\alpha$ grows from $1\u002F137$ at low energy to about $1\u002F128$ at the $Z$ mass.\n",{"path":28680,"title":28681,"module":28669,"summary":28682},"\u002Fparticle-physics\u002Fqed\u002Felectron-g-2","The Anomalous Magnetic Moment","The Dirac equation predicts $g=2$; loops shift it. Schwinger's one-loop vertex correction gives the anomaly $a=(g-2)\u002F2=\\alpha\u002F2\\pi$, and the QED series continues to five loops. The electron $a_e$ agrees with theory to better than a part in a billion, the most precise confrontation of theory and experiment in physics. The muon $a_\\mu$, heavier and so more sensitive to virtual heavy states, is dominated by hadronic uncertainty and sits at the center of a long-running comparison with the Standard Model prediction.\n",{"path":28684,"title":28685,"module":28686,"summary":28687},"\u002Fparticle-physics\u002Fweak-interaction\u002Fva-structure-weak","The V–A Charged Weak Current","The Weak Interaction","Fermi modelled beta decay as a four-fermion contact interaction, but a coupling with dimensions of inverse mass squared makes cross sections grow without bound and the theory fails near 300 GeV. The cure is a heavy mediator: the $W$ boson, whose propagator collapses to Fermi's contact term at low energy and fixes $G_F\u002F\\sqrt2 = g^2\u002F8M_W^2$. Parity violation dictates the current's form — vector minus axial-vector, coupling only to left-chiral fields — and universality of the coupling ties muon decay, beta decay, and pion decay to one constant. Pion decay's helicity suppression of the electron channel is the sharpest test.\n",{"path":28689,"title":28690,"module":28686,"summary":28691},"\u002Fparticle-physics\u002Fweak-interaction\u002Fw-z-bosons-decays","The W and Z Bosons","The contact theory hides a massive mediator. The charged $W^\\pm$ carries the current that changes flavour; the neutral $Z^0$ carries a current that does not. Both were found at CERN's proton–antiproton collider in 1983 at the masses the electroweak theory demanded. Their decay widths partition into leptonic and hadronic channels, and the $Z$ carries a decisive extra: an invisible width from decays to neutrinos that counts the number of light generations at exactly three. Beta decay and muon decay are re-read at the parton level as $W$ exchange.\n",{"path":28693,"title":28694,"module":28686,"summary":28695},"\u002Fparticle-physics\u002Fweak-interaction\u002Fckm-matrix","Quark Mixing and the CKM Matrix","The quark eigenstates the weak force acts on are not the mass eigenstates. Cabibbo captured this with one rotation angle; the GIM mechanism added a fourth quark to cancel dangerous flavour-changing neutral currents and predicted charm before its discovery. Three generations promote the rotation to the unitary Cabibbo–Kobayashi–Maskawa matrix — three angles and one irreducible complex phase, the sole source of Standard-Model CP violation. The Wolfenstein parametrization exposes its steep hierarchy, and unitarity closes into a triangle whose area measures the phase.\n",{"path":28697,"title":28698,"module":28686,"summary":28699},"\u002Fparticle-physics\u002Fweak-interaction\u002Fcp-violation-kaons-b-mesons","CP Violation in Kaons and B Mesons","The neutral kaon is its own laboratory for CP. Weak box diagrams mix $K^0$ and its antiparticle into short- and long-lived states that should be pure CP eigenstates decaying to two and three pions. In 1964 Cronin and Fitch caught the long-lived kaon decaying to two pions — CP is violated, at the two-per-mille level of $\\epsilon$. Direct violation ($\\epsilon'$) followed, and the $B$ factories turned the CKM phase into a large, clean time-dependent asymmetry measuring $\\sin 2\\beta$. The effect is real but far too small to explain why the universe is made of matter.\n",{"path":28701,"title":28702,"module":28703,"summary":28704},"\u002Fparticle-physics\u002Fqcd\u002Fcolor-su3-gluons","Color SU(3), Gluons, and the QCD Lagrangian","Quantum Chromodynamics","Color is the exact gauged SU(3) charge of the strong force. Gauging it forces eight massless gluons in the adjoint representation and, because the gauge group is non-abelian, three- and four-gluon self-couplings absent from QED. This lesson builds the QCD Lagrangian from the covariant derivative and the non-abelian field strength, states the Feynman rules with their color factors, and computes the Casimir invariants that set the strength of quark-gluon and gluon-gluon coupling.\n",{"path":28706,"title":28707,"module":28703,"summary":28708},"\u002Fparticle-physics\u002Fqcd\u002Fasymptotic-freedom-confinement","Asymptotic Freedom and Confinement","The QCD beta function is negative: gluon self-interaction antiscreens color, so the coupling weakens at short distance (asymptotic freedom) and strengthens at long distance (confinement). This lesson computes the one-loop beta coefficient, solves for the running of alpha_s and the emergent scale Lambda_QCD, and reads the strong-coupling regime as the linear quark-antiquark potential of a color flux tube that breaks by pair creation.\n",{"path":28710,"title":28711,"module":28703,"summary":28712},"\u002Fparticle-physics\u002Fqcd\u002Fdeep-inelastic-scattering-partons","Deep Inelastic Scattering and the Parton Model","Scattering electrons hard off a proton resolves pointlike constituents. This lesson sets up the deep-inelastic kinematics, defines the structure functions F1 and F2, and reads Bjorken scaling as the signature of free spin-half partons. The Callan-Gross relation fixes the parton spin, the structure function becomes a charge-weighted sum of parton distributions, and the slow logarithmic scaling violations expose the gluon through DGLAP evolution.\n",{"path":28714,"title":28715,"module":28703,"summary":28716},"\u002Fparticle-physics\u002Fqcd\u002Fjets-hadronization","Jets, Hadronization, and Testing QCD","Quarks and gluons produced in a collision fragment into collimated sprays of hadrons — jets — whose directions track the underlying partons. This lesson reads two-jet events as the quark and antiquark of electron-positron annihilation, three-jet events as direct evidence of the radiated gluon, and the hadronization step as the flux tube breaking into color singlets. Jet algorithms and event-shape variables turn the pattern into precision measurements of alpha_s.\n",{"path":28718,"title":28719,"module":28720,"summary":28721},"\u002Fparticle-physics\u002Felectroweak-higgs\u002Felectroweak-su2-u1","The Electroweak Theory","Electroweak Unification and the Higgs","The electromagnetic and weak interactions are two faces of a single gauge theory built on $SU(2)_L \\times U(1)_Y$. Left-handed fermions sit in weak-isospin doublets and right-handed fermions in singlets, each carrying a hypercharge fixed by the Gell-Mann–Nishijima relation $Q = T_3 + Y\u002F2$. The four gauge fields $W^{1,2,3}$ and $B$ mix: the charged combinations $W^\\pm$ mediate the charged current, while $W^3$ and $B$ rotate through the Weinberg angle into the massless photon and the massive $Z$. The single angle $\\theta_W$ ties the couplings, the boson masses, and the neutral-current strengths together.\n",{"path":28723,"title":28724,"module":28720,"summary":28725},"\u002Fparticle-physics\u002Felectroweak-higgs\u002Fspontaneous-symmetry-breaking","Spontaneous Symmetry Breaking","A symmetry of the Lagrangian need not be a symmetry of the ground state. When the lowest-energy configuration sits away from the symmetric point, the symmetry is spontaneously broken and the vacuum is one of a degenerate family. Breaking a continuous global symmetry produces one massless scalar — a Goldstone boson — for every broken generator, the flat direction along the vacuum manifold. The Mexican-hat potential and the ferromagnet below its Curie point are the working pictures.\n",{"path":28727,"title":28728,"module":28720,"summary":28729},"\u002Fparticle-physics\u002Felectroweak-higgs\u002Fhiggs-mechanism","The Higgs Mechanism","Gauging a spontaneously broken symmetry converts the would-be Goldstone bosons into the longitudinal polarizations of the gauge fields, which thereby acquire mass. Applied to $SU(2)_L \\times U(1)_Y$ with a single Higgs doublet, three of the four scalar degrees of freedom are eaten by the $W^\\pm$ and $Z$; the fourth survives as the physical Higgs boson, and the photon stays massless. Fermion masses come from Yukawa couplings to the same field, each mass proportional to its coupling times the vacuum expectation value $v \\approx 246$ GeV.\n",{"path":28731,"title":28732,"module":28720,"summary":28733},"\u002Fparticle-physics\u002Felectroweak-higgs\u002Fhiggs-boson-discovery","The Higgs Boson","The Higgs boson is produced at the LHC chiefly through gluon fusion, with vector-boson fusion and associated production as cleaner but rarer channels. It decays most often to $b\\bar b$ and $WW^\\ast$, but the discovery rested on two rare clean modes, $H \\to \\gamma\\gamma$ and $H \\to ZZ^\\ast \\to 4\\ell$, whose narrow invariant-mass peaks emerged over smooth backgrounds. ATLAS and CMS announced a boson near 125 GeV in 2012; its measured spin-parity $0^+$ and its couplings, which scale with particle mass, identify it as the Standard Model Higgs.\n",{"path":28735,"title":28736,"module":28720,"summary":28737},"\u002Fparticle-physics\u002Felectroweak-higgs\u002Fstandard-model","The Standard Model","The Standard Model combines the quark model, quantum chromodynamics, and the electroweak theory. SU(3) symmetry sorts the hadrons and predicted the omega; color explains why only colorless quark combinations exist; QCD gives asymptotic freedom and confinement; and spontaneous symmetry breaking through the Higgs field gives the weak bosons their mass.\n",{"path":28739,"title":28740,"module":28741,"summary":28742},"\u002Fparticle-physics\u002Fneutrinos\u002Fneutrino-oscillations","Neutrino Oscillations","Neutrino Physics","Neutrinos are produced and detected in flavour states, but they propagate as mass states, and the two bases are misaligned. A flavour therefore evolves coherently into a superposition of other flavours with a probability set by the mass-squared splitting and the ratio L\u002FE. This lesson derives the two-flavour oscillation formula, applies it to the solar and atmospheric neutrino deficits, shows how the SNO neutral-current measurement resolved the solar problem, and works out the MSW resonance that amplifies mixing inside the Sun.\n",{"path":28744,"title":28745,"module":28741,"summary":28746},"\u002Fparticle-physics\u002Fneutrinos\u002Fneutrino-mass-pmns","Neutrino Mass and the PMNS Matrix","Three-flavour mixing promotes the single oscillation angle to the unitary Pontecorvo–Maki–Nakagawa–Sakata matrix, parametrised by three angles and a Dirac CP phase. This lesson decomposes the PMNS matrix into three rotations, records the measured angles and mass-squared splittings, lays out the normal and inverted mass orderings, contrasts the large leptonic mixing with the near-diagonal CKM matrix, and collects the absolute-mass bounds from beta decay and cosmology.\n",{"path":28748,"title":28749,"module":28741,"summary":28750},"\u002Fparticle-physics\u002Fneutrinos\u002Fdirac-majorana-experiments","Dirac, Majorana, and Neutrino Experiments","A neutral fermion can carry a mass term forbidden to every charged particle, so the neutrino may be its own antiparticle. This lesson contrasts the Dirac and Majorana mass terms and their state content, derives the seesaw mechanism that ties a tiny light mass to a heavy right-handed partner, presents neutrinoless double-beta decay as the decisive lepton-number test, surveys the reactor, accelerator, solar, and atmospheric sources on a baseline–energy map, and explains why neutrino mass is physics beyond the original Standard Model.\n",{"path":28752,"title":28753,"module":28754,"summary":28755},"\u002Fparticle-physics\u002Fexperiment\u002Faccelerators-luminosity","Accelerators, Colliders, and Luminosity","Accelerators and Detectors","Fixed-target machines waste energy in the center-of-mass motion of the whole system, so the reachable $\\sqrt s$ grows only as the square root of the beam energy, while colliders put the full beam energy into the collision. Circular electron machines are limited by synchrotron radiation scaling as $E^4\u002Fm^4R$; proton machines are limited by bending fields. Luminosity, set by beam current and focusing, converts a cross section into an event rate through $R=\\mathcal L\\,\\sigma$, and integrated luminosity sets the total event count.\n",{"path":28757,"title":28758,"module":28754,"summary":28759},"\u002Fparticle-physics\u002Fexperiment\u002Fdetectors-subsystems","Particle Detectors and Subsystems","A detector reads a collision by the energy particles deposit as they cross matter. Charged particles ionize at the Bethe-Bloch rate, radiate in the field of nuclei above a critical energy, and emit Cherenkov light above a velocity threshold; electrons and photons build electromagnetic showers over a radiation length, and hadrons build wider showers over a nuclear interaction length. The onion of tracker, electromagnetic and hadronic calorimeters, and outer muon chambers turns these processes into momentum, energy, and identity, with neutrinos inferred from missing transverse momentum.\n",{"path":28761,"title":28762,"module":28754,"summary":28763},"\u002Fparticle-physics\u002Fexperiment\u002Fhow-discoveries-are-made","From Collisions to Discoveries","A discovery is a peak that survives statistics. Events are reconstructed into invariant masses, a signal accumulates as a bump over a smooth background, and its significance is judged by a p-value; the field's threshold is five sigma. The expected yield is a product — luminosity times cross section times branching ratio times acceptance and efficiency — that must be balanced by a trigger and data-reduction chain against an overwhelming rate. Worked reconstructions of $Z\\to\\ell\\ell$, the $J\u002F\\psi$, and the Higgs show the same peak-over-background logic at three scales.\n",{"path":28765,"title":28766,"module":28766,"summary":28767},"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fbeyond-standard-model","Beyond the Standard Model","The Standard Model leaves the four interactions ununified and the neutrinos massless, both now known to be wrong. Grand unification predicts the couplings merge near ten-to-the-sixteen GeV and the proton decays; supersymmetry pairs each particle with a superpartner; and the confirmed oscillation of neutrinos proves they carry mass, the first crack in the model.\n",{"path":28769,"title":28770,"module":28766,"summary":28771},"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fgrand-unified-theories","Grand Unified Theories and Proton Decay","The Standard Model gauge group is a product of three factors with three independent couplings. A grand unified theory embeds them in a single simple group — SU(5) is the minimal choice — so that one coupling runs into all three and the fractional quark charges follow from a tracelessness condition. The same embedding places quarks and leptons in shared multiplets, mediates baryon-number violation through superheavy gauge bosons, and predicts the proton decays with a lifetime that Super-Kamiokande has pushed past ten-to-the-thirty-four years.\n",{"path":28773,"title":28774,"module":28766,"summary":28775},"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fsupersymmetry","Supersymmetry","Supersymmetry relates fermions and bosons, pairing every Standard Model particle with a superpartner whose spin differs by one half. The pairing makes the scalar and fermion loop corrections to the Higgs mass cancel, removing the quadratic sensitivity to high scales; it sharpens the meeting of the three gauge couplings; and, when R-parity is conserved, it leaves the lightest superpartner stable and neutral, a natural dark-matter candidate. The LHC has excluded gluinos and light squarks below roughly two TeV.\n",{"path":28777,"title":28778,"module":28766,"summary":28779},"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fhierarchy-problem-naturalness","The Hierarchy Problem and Naturalness","The electroweak scale sits sixteen orders of magnitude below the Planck scale, and nothing in the Standard Model protects that gap. The Higgs mass squared picks up quadratic corrections proportional to the highest scale in the theory, so keeping it at the observed value requires the bare mass and its counterterm to cancel to some thirty significant figures. Naturalness treats that cancellation as a symptom of missing physics. Supersymmetry, compositeness, and extra dimensions each remove the quadratic sensitivity, but the LHC has found none of them at the predicted scale.\n",{"path":28781,"title":28782,"module":28766,"summary":28783},"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fdark-matter-candidates","Dark Matter and Particle Candidates","Flat galactic rotation curves, gravitational lensing, the cosmic microwave background, and structure formation all require about five times more matter than the visible baryons, none of it interacting electromagnetically. A stable weakly interacting particle of roughly weak-scale mass freezes out of the early universe with close to the observed abundance — the WIMP miracle — and is the leading candidate, with axions and sterile neutrinos as alternatives. Direct, indirect, and collider searches have so far only tightened the limits.\n",{"path":28785,"title":28786,"module":28766,"summary":28787},"\u002Fparticle-physics\u002Fbeyond-standard-model\u002Fmatter-antimatter-open-questions","Matter-Antimatter Asymmetry and Open Questions","The universe is made of matter, with about one extra baryon for every billion photons and no antimatter regions. Sakharov identified the three conditions any dynamical explanation must meet: baryon-number violation, C and CP violation, and a departure from thermal equilibrium. The Standard Model contains all three in principle, but its CP violation falls short by some ten orders of magnitude, so baryogenesis requires new physics — leptogenesis being the leading route. A closing survey collects the open questions and the experiments aimed at them.\n",{"path":28789,"title":28790,"module":6,"summary":6},"\u002Fparticle-physics","Particle Physics",{"path":28792,"title":28793,"module":28794,"summary":28795},"\u002Fastrophysics-cosmology\u002Forientation\u002Fthe-sun-and-stars","The Sun and the Life of Stars","Orientation","The Sun is the one star close enough to study in detail: its luminosity fixes a surface temperature of 5780 K, and the proton-proton fusion cycle in its 1.5-million-kelvin core supplies its power. Measuring other stars needs the magnitude scale, parallax, and the distance ladder; plotting luminosity against temperature builds the Hertzsprung-Russell diagram, on which a star's mass sets its lifetime and its evolutionary track off the main sequence.\n",{"path":28797,"title":28798,"module":28794,"summary":28799},"\u002Fastrophysics-cosmology\u002Forientation\u002Fstellar-death-final-states","Cataclysmic Events and the Final States of Stars","A star's death is set by its mass. In close binaries, matter poured across the Roche lobe onto a white dwarf produces novae and, at the Chandrasekhar limit of 1.4 solar masses, a Type Ia supernova; a massive star fusing to an iron core collapses into a Type II supernova. The remnant is a white dwarf held by electron degeneracy, a neutron star held by neutron degeneracy, or, above the neutron-star limit, a black hole inside its Schwarzschild radius.\n",{"path":28801,"title":28802,"module":28794,"summary":28803},"\u002Fastrophysics-cosmology\u002Forientation\u002Fgalaxies-and-cosmology","Galaxies, Cosmology, and the Evolving Universe","Galaxies come in elliptical, spiral, and irregular forms, and their redshifts obey Hubble's law, evidence that space itself is expanding. The critical density and the density parameter decide whether the universe is open, flat, or closed; baryons, dark matter, and dark energy each contribute. The cosmic microwave background and primordial helium anchor the Big Bang, whose thermal history runs from inflation through nucleosynthesis to the atoms of today.\n",{"path":28805,"title":28806,"module":28807,"summary":28808},"\u002Fastrophysics-cosmology\u002Fobservational-foundations\u002Fmagnitudes-fluxes-and-the-distance-modulus","Magnitudes, Fluxes, and the Distance Modulus","Observational Foundations","The brightness of a star reaches us as a radiant flux that falls off as the inverse square of distance. The magnitude scale encodes flux logarithmically through the Pogson ratio; the apparent and absolute magnitudes differ by the distance modulus, which converts a measured brightness into a distance. The bolometric correction folds a filtered magnitude into a total luminosity, and the difference of two magnitudes in different bands, the color index, measures surface temperature.\n",{"path":28810,"title":28811,"module":28807,"summary":28812},"\u002Fastrophysics-cosmology\u002Fobservational-foundations\u002Fstellar-spectra-and-spectral-classification","Stellar Spectra and Spectral Classification","A stellar spectrum is a continuum crossed by absorption lines whose strengths are set by the temperature of the atmosphere. The Boltzmann factor governs how atoms populate excited states, and the Saha equation governs how they ionize; their product explains why each line, such as the hydrogen Balmer series, peaks in strength at a characteristic temperature. This behavior orders stars into the OBAFGKM sequence, and the luminosity classes of the MK system add a second dimension for surface gravity.\n",{"path":28814,"title":28815,"module":28807,"summary":28816},"\u002Fastrophysics-cosmology\u002Fobservational-foundations\u002Ftelescopes-and-detectors-across-the-spectrum","Telescopes and Detectors Across the Spectrum","A telescope collects light in proportion to its collecting area and resolves detail down to the diffraction limit set by its aperture and the observing wavelength. The atmosphere blurs and blocks large parts of the spectrum, which drives the choice between ground and space and between refractors, reflectors, and radio dishes. CCDs record the light with high quantum efficiency, and interferometry synthesizes an aperture as large as the separation of two telescopes.\n",{"path":28818,"title":28819,"module":28807,"summary":28820},"\u002Fastrophysics-cosmology\u002Fobservational-foundations\u002Fthe-cosmic-distance-ladder","The Cosmic Distance Ladder","No single method measures distances from the nearest stars to the far reaches of the universe. Instead a ladder of overlapping techniques, each calibrated by the one below it, extends the scale rung by rung: trigonometric parallax, main-sequence fitting, pulsating variables, the tip of the red-giant branch, the Tully-Fisher relation, and Type Ia supernovae. Each rung inherits the uncertainty of every rung beneath it, so the whole chain sets the accuracy of the Hubble constant.\n",{"path":28822,"title":28823,"module":28824,"summary":28825},"\u002Fastrophysics-cosmology\u002Fradiation-and-matter\u002Fblackbody-radiation-and-specific-intensity","Blackbody Radiation and Specific Intensity","Radiation and Matter","Specific intensity is the fundamental measure of a radiation field: energy per unit area, time, frequency, and solid angle. It is conserved along a ray in empty space, and its angular moments give the mean intensity, flux, and radiation pressure. In thermal equilibrium the intensity equals the Planck function, whose limits and integrals reproduce the Rayleigh-Jeans law, the Wien law, Stefan-Boltzmann, and Wien's displacement law.\n",{"path":28827,"title":28828,"module":28824,"summary":28829},"\u002Fastrophysics-cosmology\u002Fradiation-and-matter\u002Fradiative-transfer-and-the-transfer-equation","Radiative Transfer and the Transfer Equation","Along a ray, matter adds intensity through emission and removes it through absorption. Measuring path length in optical depth turns this into the transfer equation, whose formal solution superposes an attenuated background on the source function integrated along the line of sight. In local thermodynamic equilibrium the source function is the Planck function, and the Eddington-Barbier relation shows that the emergent intensity samples the source function at optical depth of order unity, explaining absorption lines and solar limb darkening.\n",{"path":28831,"title":28832,"module":28824,"summary":28833},"\u002Fastrophysics-cosmology\u002Fradiation-and-matter\u002Fspectral-line-formation-and-broadening","Spectral-Line Formation and Broadening","A spectral line is a bound-bound transition whose strength is set by an oscillator strength and whose shape is set by three broadening mechanisms: the Lorentzian natural and collisional wings, the Gaussian thermal Doppler core, and their Voigt convolution. Equivalent width measures the total absorption, and the curve of growth relates it to the number of absorbers through a linear, saturated, and damping regime, turning line strengths into abundances.\n",{"path":28835,"title":28836,"module":28824,"summary":28837},"\u002Fastrophysics-cosmology\u002Fradiation-and-matter\u002Fopacity-and-the-rosseland-mean","Opacity Sources and the Rosseland Mean","Stellar opacity comes from four processes: bound-bound line absorption, bound-free photoionization, free-free absorption, and electron scattering. The bound-free and free-free terms follow a Kramers law, electron scattering sets a frequency-flat floor, and the negative hydrogen ion dominates cool photospheres. The Rosseland mean averages these harmonically, weighting transparent frequencies because they carry the flux, and its value fixes the radiative temperature gradient and decides where a star becomes convective.\n",{"path":28839,"title":28840,"module":28841,"summary":28842},"\u002Fastrophysics-cosmology\u002Fstellar-structure\u002Fhydrostatic-equilibrium-and-the-virial-theorem","Hydrostatic Equilibrium and the Virial Theorem","Stellar Structure","A star holds itself up by balancing the inward pull of gravity against an outward pressure gradient. This balance, hydrostatic equilibrium, fixes a lower bound on the central pressure and, combined with the gravitational potential energy, yields the virial theorem. The virial relation gives a star a negative heat capacity, so that losing energy makes it hotter, and sets the Kelvin-Helmholtz timescale over which contraction alone can power the Sun.\n",{"path":28844,"title":28845,"module":28841,"summary":28846},"\u002Fastrophysics-cosmology\u002Fstellar-structure\u002Fthe-equations-of-stellar-structure","The Equations of Stellar Structure","A static star is described by four coupled first-order differential equations in the interior mass or radius: mass conservation, hydrostatic equilibrium, energy generation, and energy transport. Closed with an equation of state, opacity, and reaction rates, and subject to central and surface boundary conditions, they determine the structure uniquely from mass and composition, the Vogt-Russell theorem. Energy moves by radiation until the temperature gradient exceeds the Schwarzschild limit, where convection takes over.\n",{"path":28848,"title":28849,"module":28841,"summary":28850},"\u002Fastrophysics-cosmology\u002Fstellar-structure\u002Fthe-equation-of-state-and-polytropes","The Equation of State and Polytropes","Stellar pressure comes from gas, radiation, and, at high density, degenerate electrons. When pressure depends on density as a power law, hydrostatic equilibrium reduces to the Lane-Emden equation, whose solutions describe polytropes of index n. The relativistic degenerate case, n equal to three, gives a mass independent of radius, the Chandrasekhar mass. Eddington's standard model treats a radiation-supported star as an n equal to three polytrope and yields the quartic relating radiation fraction to mass.\n",{"path":28852,"title":28853,"module":28841,"summary":28854},"\u002Fastrophysics-cosmology\u002Fstellar-structure\u002Fthe-standard-solar-model","The Standard Solar Model","The standard solar model integrates the structure equations for one solar mass and calibrates the composition and convection parameter to reproduce the Sun's observed luminosity, radius, and age. Helioseismology tests the model's sound speed through the Sun's acoustic p-mode oscillations, and the model predicts a neutrino flux by production channel. The measured deficit, the solar-neutrino problem, is resolved by matter-enhanced flavor oscillation, confirmed when SNO measured the total flux across all flavors.\n",{"path":28856,"title":28857,"module":28858,"summary":28859},"\u002Fastrophysics-cosmology\u002Fnuclear-astrophysics\u002Fthermonuclear-reaction-rates-and-the-gamow-peak","Thermonuclear Reaction Rates and the Gamow Peak","Nuclear Astrophysics","Stellar fusion proceeds only by quantum tunneling through the Coulomb barrier, because thermal energies are a thousand times smaller than the barrier height. The reaction rate is an integral over the Maxwell–Boltzmann distribution and the tunneling probability, whose product is sharply peaked at the Gamow energy. The astrophysical S-factor isolates the nuclear physics from the barrier penetration, and the steep temperature dependence follows from the width and position of the Gamow peak.\n",{"path":28861,"title":28862,"module":28858,"summary":28863},"\u002Fastrophysics-cosmology\u002Fnuclear-astrophysics\u002Fhydrogen-burning-pp-chains-and-cno","Hydrogen Burning: pp Chains and the CNO Cycle","Four protons fuse into one helium-4 nucleus, releasing 26.7 MeV, through two competing networks. The pp chain begins with a weak-interaction bottleneck and branches three ways; the CNO cycle uses carbon, nitrogen, and oxygen as catalysts and is limited by nitrogen-14 proton capture. Their steep and gentle temperature dependences cross near 1.8e7 K, which divides pp-powered lower-main-sequence stars from CNO-powered upper-main-sequence stars.\n",{"path":28865,"title":28866,"module":28858,"summary":28867},"\u002Fastrophysics-cosmology\u002Fnuclear-astrophysics\u002Fhelium-burning-and-the-triple-alpha-process","Helium Burning and the Triple-Alpha Process","Helium fuses to carbon in two steps through the unbound beryllium-8 nucleus and a resonant excited state of carbon-12, the Hoyle state, whose existence was predicted from the observed carbon abundance. The rate scales as roughly the fortieth power of temperature, and in a degenerate low-mass core this drives the runaway helium flash. A competing alpha capture on carbon-12 sets the carbon-to-oxygen ratio and the composition of the resulting white dwarf.\n",{"path":28869,"title":28870,"module":28858,"summary":28871},"\u002Fastrophysics-cosmology\u002Fnuclear-astrophysics\u002Fadvanced-burning-and-neutron-capture-nucleosynthesis","Advanced Burning, the Iron Peak, and the s\u002Fr Processes","Massive stars burn carbon, neon, oxygen, and silicon in ever-shorter stages, building an onion-shell interior and reaching nuclear statistical equilibrium at the iron peak, where the binding-energy-per-nucleon curve turns over and fusion can release no more energy. Elements beyond iron form by neutron capture: the slow s-process in AGB stars tracks the valley of stability, while the rapid r-process in supernovae and neutron-star mergers builds the heaviest nuclei far from it.\n",{"path":28873,"title":28874,"module":28875,"summary":28876},"\u002Fastrophysics-cosmology\u002Fism-and-star-formation\u002Fphases-of-the-interstellar-medium","The Phases of the Interstellar Medium","The Interstellar Medium","The gas between the stars separates into distinct thermal phases, from cold molecular clouds at 10 K to a diffuse million-degree corona, held near a common pressure by a balance of photoelectric heating and radiative cooling. Neutral hydrogen is traced by the 21-cm hyperfine line, dust reddens and extinguishes starlight along a characteristic wavelength law, and the ultraviolet output of hot stars carves ionized Strömgren spheres out of the surrounding gas.\n",{"path":28878,"title":28879,"module":28875,"summary":28880},"\u002Fastrophysics-cosmology\u002Fism-and-star-formation\u002Fmolecular-clouds-and-gravitational-collapse","Molecular Clouds and Gravitational Collapse","Stars form in cold, dense molecular clouds when self-gravity overcomes thermal and magnetic support. The virial theorem fixes the Jeans mass and length at which a clump becomes unstable, the free-fall time sets how fast it collapses, and a fragmentation cascade — cut off at a minimum mass by the onset of opacity — turns one cloud into a whole cluster, imprinting the stellar initial mass function.\n",{"path":28882,"title":28883,"module":28875,"summary":28884},"\u002Fastrophysics-cosmology\u002Fism-and-star-formation\u002Fprotostars-and-the-pre-main-sequence","Protostars and Pre-Main-Sequence Evolution","A collapsing core becomes optically thick and forms a protostar that grows by accretion through a disk while driving bipolar outflows. The newborn star appears on the birthline and contracts down the fully convective Hayashi track, then crosses the radiative Henyey track to the zero-age main sequence, powered by gravitational contraction until hydrogen ignites. Below about 0.08 solar masses degeneracy halts contraction before ignition, dividing stars from brown dwarfs.\n",{"path":28886,"title":28887,"module":28888,"summary":28889},"\u002Fastrophysics-cosmology\u002Fstellar-evolution\u002Fthe-main-sequence-and-its-structure","The Main Sequence and Its Structure","Stellar Evolution","A star settles onto the zero-age main sequence when core hydrogen ignition halts contraction. Homology scaling of the structure equations reproduces the mass–luminosity relation, and the burning mode splits the sequence into an upper branch with a convective core and a lower branch with a convective envelope. The main-sequence lifetime falls steeply with mass, and the turnoff of a coeval cluster serves as a clock.\n",{"path":28891,"title":28892,"module":28888,"summary":28893},"\u002Fastrophysics-cosmology\u002Fstellar-evolution\u002Fpost-main-sequence-low-mass-evolution","Post-Main-Sequence Evolution of Low-Mass Stars","When a low-mass star exhausts core hydrogen, burning moves to a shell, the core contracts, and the envelope swells into a red giant. A degenerate helium core ignites in a flash, settles onto the horizontal branch, and after a second contraction the star climbs the asymptotic giant branch with two burning shells. Thermal pulses and dredge-up enrich the surface, and mass loss ejects a planetary nebula, leaving a carbon–oxygen white dwarf.\n",{"path":28895,"title":28896,"module":28888,"summary":28897},"\u002Fastrophysics-cosmology\u002Fstellar-evolution\u002Fthe-evolution-of-massive-stars","The Evolution of Massive Stars","Stars above about eight solar masses burn through hydrogen, helium, carbon, neon, oxygen, and silicon in stages that grow shorter as neutrino losses accelerate contraction. The interior becomes an onion of concentric burning shells around an inert iron core. Radiation pressure near the Eddington limit drives fierce winds that can strip the hydrogen envelope entirely, and silicon burning builds an iron core toward the threshold of collapse.\n",{"path":28899,"title":28900,"module":28888,"summary":28901},"\u002Fastrophysics-cosmology\u002Fstellar-evolution\u002Fstellar-pulsation-and-the-instability-strip","Stellar Pulsation and the Instability Strip","Radial pulsation is a standing sound wave whose period scales inversely with the square root of the mean density. The kappa mechanism, an opacity valve seated in the helium partial-ionization zone, turns a star into a heat engine that pumps the oscillation. Stars in the instability strip pulsate as Cepheids, RR Lyrae, and Mira variables, and the Cepheid period–luminosity relation calibrates the distance ladder.\n",{"path":28903,"title":27319,"module":28904,"summary":28905},"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fwhite-dwarfs-and-the-chandrasekhar-limit","Stellar Death and Compact Remnants","A white dwarf is held up by the degeneracy pressure of its electrons, a quantum-mechanical stiffness that survives to zero temperature. Filling the Fermi sea sets a pressure that scales as density to the five-thirds power when the electrons are slow and only four-thirds when they are relativistic. The softer relativistic law produces the inverted mass-radius relation and a maximum mass, the Chandrasekhar limit near 1.4 solar masses, above which no cold equilibrium exists. Cooling and crystallization then turn the white-dwarf population into a clock for the Galactic disk.\n",{"path":28907,"title":28908,"module":28904,"summary":28909},"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fcore-collapse-supernovae","Core-Collapse Supernovae","When a massive star builds an iron core past the Chandrasekhar mass, degeneracy fails and the core collapses in less than a second. Photodisintegration and electron capture remove pressure support and neutronize the matter; the collapse halts abruptly at nuclear density, launching a shock that stalls and is revived by neutrino heating. The event is a Type II or stripped-envelope Ib\u002FIc supernova, and the neutrinos from SN 1987A confirmed the picture directly.\n",{"path":28911,"title":28912,"module":28904,"summary":28913},"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fthermonuclear-supernovae-type-ia","Thermonuclear Supernovae","A carbon-oxygen white dwarf driven toward the Chandrasekhar mass ignites its degenerate fuel and unbinds itself in a thermonuclear runaway, the Type Ia supernova. The light curve is powered by the radioactive decay of nickel-56 to cobalt-56 to iron-56, and the Phillips relation between peak brightness and decline rate makes these events standardizable candles. Their near-uniform luminosity turns them into the distance indicators that revealed cosmic acceleration.\n",{"path":28915,"title":28916,"module":28904,"summary":28917},"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fneutron-stars-and-pulsars","Neutron Stars and Pulsars","A neutron star is held up by neutron degeneracy and the repulsive nuclear force, with a maximum mass, the Tolman-Oppenheimer-Volkoff limit, set by an uncertain dense-matter equation of state. Its rotating magnetic dipole sweeps a beam past Earth as a pulsar, and magnetic braking traces a track across the period-period- derivative diagram. Millisecond pulsars, magnetars, glitches, and the orbital decay of the Hulse-Taylor binary follow from the same structure.\n",{"path":28919,"title":28920,"module":28904,"summary":28921},"\u002Fastrophysics-cosmology\u002Fstellar-death-and-compact-remnants\u002Fblack-holes-schwarzschild-and-kerr","Black Holes, Schwarzschild and Kerr","Above the neutron-star mass limit gravity wins completely and the remnant is a black hole. The Schwarzschild solution gives the event horizon, gravitational redshift, and time dilation; the innermost stable circular orbit sets the efficiency of accretion. Rotating Kerr black holes drag spacetime and carry an ergosphere. Stellar-mass black holes are found in X-ray binaries, and the Event Horizon Telescope has imaged the shadow of a supermassive one.\n",{"path":28923,"title":28924,"module":28925,"summary":28926},"\u002Fastrophysics-cosmology\u002Fbinaries-and-gravitational-waves\u002Fbinary-systems-and-mass-transfer","Binary Systems and Mass Transfer","Binaries and Gravitational Waves","Most stars are born in pairs, and a binary is the only setting where a stellar mass can be measured directly. Visual, spectroscopic, and eclipsing binaries each expose a different combination of the orbital elements, and together they calibrate the mass-luminosity relation. When one star swells to fill its Roche lobe, gas streams through the inner Lagrange point onto its companion. Conservative transfer widens or shrinks the orbit depending on the mass ratio, and the sign of that response explains the Algol paradox.\n",{"path":28928,"title":28929,"module":28925,"summary":28930},"\u002Fastrophysics-cosmology\u002Fbinaries-and-gravitational-waves\u002Faccreting-compact-objects","Accreting Compact Objects","Gas falling onto a compact object converts gravitational binding energy into radiation with an efficiency set by the depth of the potential well, up to tens of percent of the rest mass for a neutron star or black hole. Angular momentum forces the flow into a disk, and viscous dissipation gives a temperature profile that falls as radius to the minus three-quarters, producing a multicolor blackbody spectrum. Radiation pressure caps the steady luminosity at the Eddington limit. Unstable nuclear burning of the accreted fuel powers classical novae on white dwarfs and Type I X-ray bursts on neutron stars.\n",{"path":28932,"title":28933,"module":28925,"summary":28934},"\u002Fastrophysics-cosmology\u002Fbinaries-and-gravitational-waves\u002Fgravitational-waves-from-inspiraling-binaries","Gravitational Waves from Inspiraling Binaries","A time-varying mass quadrupole radiates gravitational waves, ripples in spacetime that stretch and squeeze a ring of free masses along two polarizations. The radiated power drains a binary's orbital energy, shrinking the orbit and sweeping the wave frequency upward in a chirp whose rate fixes the chirp mass. Laser interferometers with kilometre arms measure the resulting strain of order ten to the minus twenty-one. The first detection, GW150914, matched a template for two merging black holes near thirty solar masses each.\n",{"path":28936,"title":28937,"module":28925,"summary":28938},"\u002Fastrophysics-cosmology\u002Fbinaries-and-gravitational-waves\u002Fmultimessenger-astronomy-and-gamma-ray-bursts","Multimessenger Astronomy and Gamma-Ray Bursts","Gamma-ray bursts split into two populations: long bursts from the collapse of massive stars and short bursts from merging compact objects. The compactness problem forces the emitting plasma to move at ultra-relativistic speed, beaming the radiation into a narrow jet. The neutron-star merger GW170817 tied a gravitational chirp to a short gamma-ray burst, a radioactive kilonova, and a broadband afterglow, confirming that mergers forge r-process elements. A merger with a measured redshift is a standard siren that reads the Hubble constant from gravitational data alone.\n",{"path":28940,"title":28941,"module":28942,"summary":28943},"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Fthe-milky-way","The Milky Way Galaxy","Galaxies and Dark Matter","The Galaxy resolves into a thin disk of gas and young stars, a central bar and bulge, and a diffuse old halo studded with globular clusters. Star counts and the reddening of distant light map these components, while the differential rotation of the disk — encoded in the Oort constants and the flat rotation curve — measures the enclosed mass and reveals more than the stars can account for. Spiral arms are density waves, not material structures, and the innermost stellar orbits around Sgr A* weigh a four-million-solar-mass black hole.\n",{"path":28945,"title":28946,"module":28942,"summary":28947},"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Fgalaxy-morphology-and-classification","Galaxy Morphology and Classification","Galaxies sort along the Hubble tuning fork from smooth ellipticals through lenticulars to grand-design and barred spirals, with irregulars off the end. The light of a spheroid follows the de Vaucouleurs quarter-power law while a disk fades exponentially, and the general Sérsic profile interpolates between them. Virial scaling relations — Tully–Fisher for disks, Faber–Jackson and the fundamental plane for spheroids — tie luminosity to internal motions, and the Schechter function fixes the abundance of galaxies as a function of luminosity.\n",{"path":28949,"title":28950,"module":28942,"summary":28951},"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Fgalaxy-rotation-curves-and-dark-matter","Galaxy Rotation Curves and Dark Matter","The rotation curves of disk galaxies stay flat far beyond the light, demanding an extended halo whose density falls as the inverse square of radius. Decomposing the curve into disk, bulge, and halo, and fitting isothermal or NFW profiles, quantifies the missing mass. Gravitational lensing weighs the same mass without dynamics, the mass-to-light ratio climbs from stars to clusters, and the Bullet Cluster separates the collisionless dark matter from the colliding gas — evidence that MOND strains to match.\n",{"path":28953,"title":28954,"module":28942,"summary":28955},"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Factive-galactic-nuclei-and-supermassive-black-holes","Active Galactic Nuclei","A small fraction of galaxies pour out enormous luminosity from a region smaller than the solar system. Accretion onto a supermassive black hole, limited by the Eddington balance of radiation pressure and gravity, powers the Seyferts, quasars, radio galaxies, and blazars — one engine seen from different angles through an obscuring torus. Relativistic jets produce apparent superluminal motion, reverberation mapping and stellar dynamics weigh the central mass, and the M–sigma relation ties that mass to the host bulge.\n",{"path":28957,"title":28958,"module":28942,"summary":28959},"\u002Fastrophysics-cosmology\u002Fgalaxies\u002Fgalaxy-clusters-and-large-scale-structure","Galaxy Clusters and Large-Scale Structure","Galaxies gather into groups and rich clusters bound by a common dark halo and filled with hot X-ray gas. Three independent probes — the virial theorem, the hydrostatic X-ray temperature, and gravitational lensing — agree on a mass that dwarfs the stars. On the largest scales galaxies trace a cosmic web of filaments, walls, and voids, quantified by the two-point correlation function, whose baryon acoustic oscillation bump provides a standard ruler for cosmology.\n",{"path":28961,"title":28962,"module":28963,"summary":28964},"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fthe-expanding-universe-and-hubbles-law","The Expanding Universe and Hubble's Law","Cosmic Expansion and Dynamics","The universe is homogeneous and isotropic on large scales, so its expansion is captured by a single function of time, the scale factor. Comoving coordinates stay fixed while proper distances grow in proportion to the scale factor, producing Hubble's law and a cosmological redshift that measures stretched space rather than a Doppler shift. A Newtonian energy argument reproduces the dynamics, and the same finite, expanding cosmos resolves Olbers' paradox.\n",{"path":28966,"title":28967,"module":28963,"summary":28968},"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fthe-frw-metric-and-cosmological-redshift","The FRW Metric and Cosmological Redshift","The geometry of a homogeneous, isotropic universe is fixed by symmetry to the Robertson-Walker metric, with the entire freedom reduced to a scale factor and a single curvature constant selecting an open, flat, or closed space. From the metric the null geodesic of light gives comoving distance, the exact cosmological redshift, and the distinction between the proper distance we cannot measure and the redshift we can.\n",{"path":28970,"title":26047,"module":28963,"summary":28971},"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fthe-friedmann-equations-and-cosmic-dynamics","The scale factor obeys the Friedmann equation, the acceleration equation, and the fluid equation, only two of which are independent. An equation of state fixes how each component behaves under expansion, so radiation dilutes as the inverse fourth power of the scale factor, matter as the inverse cube, and vacuum energy not at all. The critical density defines the density parameters, and the deceleration parameter encodes whether gravity or dark energy is winning.\n",{"path":28973,"title":28974,"module":28963,"summary":28975},"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fcosmological-models-and-distances","Cosmological Models and Distances","Integrating the Friedmann equation for particular mixtures gives the benchmark models, from the matter-only Einstein-de Sitter universe to the concordance Lambda-CDM, each with its own scale-factor history and age. Because the redshift is the only direct observable, several distance measures diverge at high redshift, and the angular-diameter distance even turns over so that the most distant objects look larger. The horizon and lookback time set what is causally and observationally reachable.\n",{"path":28977,"title":28978,"module":28963,"summary":28979},"\u002Fastrophysics-cosmology\u002Fcosmology-expansion-and-dynamics\u002Fdark-energy-and-the-accelerating-universe","Dark Energy and the Accelerating Universe","In 1998 two teams found that distant Type Ia supernovae are fainter than a decelerating universe predicts, revealing that the expansion is accelerating and that a component with negative pressure dominates the energy budget. The simplest candidate is the cosmological constant, or vacuum energy, with an equation of state near minus one. It works observationally but leaves two deep puzzles: why the vacuum energy is a hundred and twenty orders of magnitude smaller than expected, and why it is comparable to the matter density just now.\n",{"path":28981,"title":28982,"module":28983,"summary":28984},"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fthe-thermal-history-of-the-universe","The Thermal History of the Universe","The Hot Big Bang","Running the expansion backward compresses and heats the universe, so its past is a sequence of thermal epochs set by temperature. Temperature scales as the inverse scale factor; species stay in equilibrium while their interaction rate exceeds the expansion rate and freeze out when it drops below. The effective degrees of freedom count the relativistic species and step down through mass thresholds, and neutrino decoupling just before electron-positron annihilation leaves a relic neutrino background slightly cooler than the photons.\n",{"path":28986,"title":28987,"module":28983,"summary":28988},"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fbig-bang-nucleosynthesis","Big Bang Nucleosynthesis","In the first three minutes the weak interactions freeze out the neutron-to-proton ratio, and once deuterium survives photodissociation a fast reaction network converts nearly all free neutrons into helium-4. The primordial abundances of deuterium, helium-3, helium-4, and lithium-7 depend on a single free parameter, the baryon-to-photon ratio, so measuring them fixes the baryon density. The predictions match observation across nine decades of abundance, with a persistent discrepancy in lithium-7.\n",{"path":28990,"title":28991,"module":28983,"summary":28992},"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Frecombination-and-the-cosmic-microwave-background","Recombination and the Cosmic Microwave Background","As the universe cooled through a few thousand kelvin the free electrons bound to protons, and the Saha equation tracks the falling ionization fraction. Once the plasma neutralized, photons stopped scattering and streamed freely from a spherical surface of last scattering at redshift about 1100. Those photons are the cosmic microwave background, an almost perfect blackbody at 2.725 kelvin with a dipole from our motion through it.\n",{"path":28994,"title":28995,"module":28983,"summary":28996},"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fcmb-anisotropies-and-cosmological-parameters","CMB Anisotropies and Cosmological Parameters","The cosmic microwave background carries temperature fluctuations at the ten-parts-per-million level, imprinted by sound waves in the photon-baryon plasma before recombination. Decomposed into spherical harmonics, the fluctuations form an angular power spectrum whose acoustic peaks encode the geometry and contents of the universe: the first peak fixes spatial flatness, the odd-even peak ratio the baryon density, and the third peak the dark-matter density. Polarization adds an independent channel, and the Planck measurements pin the concordance parameters.\n",{"path":28998,"title":28999,"module":28983,"summary":29000},"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fcosmic-inflation","Cosmic Inflation","The hot Big Bang leaves three initial-condition puzzles unexplained: why causally disconnected patches share a temperature, why the geometry is so nearly flat, and why no magnetic monopoles are seen. A brief epoch of accelerated expansion driven by a slowly rolling scalar field solves all three by stretching a small causal patch across the observable universe. The same accelerated expansion freezes quantum fluctuations into a near-scale-invariant spectrum of density perturbations, seeding all later structure.\n",{"path":29002,"title":29003,"module":28983,"summary":29004},"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fstructure-formation-and-the-growth-of-perturbations","Structure Formation and the Growth of Perturbations","The near-uniform early universe grew its galaxies and clusters by gravitational instability acting on the tiny inflationary perturbations. In an expanding background the growth is slowed to a power law rather than the exponential of a static medium; perturbations stall during radiation domination and grow with the scale factor once matter dominates. The transfer function turns the primordial spectrum into the processed matter power spectrum, and cold dark matter builds structure from the bottom up.\n",{"path":29006,"title":29007,"module":28983,"summary":29008},"\u002Fastrophysics-cosmology\u002Fthe-hot-big-bang\u002Fdark-matter-dark-energy-and-open-questions","Dark Matter, Dark Energy, and Open Questions","Five independent lines of evidence converge on a universe whose energy budget is dominated by dark energy and dark matter, with ordinary baryons a small remainder. The candidate particles for dark matter range from WIMPs to axions to sterile neutrinos, each with its own detection strategy. The concordance model fits the data with six parameters but leaves the nature of dark energy, the Hubble tension, small-scale structure, and the matter-antimatter asymmetry unexplained.\n",{"path":29010,"title":29011,"module":6,"summary":6},"\u002Fastrophysics-cosmology","Astrophysics & Cosmology",{"path":29013,"title":29014,"module":6,"summary":6},"\u002Fcolophon","Colophon",{"path":8240,"title":29016,"module":6,"summary":6},"Study Notes",[29018,29029,29038,29051,29064,29075,29096,29109,29126,29135,29153,29179],{"module":26893,"moduleNumber":8365,"slug":29019,"lessons":29020},"mathematical-background",[29021,29023,29025,29027],{"title":25534,"path":26892,"lessonNumber":8365,"topics":29022,"summary":26894},[26893],{"title":26897,"path":26896,"lessonNumber":8371,"topics":29024,"summary":26898},[26893],{"title":26901,"path":26900,"lessonNumber":8377,"topics":29026,"summary":26902},[26893],{"title":24893,"path":26904,"lessonNumber":8383,"topics":29028,"summary":26905},[26893],{"module":24340,"moduleNumber":8371,"slug":29030,"lessons":29031},"foundations",[29032,29034,29036],{"title":26908,"path":26907,"lessonNumber":8365,"topics":29033,"summary":26909},[24340],{"title":26912,"path":26911,"lessonNumber":8371,"topics":29035,"summary":26913},[24340],{"title":26916,"path":26915,"lessonNumber":8377,"topics":29037,"summary":26917},[24340],{"module":26921,"moduleNumber":8377,"slug":29039,"lessons":29040},"neural-networks",[29041,29043,29045,29047,29049],{"title":26920,"path":26919,"lessonNumber":8365,"topics":29042,"summary":26922},[26921],{"title":26925,"path":26924,"lessonNumber":8371,"topics":29044,"summary":26926},[26921],{"title":26929,"path":26928,"lessonNumber":8377,"topics":29046,"summary":26930},[26921],{"title":26933,"path":26932,"lessonNumber":8383,"topics":29048,"summary":26934},[26921],{"title":26937,"path":26936,"lessonNumber":8389,"topics":29050,"summary":26938},[26921],{"module":26942,"moduleNumber":8383,"slug":29052,"lessons":29053},"optimization",[29054,29056,29058,29060,29062],{"title":26941,"path":26940,"lessonNumber":8365,"topics":29055,"summary":26943},[26942],{"title":26946,"path":26945,"lessonNumber":8371,"topics":29057,"summary":26947},[26942],{"title":26950,"path":26949,"lessonNumber":8377,"topics":29059,"summary":26951},[26942],{"title":26954,"path":26953,"lessonNumber":8383,"topics":29061,"summary":26955},[26942],{"title":26958,"path":26957,"lessonNumber":8389,"topics":29063,"summary":26959},[26942],{"module":26963,"moduleNumber":8389,"slug":29065,"lessons":29066},"regularization",[29067,29069,29071,29073],{"title":26962,"path":26961,"lessonNumber":8365,"topics":29068,"summary":26964},[26963],{"title":26967,"path":26966,"lessonNumber":8371,"topics":29070,"summary":26968},[26963],{"title":26971,"path":26970,"lessonNumber":8377,"topics":29072,"summary":26972},[26963],{"title":26974,"path":18525,"lessonNumber":8383,"topics":29074,"summary":26975},[26963],{"module":23403,"moduleNumber":8395,"slug":29076,"lessons":29077},"architectures",[29078,29080,29082,29084,29086,29088,29090,29092,29094],{"title":26978,"path":26977,"lessonNumber":8365,"topics":29079,"summary":26979},[23403],{"title":26981,"path":18943,"lessonNumber":8371,"topics":29081,"summary":26982},[23403],{"title":26984,"path":17,"lessonNumber":8377,"topics":29083,"summary":26985},[23403],{"title":26987,"path":154,"lessonNumber":8383,"topics":29085,"summary":26988},[23403],{"title":5,"path":23405,"lessonNumber":8389,"topics":29087,"summary":23421},[23403],{"title":26991,"path":22458,"lessonNumber":8395,"topics":29089,"summary":26992},[23403],{"title":26995,"path":26994,"lessonNumber":8401,"topics":29091,"summary":26996},[23403],{"title":26999,"path":26998,"lessonNumber":8407,"topics":29093,"summary":27000},[23403],{"title":27003,"path":27002,"lessonNumber":8413,"topics":29095,"summary":27004},[23403],{"module":27008,"moduleNumber":8401,"slug":29097,"lessons":29098},"theory",[29099,29101,29103,29105,29107],{"title":27007,"path":27006,"lessonNumber":8365,"topics":29100,"summary":27009},[27008],{"title":27012,"path":27011,"lessonNumber":8371,"topics":29102,"summary":27013},[27008],{"title":27016,"path":27015,"lessonNumber":8377,"topics":29104,"summary":27017},[27008],{"title":27020,"path":27019,"lessonNumber":8383,"topics":29106,"summary":27021},[27008],{"title":27024,"path":27023,"lessonNumber":8389,"topics":29108,"summary":27025},[27008],{"module":27029,"moduleNumber":8407,"slug":29110,"lessons":29111},"generative-models",[29112,29114,29116,29118,29120,29122,29124],{"title":27028,"path":27027,"lessonNumber":8365,"topics":29113,"summary":27030},[27029],{"title":27033,"path":27032,"lessonNumber":8371,"topics":29115,"summary":27034},[27029],{"title":27037,"path":27036,"lessonNumber":8377,"topics":29117,"summary":27038},[27029],{"title":27041,"path":27040,"lessonNumber":8383,"topics":29119,"summary":27042},[27029],{"title":27045,"path":27044,"lessonNumber":8389,"topics":29121,"summary":27046},[27029],{"title":27049,"path":27048,"lessonNumber":8395,"topics":29123,"summary":27050},[27029],{"title":27053,"path":27052,"lessonNumber":8401,"topics":29125,"summary":27054},[27029],{"module":27058,"moduleNumber":8413,"slug":29127,"lessons":29128},"probabilistic-methods",[29129,29131,29133],{"title":27057,"path":27056,"lessonNumber":8365,"topics":29130,"summary":27059},[27058],{"title":27062,"path":27061,"lessonNumber":8371,"topics":29132,"summary":27063},[27058],{"title":27066,"path":27065,"lessonNumber":8377,"topics":29134,"summary":27067},[27058],{"module":27071,"moduleNumber":29136,"slug":29137,"lessons":29138},10,"practical",[29139,29141,29143,29145,29147,29149,29151],{"title":27070,"path":27069,"lessonNumber":8365,"topics":29140,"summary":27072},[27071],{"title":27075,"path":27074,"lessonNumber":8371,"topics":29142,"summary":27076},[27071],{"title":27079,"path":27078,"lessonNumber":8377,"topics":29144,"summary":27080},[27071],{"title":27083,"path":27082,"lessonNumber":8383,"topics":29146,"summary":27084},[27071],{"title":27087,"path":27086,"lessonNumber":8389,"topics":29148,"summary":27088},[27071],{"title":27091,"path":27090,"lessonNumber":8395,"topics":29150,"summary":27092},[27071],{"title":27095,"path":27094,"lessonNumber":8401,"topics":29152,"summary":27096},[27071],{"module":27100,"moduleNumber":29154,"slug":29155,"lessons":29156},11,"large-models-and-agents",[29157,29159,29161,29163,29165,29167,29169,29171,29173,29175,29177],{"title":27099,"path":27098,"lessonNumber":8365,"topics":29158,"summary":27101},[27100],{"title":27104,"path":27103,"lessonNumber":8371,"topics":29160,"summary":27105},[27100],{"title":27108,"path":27107,"lessonNumber":8377,"topics":29162,"summary":27109},[27100],{"title":27112,"path":27111,"lessonNumber":8383,"topics":29164,"summary":27113},[27100],{"title":27116,"path":27115,"lessonNumber":8389,"topics":29166,"summary":27117},[27100],{"title":27120,"path":27119,"lessonNumber":8395,"topics":29168,"summary":27121},[27100],{"title":27124,"path":27123,"lessonNumber":8401,"topics":29170,"summary":27125},[27100],{"title":27128,"path":27127,"lessonNumber":8407,"topics":29172,"summary":27129},[27100],{"title":27132,"path":27131,"lessonNumber":8413,"topics":29174,"summary":27133},[27100],{"title":27136,"path":27135,"lessonNumber":29136,"topics":29176,"summary":27137},[27100],{"title":27140,"path":27139,"lessonNumber":29154,"topics":29178,"summary":27141},[27100],{"module":27145,"moduleNumber":29180,"slug":29181,"lessons":29182},12,"reinforcement-learning",[29183,29185,29187,29189,29191],{"title":27144,"path":27143,"lessonNumber":8365,"topics":29184,"summary":27146},[27145],{"title":27149,"path":27148,"lessonNumber":8371,"topics":29186,"summary":27150},[27145],{"title":27153,"path":27152,"lessonNumber":8377,"topics":29188,"summary":27154},[27145],{"title":27157,"path":27156,"lessonNumber":8383,"topics":29190,"summary":27158},[27145],{"title":27161,"path":27160,"lessonNumber":8389,"topics":29192,"summary":27162},[27145],"\u003Csvg style=\"width:100%;max-width:571.302px;height:auto\" xmlns=\"http:\u002F\u002Fwww.w3.org\u002F2000\u002Fsvg\" viewBox=\"-75 -75 428.477 179.962\">\u003Cg stroke=\"currentColor\" style=\"stroke-miterlimit:10;stroke-width:.4\">\u003Cg stroke=\"var(--tk-accent)\">\u003Cpath fill=\"none\" d=\"M-65.403-46.463h45.524V-72.07h-45.524Z\"\u002F>\u003Cg transform=\"translate(-3.306 -72.022)\">\u003Cpath d=\"M-39.336 14.879Q-39.918 14.879-40.436 14.652Q-40.953 14.426-41.342 14.027Q-41.731 13.629-41.950 13.104Q-42.168 12.578-42.168 12.008Q-42.168 11.238-41.793 10.561Q-41.418 9.883-40.768 9.481Q-40.118 9.078-39.336 9.078Q-38.563 9.078-37.912 9.481Q-37.262 9.883-36.887 10.561Q-36.512 11.238-36.512 12.008Q-36.512 12.570-36.727 13.094Q-36.942 13.617-37.332 14.024Q-37.723 14.430-38.227 14.649Q-38.059 15.008-37.891 15.203Q-37.723 15.399-37.442 15.399Q-37.243 15.399-37.069 15.307Q-36.895 15.215-36.791 15.053Q-36.688 14.891-36.688 14.695Q-36.661 14.598-36.578 14.598Q-36.524 14.598-36.495 14.633Q-36.465 14.668-36.465 14.727Q-36.465 15.078-36.569 15.430Q-36.672 15.781-36.905 16.022Q-37.137 16.262-37.496 16.262Q-37.786 16.262-37.975 16.123Q-38.164 15.984-38.274 15.764Q-38.383 15.543-38.446 15.279Q-38.508 15.016-38.563 14.766Q-38.938 14.879-39.336 14.879M-39.801 14.176Q-39.801 14.367-39.670 14.482Q-39.539 14.598-39.336 14.598Q-38.977 14.598-38.641 14.445Q-38.731 14.129-38.889 13.936Q-39.047 13.742-39.336 13.742Q-39.524 13.742-39.662 13.865Q-39.801 13.988-39.801 14.176M-40.055 14.176Q-40.055 13.891-39.842 13.690Q-39.629 13.488-39.336 13.488Q-38.985 13.488-38.768 13.713Q-38.551 13.938-38.375 14.313Q-38.012 14.078-37.784 13.709Q-37.555 13.340-37.453 12.906Q-37.352 12.473-37.352 12.008Q-37.352 11.527-37.465 11.063Q-37.578 10.598-37.823 10.209Q-38.067 9.820-38.448 9.590Q-38.828 9.359-39.336 9.359Q-39.844 9.359-40.231 9.592Q-40.618 9.824-40.854 10.203Q-41.090 10.582-41.209 11.055Q-41.328 11.527-41.328 12.008Q-41.328 12.531-41.196 13.024Q-41.063 13.516-40.762 13.899Q-40.461 14.281-39.985 14.473Q-40.055 14.340-40.055 14.176\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"var(--tk-accent)\">\u003Cpath fill=\"none\" d=\"M-65.403 27.515h45.524V1.907h-45.524Z\"\u002F>\u003Cg transform=\"translate(-3.301 2.733)\">\u003Cpath d=\"M-39.922 14.711L-42.305 14.711L-42.305 14.414Q-41.985 14.414-41.741 14.367Q-41.496 14.320-41.496 14.152L-41.496 9.809Q-41.496 9.637-41.741 9.590Q-41.985 9.543-42.305 9.543L-42.305 9.246L-39.922 9.246L-39.922 9.543Q-40.243 9.543-40.487 9.592Q-40.731 9.641-40.731 9.809L-40.731 12.359L-38.114 10.024Q-37.977 9.902-37.977 9.774Q-37.977 9.664-38.065 9.604Q-38.153 9.543-38.274 9.543L-38.274 9.246L-36.520 9.246L-36.520 9.543Q-37.157 9.543-37.688 10.024L-39.207 11.383L-37.344 13.934Q-37.200 14.137-37.084 14.240Q-36.969 14.344-36.821 14.379Q-36.672 14.414-36.411 14.414L-36.411 14.711L-38.489 14.711L-38.489 14.414Q-38.075 14.414-38.075 14.207Q-38.075 14.098-38.200 13.934L-39.731 11.840L-40.731 12.734L-40.731 14.152Q-40.731 14.316-40.487 14.365Q-40.243 14.414-39.922 14.414\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"var(--tk-accent)\">\u003Cpath fill=\"none\" d=\"M-65.403 101.492h45.524V75.885h-45.524Z\"\u002F>\u003Cg transform=\"translate(-3.183 76.71)\">\u003Cpath d=\"M-39.664 14.766L-41.688 9.809Q-41.770 9.637-41.963 9.590Q-42.157 9.543-42.465 9.543L-42.465 9.246L-40.274 9.246L-40.274 9.543Q-40.899 9.543-40.899 9.758Q-40.895 9.774-40.893 9.787Q-40.891 9.801-40.887 9.809L-39.227 13.902L-37.641 10.024Q-37.618 9.977-37.618 9.910Q-37.618 9.727-37.787 9.635Q-37.957 9.543-38.168 9.543L-38.168 9.246L-36.457 9.246L-36.457 9.543Q-36.754 9.543-36.985 9.658Q-37.215 9.774-37.321 10.024L-39.258 14.766Q-39.297 14.879-39.426 14.879L-39.496 14.879Q-39.625 14.879-39.664 14.766\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"M15.687-9.474h42.68V-35.08h-42.68Z\"\u002F>\u003Cg transform=\"translate(66.918 -34.544)\">\u003Cpath d=\"M-42.696 14.473L-42.696 14.383Q-42.645 14.172-42.450 14.152L-42.250 14.152L-42.250 11.824L-42.450 11.824Q-42.657 11.801-42.696 11.582L-42.696 11.496Q-42.645 11.281-42.450 11.262L-41.969 11.262Q-41.778 11.285-41.719 11.496Q-41.399 11.223-40.985 11.223Q-40.793 11.223-40.625 11.332Q-40.457 11.441-40.383 11.613Q-40.207 11.422-39.985 11.322Q-39.762 11.223-39.520 11.223Q-39.102 11.223-38.948 11.565Q-38.793 11.906-38.793 12.375L-38.793 14.152L-38.594 14.152Q-38.383 14.176-38.344 14.383L-38.344 14.473Q-38.395 14.688-38.594 14.711L-39.411 14.711Q-39.606 14.688-39.657 14.473L-39.657 14.383Q-39.606 14.176-39.411 14.152L-39.321 14.152L-39.321 12.406Q-39.321 11.781-39.571 11.781Q-39.903 11.781-40.080 12.092Q-40.258 12.402-40.258 12.766L-40.258 14.152L-40.055 14.152Q-39.848 14.176-39.809 14.383L-39.809 14.473Q-39.860 14.688-40.055 14.711L-40.871 14.711Q-41.071 14.688-41.121 14.473L-41.121 14.383Q-41.071 14.176-40.871 14.152L-40.786 14.152L-40.786 12.406Q-40.786 11.781-41.032 11.781Q-41.364 11.781-41.541 12.094Q-41.719 12.406-41.719 12.766L-41.719 14.152L-41.520 14.152Q-41.313 14.176-41.274 14.383L-41.274 14.473Q-41.325 14.691-41.520 14.711L-42.450 14.711Q-42.657 14.688-42.696 14.473M-37.989 13.598Q-37.989 13.152-37.575 12.895Q-37.161 12.637-36.620 12.537Q-36.078 12.438-35.571 12.430Q-35.571 12.215-35.705 12.063Q-35.840 11.910-36.047 11.834Q-36.254 11.758-36.465 11.758Q-36.809 11.758-36.969 11.781L-36.969 11.840Q-36.969 12.008-37.088 12.123Q-37.207 12.238-37.371 12.238Q-37.547 12.238-37.662 12.115Q-37.778 11.992-37.778 11.824Q-37.778 11.418-37.397 11.309Q-37.016 11.199-36.457 11.199Q-36.188 11.199-35.920 11.277Q-35.653 11.356-35.428 11.506Q-35.203 11.656-35.067 11.877Q-34.930 12.098-34.930 12.375L-34.930 14.094Q-34.930 14.152-34.403 14.152Q-34.207 14.172-34.157 14.383L-34.157 14.473Q-34.207 14.688-34.403 14.711L-34.547 14.711Q-34.891 14.711-35.120 14.664Q-35.348 14.617-35.493 14.430Q-35.953 14.750-36.661 14.750Q-36.996 14.750-37.301 14.609Q-37.606 14.469-37.797 14.207Q-37.989 13.945-37.989 13.598M-37.348 13.606Q-37.348 13.879-37.106 14.035Q-36.864 14.191-36.578 14.191Q-36.360 14.191-36.127 14.133Q-35.895 14.074-35.733 13.936Q-35.571 13.797-35.571 13.574L-35.571 12.984Q-35.852 12.984-36.268 13.041Q-36.684 13.098-37.016 13.236Q-37.348 13.375-37.348 13.606M-32.950 13.606L-32.950 11.824L-33.700 11.824Q-33.899 11.801-33.950 11.582L-33.950 11.496Q-33.899 11.285-33.700 11.262L-32.950 11.262L-32.950 10.512Q-32.899 10.305-32.700 10.277L-32.555 10.277Q-32.360 10.305-32.309 10.512L-32.309 11.262L-30.950 11.262Q-30.758 11.281-30.700 11.496L-30.700 11.582Q-30.754 11.801-30.950 11.824L-32.309 11.824L-32.309 13.574Q-32.309 14.191-31.735 14.191Q-31.485 14.191-31.321 14.006Q-31.157 13.820-31.157 13.574Q-31.157 13.481-31.084 13.410Q-31.012 13.340-30.911 13.328L-30.766 13.328Q-30.567 13.352-30.516 13.559L-30.516 13.606Q-30.516 13.930-30.700 14.193Q-30.883 14.457-31.176 14.604Q-31.469 14.750-31.789 14.750Q-32.301 14.750-32.625 14.440Q-32.950 14.129-32.950 13.606M-29.957 14.473L-29.957 14.383Q-29.907 14.172-29.711 14.152L-29.512 14.152L-29.512 11.824L-29.711 11.824Q-29.918 11.801-29.957 11.582L-29.957 11.496Q-29.907 11.281-29.711 11.262L-29.231 11.262Q-29.039 11.285-28.981 11.496Q-28.661 11.223-28.246 11.223Q-28.055 11.223-27.887 11.332Q-27.719 11.441-27.645 11.613Q-27.469 11.422-27.246 11.322Q-27.024 11.223-26.782 11.223Q-26.364 11.223-26.209 11.565Q-26.055 11.906-26.055 12.375L-26.055 14.152L-25.856 14.152Q-25.645 14.176-25.606 14.383L-25.606 14.473Q-25.657 14.688-25.856 14.711L-26.672 14.711Q-26.868 14.688-26.918 14.473L-26.918 14.383Q-26.868 14.176-26.672 14.152L-26.582 14.152L-26.582 12.406Q-26.582 11.781-26.832 11.781Q-27.164 11.781-27.342 12.092Q-27.520 12.402-27.520 12.766L-27.520 14.152L-27.317 14.152Q-27.110 14.176-27.071 14.383L-27.071 14.473Q-27.121 14.688-27.317 14.711L-28.133 14.711Q-28.332 14.688-28.383 14.473L-28.383 14.383Q-28.332 14.176-28.133 14.152L-28.047 14.152L-28.047 12.406Q-28.047 11.781-28.293 11.781Q-28.625 11.781-28.803 12.094Q-28.981 12.406-28.981 12.766L-28.981 14.152L-28.782 14.152Q-28.575 14.176-28.536 14.383L-28.536 14.473Q-28.586 14.691-28.782 14.711L-29.711 14.711Q-29.918 14.688-29.957 14.473M-24.914 13.856L-24.914 11.824L-25.336 11.824Q-25.543 11.801-25.586 11.582L-25.586 11.496Q-25.539 11.285-25.336 11.262L-24.520 11.262Q-24.325 11.285-24.274 11.496L-24.274 13.824Q-24.274 14.059-24.104 14.125Q-23.934 14.191-23.649 14.191Q-23.442 14.191-23.246 14.115Q-23.051 14.039-22.926 13.889Q-22.801 13.738-22.801 13.527L-22.801 11.824L-23.223 11.824Q-23.434 11.801-23.473 11.582L-23.473 11.496Q-23.434 11.285-23.223 11.262L-22.411 11.262Q-22.211 11.285-22.161 11.496L-22.161 14.152L-21.735 14.152Q-21.528 14.176-21.489 14.383L-21.489 14.473Q-21.528 14.688-21.735 14.711L-22.551 14.711Q-22.750 14.688-22.801 14.488Q-23.203 14.750-23.711 14.750Q-23.946 14.750-24.161 14.709Q-24.375 14.668-24.541 14.566Q-24.707 14.465-24.811 14.287Q-24.914 14.109-24.914 13.856M-20.961 14.473L-20.961 14.383Q-20.911 14.176-20.715 14.152L-19.610 14.152L-19.610 10.383L-20.715 10.383Q-20.911 10.359-20.961 10.145L-20.961 10.055Q-20.911 9.848-20.715 9.824L-19.219 9.824Q-19.028 9.848-18.969 10.055L-18.969 14.152L-17.868 14.152Q-17.668 14.176-17.618 14.383L-17.618 14.473Q-17.668 14.688-17.868 14.711L-20.715 14.711Q-20.911 14.688-20.961 14.473\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"M88.241-9.474h45.525V-35.08H88.24Z\"\u002F>\u003Cg transform=\"translate(143.02 -34.544)\">\u003Cpath d=\"M-42.055 14.512L-42.055 13.598Q-42.028 13.391-41.817 13.367L-41.649 13.367Q-41.485 13.391-41.426 13.551Q-41.223 14.191-40.496 14.191Q-40.289 14.191-40.061 14.156Q-39.832 14.121-39.664 14.006Q-39.496 13.891-39.496 13.688Q-39.496 13.477-39.719 13.363Q-39.942 13.250-40.215 13.207L-40.914 13.094Q-42.055 12.883-42.055 12.160Q-42.055 11.871-41.911 11.682Q-41.766 11.492-41.526 11.385Q-41.286 11.277-41.030 11.238Q-40.774 11.199-40.496 11.199Q-40.246 11.199-40.053 11.229Q-39.860 11.258-39.696 11.336Q-39.618 11.219-39.489 11.199L-39.411 11.199Q-39.313 11.211-39.250 11.274Q-39.188 11.336-39.176 11.430L-39.176 12.137Q-39.188 12.231-39.250 12.297Q-39.313 12.363-39.411 12.375L-39.578 12.375Q-39.672 12.363-39.739 12.297Q-39.805 12.231-39.817 12.137Q-39.817 11.758-40.512 11.758Q-40.860 11.758-41.178 11.840Q-41.496 11.922-41.496 12.168Q-41.496 12.434-40.825 12.543L-40.121 12.664Q-39.637 12.746-39.287 12.994Q-38.938 13.242-38.938 13.688Q-38.938 14.078-39.174 14.320Q-39.411 14.563-39.760 14.656Q-40.110 14.750-40.496 14.750Q-41.075 14.750-41.473 14.496Q-41.543 14.621-41.592 14.678Q-41.641 14.734-41.746 14.750L-41.817 14.750Q-42.032 14.727-42.055 14.512M-37.836 12.984Q-37.836 12.504-37.592 12.090Q-37.348 11.676-36.932 11.438Q-36.516 11.199-36.036 11.199Q-35.481 11.199-35.102 11.309Q-34.723 11.418-34.723 11.824Q-34.723 11.992-34.836 12.115Q-34.950 12.238-35.121 12.238Q-35.293 12.238-35.412 12.123Q-35.532 12.008-35.532 11.840L-35.532 11.781Q-35.692 11.758-36.028 11.758Q-36.356 11.758-36.623 11.928Q-36.891 12.098-37.043 12.381Q-37.196 12.664-37.196 12.984Q-37.196 13.305-37.024 13.586Q-36.852 13.867-36.567 14.029Q-36.282 14.191-35.953 14.191Q-35.641 14.191-35.514 14.088Q-35.387 13.984-35.270 13.795Q-35.153 13.606-35.036 13.590L-34.868 13.590Q-34.762 13.602-34.698 13.670Q-34.633 13.738-34.633 13.840Q-34.633 13.887-34.653 13.926Q-34.762 14.219-34.965 14.399Q-35.168 14.578-35.444 14.664Q-35.719 14.750-36.036 14.750Q-36.520 14.750-36.936 14.512Q-37.352 14.274-37.594 13.871Q-37.836 13.469-37.836 12.984M-33.743 13.598Q-33.743 13.152-33.328 12.895Q-32.914 12.637-32.373 12.537Q-31.832 12.438-31.325 12.430Q-31.325 12.215-31.459 12.063Q-31.594 11.910-31.801 11.834Q-32.008 11.758-32.219 11.758Q-32.563 11.758-32.723 11.781L-32.723 11.840Q-32.723 12.008-32.842 12.123Q-32.961 12.238-33.125 12.238Q-33.301 12.238-33.416 12.115Q-33.532 11.992-33.532 11.824Q-33.532 11.418-33.151 11.309Q-32.770 11.199-32.211 11.199Q-31.942 11.199-31.674 11.277Q-31.407 11.356-31.182 11.506Q-30.957 11.656-30.821 11.877Q-30.684 12.098-30.684 12.375L-30.684 14.094Q-30.684 14.152-30.157 14.152Q-29.961 14.172-29.911 14.383L-29.911 14.473Q-29.961 14.688-30.157 14.711L-30.301 14.711Q-30.645 14.711-30.873 14.664Q-31.102 14.617-31.246 14.430Q-31.707 14.750-32.414 14.750Q-32.750 14.750-33.055 14.609Q-33.360 14.469-33.551 14.207Q-33.743 13.945-33.743 13.598M-33.102 13.606Q-33.102 13.879-32.860 14.035Q-32.618 14.191-32.332 14.191Q-32.114 14.191-31.881 14.133Q-31.649 14.074-31.487 13.936Q-31.325 13.797-31.325 13.574L-31.325 12.984Q-31.606 12.984-32.022 13.041Q-32.438 13.098-32.770 13.236Q-33.102 13.375-33.102 13.606M-29.453 14.473L-29.453 14.383Q-29.403 14.176-29.207 14.152L-28.102 14.152L-28.102 10.383L-29.207 10.383Q-29.403 10.359-29.453 10.145L-29.453 10.055Q-29.403 9.848-29.207 9.824L-27.711 9.824Q-27.520 9.848-27.461 10.055L-27.461 14.152L-26.360 14.152Q-26.161 14.176-26.110 14.383L-26.110 14.473Q-26.161 14.688-26.360 14.711L-29.207 14.711Q-29.403 14.688-29.453 14.473M-22.145 13.223L-24.586 13.223Q-24.532 13.500-24.334 13.723Q-24.137 13.945-23.860 14.068Q-23.582 14.191-23.297 14.191Q-22.825 14.191-22.602 13.902Q-22.594 13.891-22.537 13.785Q-22.481 13.680-22.432 13.637Q-22.383 13.594-22.289 13.582L-22.145 13.582Q-21.953 13.602-21.895 13.816L-21.895 13.871Q-21.961 14.172-22.192 14.369Q-22.422 14.566-22.735 14.658Q-23.047 14.750-23.352 14.750Q-23.836 14.750-24.276 14.522Q-24.715 14.293-24.983 13.893Q-25.250 13.492-25.250 13L-25.250 12.941Q-25.250 12.473-25.004 12.070Q-24.758 11.668-24.350 11.434Q-23.942 11.199-23.473 11.199Q-22.969 11.199-22.616 11.422Q-22.262 11.645-22.078 12.033Q-21.895 12.422-21.895 12.926L-21.895 12.984Q-21.953 13.199-22.145 13.223M-24.578 12.672L-22.551 12.672Q-22.598 12.262-22.836 12.010Q-23.075 11.758-23.473 11.758Q-23.868 11.758-24.174 12.020Q-24.481 12.281-24.578 12.672\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"M159.373-9.474h45.525V-35.08h-45.525Z\"\u002F>\u003Cg transform=\"translate(209.901 -34.544)\">\u003Cpath d=\"M-42.055 14.512L-42.055 13.598Q-42.028 13.391-41.817 13.367L-41.649 13.367Q-41.485 13.391-41.426 13.551Q-41.223 14.191-40.496 14.191Q-40.289 14.191-40.061 14.156Q-39.832 14.121-39.664 14.006Q-39.496 13.891-39.496 13.688Q-39.496 13.477-39.719 13.363Q-39.942 13.250-40.215 13.207L-40.914 13.094Q-42.055 12.883-42.055 12.160Q-42.055 11.871-41.911 11.682Q-41.766 11.492-41.526 11.385Q-41.286 11.277-41.030 11.238Q-40.774 11.199-40.496 11.199Q-40.246 11.199-40.053 11.229Q-39.860 11.258-39.696 11.336Q-39.618 11.219-39.489 11.199L-39.411 11.199Q-39.313 11.211-39.250 11.274Q-39.188 11.336-39.176 11.430L-39.176 12.137Q-39.188 12.231-39.250 12.297Q-39.313 12.363-39.411 12.375L-39.578 12.375Q-39.672 12.363-39.739 12.297Q-39.805 12.231-39.817 12.137Q-39.817 11.758-40.512 11.758Q-40.860 11.758-41.178 11.840Q-41.496 11.922-41.496 12.168Q-41.496 12.434-40.825 12.543L-40.121 12.664Q-39.637 12.746-39.287 12.994Q-38.938 13.242-38.938 13.688Q-38.938 14.078-39.174 14.320Q-39.411 14.563-39.760 14.656Q-40.110 14.750-40.496 14.750Q-41.075 14.750-41.473 14.496Q-41.543 14.621-41.592 14.678Q-41.641 14.734-41.746 14.750L-41.817 14.750Q-42.032 14.727-42.055 14.512M-36.274 14.750Q-36.746 14.750-37.131 14.506Q-37.516 14.262-37.739 13.852Q-37.961 13.441-37.961 12.984Q-37.961 12.641-37.836 12.318Q-37.711 11.996-37.481 11.742Q-37.250 11.488-36.944 11.344Q-36.637 11.199-36.274 11.199Q-35.911 11.199-35.598 11.346Q-35.286 11.492-35.063 11.738Q-34.840 11.984-34.713 12.305Q-34.586 12.625-34.586 12.984Q-34.586 13.441-34.811 13.854Q-35.036 14.266-35.420 14.508Q-35.805 14.750-36.274 14.750M-36.274 14.191Q-35.809 14.191-35.518 13.797Q-35.227 13.402-35.227 12.918Q-35.227 12.625-35.362 12.357Q-35.496 12.090-35.737 11.924Q-35.977 11.758-36.274 11.758Q-36.578 11.758-36.817 11.924Q-37.055 12.090-37.190 12.357Q-37.325 12.625-37.325 12.918Q-37.325 13.399-37.032 13.795Q-36.739 14.191-36.274 14.191M-33.828 14.473L-33.828 14.383Q-33.778 14.176-33.582 14.152L-32.700 14.152L-32.700 11.824L-33.555 11.824Q-33.754 11.801-33.805 11.582L-33.805 11.496Q-33.754 11.285-33.555 11.262L-32.700 11.262L-32.700 10.809Q-32.700 10.344-32.293 10.063Q-31.887 9.781-31.407 9.781Q-31.094 9.781-30.850 9.902Q-30.606 10.024-30.606 10.305Q-30.606 10.469-30.715 10.586Q-30.825 10.703-30.989 10.703Q-31.137 10.703-31.258 10.598Q-31.379 10.492-31.379 10.344L-31.461 10.344Q-31.614 10.344-31.750 10.404Q-31.887 10.465-31.973 10.580Q-32.059 10.695-32.059 10.840L-32.059 11.262L-31.028 11.262Q-30.832 11.281-30.782 11.496L-30.782 11.582Q-30.832 11.801-31.028 11.824L-32.059 11.824L-32.059 14.152L-31.180 14.152Q-30.985 14.176-30.934 14.383L-30.934 14.473Q-30.985 14.688-31.180 14.711L-33.582 14.711Q-33.778 14.688-33.828 14.473M-28.703 13.606L-28.703 11.824L-29.453 11.824Q-29.653 11.801-29.703 11.582L-29.703 11.496Q-29.653 11.285-29.453 11.262L-28.703 11.262L-28.703 10.512Q-28.653 10.305-28.453 10.277L-28.309 10.277Q-28.114 10.305-28.063 10.512L-28.063 11.262L-26.703 11.262Q-26.512 11.281-26.453 11.496L-26.453 11.582Q-26.508 11.801-26.703 11.824L-28.063 11.824L-28.063 13.574Q-28.063 14.191-27.489 14.191Q-27.239 14.191-27.075 14.006Q-26.911 13.820-26.911 13.574Q-26.911 13.481-26.838 13.410Q-26.766 13.340-26.664 13.328L-26.520 13.328Q-26.321 13.352-26.270 13.559L-26.270 13.606Q-26.270 13.930-26.453 14.193Q-26.637 14.457-26.930 14.604Q-27.223 14.750-27.543 14.750Q-28.055 14.750-28.379 14.440Q-28.703 14.129-28.703 13.606M-25.711 14.473L-25.711 14.383Q-25.661 14.172-25.465 14.152L-25.266 14.152L-25.266 11.824L-25.465 11.824Q-25.672 11.801-25.711 11.582L-25.711 11.496Q-25.661 11.281-25.465 11.262L-24.985 11.262Q-24.793 11.285-24.735 11.496Q-24.414 11.223-24 11.223Q-23.809 11.223-23.641 11.332Q-23.473 11.441-23.399 11.613Q-23.223 11.422-23 11.322Q-22.778 11.223-22.536 11.223Q-22.118 11.223-21.963 11.565Q-21.809 11.906-21.809 12.375L-21.809 14.152L-21.610 14.152Q-21.399 14.176-21.360 14.383L-21.360 14.473Q-21.411 14.688-21.610 14.711L-22.426 14.711Q-22.621 14.688-22.672 14.473L-22.672 14.383Q-22.621 14.176-22.426 14.152L-22.336 14.152L-22.336 12.406Q-22.336 11.781-22.586 11.781Q-22.918 11.781-23.096 12.092Q-23.274 12.402-23.274 12.766L-23.274 14.152L-23.071 14.152Q-22.864 14.176-22.825 14.383L-22.825 14.473Q-22.875 14.688-23.071 14.711L-23.887 14.711Q-24.086 14.688-24.137 14.473L-24.137 14.383Q-24.086 14.176-23.887 14.152L-23.801 14.152L-23.801 12.406Q-23.801 11.781-24.047 11.781Q-24.379 11.781-24.557 12.094Q-24.735 12.406-24.735 12.766L-24.735 14.152L-24.536 14.152Q-24.328 14.176-24.289 14.383L-24.289 14.473Q-24.340 14.691-24.536 14.711L-25.465 14.711Q-25.672 14.688-25.711 14.473M-21.004 13.598Q-21.004 13.152-20.590 12.895Q-20.176 12.637-19.635 12.537Q-19.094 12.438-18.586 12.430Q-18.586 12.215-18.721 12.063Q-18.856 11.910-19.063 11.834Q-19.270 11.758-19.481 11.758Q-19.825 11.758-19.985 11.781L-19.985 11.840Q-19.985 12.008-20.104 12.123Q-20.223 12.238-20.387 12.238Q-20.563 12.238-20.678 12.115Q-20.793 11.992-20.793 11.824Q-20.793 11.418-20.412 11.309Q-20.032 11.199-19.473 11.199Q-19.203 11.199-18.936 11.277Q-18.668 11.356-18.444 11.506Q-18.219 11.656-18.082 11.877Q-17.946 12.098-17.946 12.375L-17.946 14.094Q-17.946 14.152-17.418 14.152Q-17.223 14.172-17.172 14.383L-17.172 14.473Q-17.223 14.688-17.418 14.711L-17.563 14.711Q-17.907 14.711-18.135 14.664Q-18.364 14.617-18.508 14.430Q-18.969 14.750-19.676 14.750Q-20.012 14.750-20.317 14.609Q-20.621 14.469-20.813 14.207Q-21.004 13.945-21.004 13.598M-20.364 13.606Q-20.364 13.879-20.121 14.035Q-19.879 14.191-19.594 14.191Q-19.375 14.191-19.143 14.133Q-18.911 14.074-18.748 13.936Q-18.586 13.797-18.586 13.574L-18.586 12.984Q-18.868 12.984-19.284 13.041Q-19.700 13.098-20.032 13.236Q-20.364 13.375-20.364 13.606M-16.907 14.473L-16.907 14.383Q-16.868 14.176-16.661 14.152L-16.254 14.152L-15.325 12.934L-16.196 11.824L-16.614 11.824Q-16.809 11.805-16.860 11.582L-16.860 11.496Q-16.809 11.285-16.614 11.262L-15.453 11.262Q-15.254 11.285-15.203 11.496L-15.203 11.582Q-15.254 11.801-15.453 11.824L-15.563 11.824L-15.067 12.481L-14.590 11.824L-14.707 11.824Q-14.907 11.801-14.957 11.582L-14.957 11.496Q-14.907 11.285-14.707 11.262L-13.547 11.262Q-13.352 11.281-13.301 11.496L-13.301 11.582Q-13.352 11.801-13.547 11.824L-13.957 11.824L-14.805 12.934L-13.844 14.152L-13.438 14.152Q-13.239 14.176-13.188 14.383L-13.188 14.473Q-13.227 14.688-13.438 14.711L-14.590 14.711Q-14.797 14.688-14.836 14.473L-14.836 14.383Q-14.797 14.176-14.590 14.152L-14.461 14.152L-15.067 13.297L-15.653 14.152L-15.508 14.152Q-15.301 14.176-15.262 14.383L-15.262 14.473Q-15.301 14.688-15.508 14.711L-16.661 14.711Q-16.856 14.691-16.907 14.473\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"M231.928 44.587h42.679V18.979h-42.68Z\"\u002F>\u003Cg transform=\"translate(283.158 19.516)\">\u003Cpath d=\"M-42.696 14.473L-42.696 14.383Q-42.645 14.172-42.450 14.152L-42.250 14.152L-42.250 11.824L-42.450 11.824Q-42.657 11.801-42.696 11.582L-42.696 11.496Q-42.645 11.281-42.450 11.262L-41.969 11.262Q-41.778 11.285-41.719 11.496Q-41.399 11.223-40.985 11.223Q-40.793 11.223-40.625 11.332Q-40.457 11.441-40.383 11.613Q-40.207 11.422-39.985 11.322Q-39.762 11.223-39.520 11.223Q-39.102 11.223-38.948 11.565Q-38.793 11.906-38.793 12.375L-38.793 14.152L-38.594 14.152Q-38.383 14.176-38.344 14.383L-38.344 14.473Q-38.395 14.688-38.594 14.711L-39.411 14.711Q-39.606 14.688-39.657 14.473L-39.657 14.383Q-39.606 14.176-39.411 14.152L-39.321 14.152L-39.321 12.406Q-39.321 11.781-39.571 11.781Q-39.903 11.781-40.080 12.092Q-40.258 12.402-40.258 12.766L-40.258 14.152L-40.055 14.152Q-39.848 14.176-39.809 14.383L-39.809 14.473Q-39.860 14.688-40.055 14.711L-40.871 14.711Q-41.071 14.688-41.121 14.473L-41.121 14.383Q-41.071 14.176-40.871 14.152L-40.786 14.152L-40.786 12.406Q-40.786 11.781-41.032 11.781Q-41.364 11.781-41.541 12.094Q-41.719 12.406-41.719 12.766L-41.719 14.152L-41.520 14.152Q-41.313 14.176-41.274 14.383L-41.274 14.473Q-41.325 14.691-41.520 14.711L-42.450 14.711Q-42.657 14.688-42.696 14.473M-37.989 13.598Q-37.989 13.152-37.575 12.895Q-37.161 12.637-36.620 12.537Q-36.078 12.438-35.571 12.430Q-35.571 12.215-35.705 12.063Q-35.840 11.910-36.047 11.834Q-36.254 11.758-36.465 11.758Q-36.809 11.758-36.969 11.781L-36.969 11.840Q-36.969 12.008-37.088 12.123Q-37.207 12.238-37.371 12.238Q-37.547 12.238-37.662 12.115Q-37.778 11.992-37.778 11.824Q-37.778 11.418-37.397 11.309Q-37.016 11.199-36.457 11.199Q-36.188 11.199-35.920 11.277Q-35.653 11.356-35.428 11.506Q-35.203 11.656-35.067 11.877Q-34.930 12.098-34.930 12.375L-34.930 14.094Q-34.930 14.152-34.403 14.152Q-34.207 14.172-34.157 14.383L-34.157 14.473Q-34.207 14.688-34.403 14.711L-34.547 14.711Q-34.891 14.711-35.120 14.664Q-35.348 14.617-35.493 14.430Q-35.953 14.750-36.661 14.750Q-36.996 14.750-37.301 14.609Q-37.606 14.469-37.797 14.207Q-37.989 13.945-37.989 13.598M-37.348 13.606Q-37.348 13.879-37.106 14.035Q-36.864 14.191-36.578 14.191Q-36.360 14.191-36.127 14.133Q-35.895 14.074-35.733 13.936Q-35.571 13.797-35.571 13.574L-35.571 12.984Q-35.852 12.984-36.268 13.041Q-36.684 13.098-37.016 13.236Q-37.348 13.375-37.348 13.606M-32.950 13.606L-32.950 11.824L-33.700 11.824Q-33.899 11.801-33.950 11.582L-33.950 11.496Q-33.899 11.285-33.700 11.262L-32.950 11.262L-32.950 10.512Q-32.899 10.305-32.700 10.277L-32.555 10.277Q-32.360 10.305-32.309 10.512L-32.309 11.262L-30.950 11.262Q-30.758 11.281-30.700 11.496L-30.700 11.582Q-30.754 11.801-30.950 11.824L-32.309 11.824L-32.309 13.574Q-32.309 14.191-31.735 14.191Q-31.485 14.191-31.321 14.006Q-31.157 13.820-31.157 13.574Q-31.157 13.481-31.084 13.410Q-31.012 13.340-30.911 13.328L-30.766 13.328Q-30.567 13.352-30.516 13.559L-30.516 13.606Q-30.516 13.930-30.700 14.193Q-30.883 14.457-31.176 14.604Q-31.469 14.750-31.789 14.750Q-32.301 14.750-32.625 14.440Q-32.950 14.129-32.950 13.606M-29.957 14.473L-29.957 14.383Q-29.907 14.172-29.711 14.152L-29.512 14.152L-29.512 11.824L-29.711 11.824Q-29.918 11.801-29.957 11.582L-29.957 11.496Q-29.907 11.281-29.711 11.262L-29.231 11.262Q-29.039 11.285-28.981 11.496Q-28.661 11.223-28.246 11.223Q-28.055 11.223-27.887 11.332Q-27.719 11.441-27.645 11.613Q-27.469 11.422-27.246 11.322Q-27.024 11.223-26.782 11.223Q-26.364 11.223-26.209 11.565Q-26.055 11.906-26.055 12.375L-26.055 14.152L-25.856 14.152Q-25.645 14.176-25.606 14.383L-25.606 14.473Q-25.657 14.688-25.856 14.711L-26.672 14.711Q-26.868 14.688-26.918 14.473L-26.918 14.383Q-26.868 14.176-26.672 14.152L-26.582 14.152L-26.582 12.406Q-26.582 11.781-26.832 11.781Q-27.164 11.781-27.342 12.092Q-27.520 12.402-27.520 12.766L-27.520 14.152L-27.317 14.152Q-27.110 14.176-27.071 14.383L-27.071 14.473Q-27.121 14.688-27.317 14.711L-28.133 14.711Q-28.332 14.688-28.383 14.473L-28.383 14.383Q-28.332 14.176-28.133 14.152L-28.047 14.152L-28.047 12.406Q-28.047 11.781-28.293 11.781Q-28.625 11.781-28.803 12.094Q-28.981 12.406-28.981 12.766L-28.981 14.152L-28.782 14.152Q-28.575 14.176-28.536 14.383L-28.536 14.473Q-28.586 14.691-28.782 14.711L-29.711 14.711Q-29.918 14.688-29.957 14.473M-24.914 13.856L-24.914 11.824L-25.336 11.824Q-25.543 11.801-25.586 11.582L-25.586 11.496Q-25.539 11.285-25.336 11.262L-24.520 11.262Q-24.325 11.285-24.274 11.496L-24.274 13.824Q-24.274 14.059-24.104 14.125Q-23.934 14.191-23.649 14.191Q-23.442 14.191-23.246 14.115Q-23.051 14.039-22.926 13.889Q-22.801 13.738-22.801 13.527L-22.801 11.824L-23.223 11.824Q-23.434 11.801-23.473 11.582L-23.473 11.496Q-23.434 11.285-23.223 11.262L-22.411 11.262Q-22.211 11.285-22.161 11.496L-22.161 14.152L-21.735 14.152Q-21.528 14.176-21.489 14.383L-21.489 14.473Q-21.528 14.688-21.735 14.711L-22.551 14.711Q-22.750 14.688-22.801 14.488Q-23.203 14.750-23.711 14.750Q-23.946 14.750-24.161 14.709Q-24.375 14.668-24.541 14.566Q-24.707 14.465-24.811 14.287Q-24.914 14.109-24.914 13.856M-20.961 14.473L-20.961 14.383Q-20.911 14.176-20.715 14.152L-19.610 14.152L-19.610 10.383L-20.715 10.383Q-20.911 10.359-20.961 10.145L-20.961 10.055Q-20.911 9.848-20.715 9.824L-19.219 9.824Q-19.028 9.848-18.969 10.055L-18.969 14.152L-17.868 14.152Q-17.668 14.176-17.618 14.383L-17.618 14.473Q-17.668 14.688-17.868 14.711L-20.715 14.711Q-20.911 14.688-20.961 14.473\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg stroke=\"var(--tk-accent)\">\u003Cpath fill=\"none\" d=\"M304.482 44.587h45.525V18.979h-45.525Z\"\u002F>\u003Cg transform=\"translate(357.135 18.397)\">\u003Cpath d=\"M-40.520 14.750Q-40.993 14.750-41.377 14.506Q-41.762 14.262-41.985 13.852Q-42.207 13.441-42.207 12.984Q-42.207 12.641-42.082 12.318Q-41.957 11.996-41.727 11.742Q-41.496 11.488-41.190 11.344Q-40.883 11.199-40.520 11.199Q-40.157 11.199-39.844 11.346Q-39.532 11.492-39.309 11.738Q-39.086 11.984-38.959 12.305Q-38.832 12.625-38.832 12.984Q-38.832 13.441-39.057 13.854Q-39.282 14.266-39.666 14.508Q-40.051 14.750-40.520 14.750M-40.520 14.191Q-40.055 14.191-39.764 13.797Q-39.473 13.402-39.473 12.918Q-39.473 12.625-39.608 12.357Q-39.743 12.090-39.983 11.924Q-40.223 11.758-40.520 11.758Q-40.825 11.758-41.063 11.924Q-41.301 12.090-41.436 12.357Q-41.571 12.625-41.571 12.918Q-41.571 13.399-41.278 13.795Q-40.985 14.191-40.520 14.191M-37.653 13.856L-37.653 11.824L-38.075 11.824Q-38.282 11.801-38.325 11.582L-38.325 11.496Q-38.278 11.285-38.075 11.262L-37.258 11.262Q-37.063 11.285-37.012 11.496L-37.012 13.824Q-37.012 14.059-36.842 14.125Q-36.672 14.191-36.387 14.191Q-36.180 14.191-35.985 14.115Q-35.789 14.039-35.664 13.889Q-35.539 13.738-35.539 13.527L-35.539 11.824L-35.961 11.824Q-36.172 11.801-36.211 11.582L-36.211 11.496Q-36.172 11.285-35.961 11.262L-35.149 11.262Q-34.950 11.285-34.899 11.496L-34.899 14.152L-34.473 14.152Q-34.266 14.176-34.227 14.383L-34.227 14.473Q-34.266 14.688-34.473 14.711L-35.289 14.711Q-35.489 14.688-35.539 14.488Q-35.942 14.750-36.450 14.750Q-36.684 14.750-36.899 14.709Q-37.114 14.668-37.280 14.566Q-37.446 14.465-37.549 14.287Q-37.653 14.109-37.653 13.856M-32.950 13.606L-32.950 11.824L-33.700 11.824Q-33.899 11.801-33.950 11.582L-33.950 11.496Q-33.899 11.285-33.700 11.262L-32.950 11.262L-32.950 10.512Q-32.899 10.305-32.700 10.277L-32.555 10.277Q-32.360 10.305-32.309 10.512L-32.309 11.262L-30.950 11.262Q-30.758 11.281-30.700 11.496L-30.700 11.582Q-30.754 11.801-30.950 11.824L-32.309 11.824L-32.309 13.574Q-32.309 14.191-31.735 14.191Q-31.485 14.191-31.321 14.006Q-31.157 13.820-31.157 13.574Q-31.157 13.481-31.084 13.410Q-31.012 13.340-30.911 13.328L-30.766 13.328Q-30.567 13.352-30.516 13.559L-30.516 13.606Q-30.516 13.930-30.700 14.193Q-30.883 14.457-31.176 14.604Q-31.469 14.750-31.789 14.750Q-32.301 14.750-32.625 14.440Q-32.950 14.129-32.950 13.606M-29.832 16.254L-29.832 16.168Q-29.789 15.949-29.582 15.926L-29.161 15.926L-29.161 11.824L-29.582 11.824Q-29.789 11.801-29.832 11.582L-29.832 11.496Q-29.786 11.285-29.582 11.262L-28.766 11.262Q-28.571 11.281-28.520 11.496L-28.520 11.566Q-28.309 11.399-28.045 11.311Q-27.782 11.223-27.512 11.223Q-27.172 11.223-26.875 11.367Q-26.578 11.512-26.364 11.764Q-26.149 12.016-26.034 12.328Q-25.918 12.641-25.918 12.984Q-25.918 13.449-26.145 13.859Q-26.371 14.270-26.758 14.510Q-27.145 14.750-27.614 14.750Q-28.133 14.750-28.520 14.383L-28.520 15.926L-28.094 15.926Q-27.887 15.949-27.848 16.168L-27.848 16.254Q-27.887 16.465-28.094 16.488L-29.582 16.488Q-29.786 16.465-29.832 16.254M-27.657 14.191Q-27.348 14.191-27.098 14.022Q-26.848 13.852-26.703 13.570Q-26.559 13.289-26.559 12.984Q-26.559 12.695-26.684 12.416Q-26.809 12.137-27.041 11.959Q-27.274 11.781-27.575 11.781Q-27.895 11.781-28.157 11.967Q-28.418 12.152-28.520 12.453L-28.520 13.305Q-28.430 13.672-28.209 13.932Q-27.989 14.191-27.657 14.191M-24.914 13.856L-24.914 11.824L-25.336 11.824Q-25.543 11.801-25.586 11.582L-25.586 11.496Q-25.539 11.285-25.336 11.262L-24.520 11.262Q-24.325 11.285-24.274 11.496L-24.274 13.824Q-24.274 14.059-24.104 14.125Q-23.934 14.191-23.649 14.191Q-23.442 14.191-23.246 14.115Q-23.051 14.039-22.926 13.889Q-22.801 13.738-22.801 13.527L-22.801 11.824L-23.223 11.824Q-23.434 11.801-23.473 11.582L-23.473 11.496Q-23.434 11.285-23.223 11.262L-22.411 11.262Q-22.211 11.285-22.161 11.496L-22.161 14.152L-21.735 14.152Q-21.528 14.176-21.489 14.383L-21.489 14.473Q-21.528 14.688-21.735 14.711L-22.551 14.711Q-22.750 14.688-22.801 14.488Q-23.203 14.750-23.711 14.750Q-23.946 14.750-24.161 14.709Q-24.375 14.668-24.541 14.566Q-24.707 14.465-24.811 14.287Q-24.914 14.109-24.914 13.856M-20.211 13.606L-20.211 11.824L-20.961 11.824Q-21.161 11.801-21.211 11.582L-21.211 11.496Q-21.161 11.285-20.961 11.262L-20.211 11.262L-20.211 10.512Q-20.161 10.305-19.961 10.277L-19.817 10.277Q-19.621 10.305-19.571 10.512L-19.571 11.262L-18.211 11.262Q-18.020 11.281-17.961 11.496L-17.961 11.582Q-18.016 11.801-18.211 11.824L-19.571 11.824L-19.571 13.574Q-19.571 14.191-18.996 14.191Q-18.746 14.191-18.582 14.006Q-18.418 13.820-18.418 13.574Q-18.418 13.481-18.346 13.410Q-18.274 13.340-18.172 13.328L-18.028 13.328Q-17.828 13.352-17.778 13.559L-17.778 13.606Q-17.778 13.930-17.961 14.193Q-18.145 14.457-18.438 14.604Q-18.731 14.750-19.051 14.750Q-19.563 14.750-19.887 14.440Q-20.211 14.129-20.211 13.606\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"m-19.679-48.606 32.808 15.24\"\u002F>\u003Cpath stroke=\"none\" d=\"m15.487-32.27-2.896-3.64.538 2.544-2.29 1.23\"\u002F>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"m-19.679 4.051 32.808-15.24\"\u002F>\u003Cpath stroke=\"none\" d=\"m15.487-12.285-4.649-.134 2.291 1.23-.538 2.543\"\u002F>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M58.566-22.278h26.875\"\u002F>\u003Cpath stroke=\"none\" d=\"m88.041-22.278-4.16-2.08 1.56 2.08-1.56 2.08\"\u002F>\u003Cg transform=\"translate(103.195 -40.722)\">\u003Cpath d=\"M-42.055 14.512L-42.055 13.598Q-42.028 13.391-41.817 13.367L-41.649 13.367Q-41.485 13.391-41.426 13.551Q-41.223 14.191-40.496 14.191Q-40.289 14.191-40.061 14.156Q-39.832 14.121-39.664 14.006Q-39.496 13.891-39.496 13.688Q-39.496 13.477-39.719 13.363Q-39.942 13.250-40.215 13.207L-40.914 13.094Q-42.055 12.883-42.055 12.160Q-42.055 11.871-41.911 11.682Q-41.766 11.492-41.526 11.385Q-41.286 11.277-41.030 11.238Q-40.774 11.199-40.496 11.199Q-40.246 11.199-40.053 11.229Q-39.860 11.258-39.696 11.336Q-39.618 11.219-39.489 11.199L-39.411 11.199Q-39.313 11.211-39.250 11.274Q-39.188 11.336-39.176 11.430L-39.176 12.137Q-39.188 12.231-39.250 12.297Q-39.313 12.363-39.411 12.375L-39.578 12.375Q-39.672 12.363-39.739 12.297Q-39.805 12.231-39.817 12.137Q-39.817 11.758-40.512 11.758Q-40.860 11.758-41.178 11.840Q-41.496 11.922-41.496 12.168Q-41.496 12.434-40.825 12.543L-40.121 12.664Q-39.637 12.746-39.287 12.994Q-38.938 13.242-38.938 13.688Q-38.938 14.078-39.174 14.320Q-39.411 14.563-39.760 14.656Q-40.110 14.750-40.496 14.750Q-41.075 14.750-41.473 14.496Q-41.543 14.621-41.592 14.678Q-41.641 14.734-41.746 14.750L-41.817 14.750Q-42.032 14.727-42.055 14.512M-37.836 12.984Q-37.836 12.504-37.592 12.090Q-37.348 11.676-36.932 11.438Q-36.516 11.199-36.036 11.199Q-35.481 11.199-35.102 11.309Q-34.723 11.418-34.723 11.824Q-34.723 11.992-34.836 12.115Q-34.950 12.238-35.121 12.238Q-35.293 12.238-35.412 12.123Q-35.532 12.008-35.532 11.840L-35.532 11.781Q-35.692 11.758-36.028 11.758Q-36.356 11.758-36.623 11.928Q-36.891 12.098-37.043 12.381Q-37.196 12.664-37.196 12.984Q-37.196 13.305-37.024 13.586Q-36.852 13.867-36.567 14.029Q-36.282 14.191-35.953 14.191Q-35.641 14.191-35.514 14.088Q-35.387 13.984-35.270 13.795Q-35.153 13.606-35.036 13.590L-34.868 13.590Q-34.762 13.602-34.698 13.670Q-34.633 13.738-34.633 13.840Q-34.633 13.887-34.653 13.926Q-34.762 14.219-34.965 14.399Q-35.168 14.578-35.444 14.664Q-35.719 14.750-36.036 14.750Q-36.520 14.750-36.936 14.512Q-37.352 14.274-37.594 13.871Q-37.836 13.469-37.836 12.984M-32.028 14.750Q-32.500 14.750-32.885 14.506Q-33.270 14.262-33.493 13.852Q-33.715 13.441-33.715 12.984Q-33.715 12.641-33.590 12.318Q-33.465 11.996-33.235 11.742Q-33.004 11.488-32.698 11.344Q-32.391 11.199-32.028 11.199Q-31.664 11.199-31.352 11.346Q-31.039 11.492-30.817 11.738Q-30.594 11.984-30.467 12.305Q-30.340 12.625-30.340 12.984Q-30.340 13.441-30.565 13.854Q-30.789 14.266-31.174 14.508Q-31.559 14.750-32.028 14.750M-32.028 14.191Q-31.563 14.191-31.272 13.797Q-30.981 13.402-30.981 12.918Q-30.981 12.625-31.116 12.357Q-31.250 12.090-31.491 11.924Q-31.731 11.758-32.028 11.758Q-32.332 11.758-32.571 11.924Q-32.809 12.090-32.944 12.357Q-33.078 12.625-33.078 12.918Q-33.078 13.399-32.786 13.795Q-32.493 14.191-32.028 14.191M-29.680 14.473L-29.680 14.383Q-29.621 14.176-29.430 14.152L-28.719 14.152L-28.719 11.824L-29.430 11.824Q-29.625 11.801-29.680 11.582L-29.680 11.496Q-29.621 11.285-29.430 11.262L-28.328 11.262Q-28.129 11.281-28.078 11.496L-28.078 11.824Q-27.817 11.539-27.461 11.381Q-27.106 11.223-26.719 11.223Q-26.426 11.223-26.192 11.357Q-25.957 11.492-25.957 11.758Q-25.957 11.926-26.067 12.043Q-26.176 12.160-26.344 12.160Q-26.496 12.160-26.612 12.049Q-26.727 11.938-26.727 11.781Q-27.102 11.781-27.416 11.982Q-27.731 12.184-27.905 12.518Q-28.078 12.852-28.078 13.231L-28.078 14.152L-27.133 14.152Q-26.926 14.176-26.887 14.383L-26.887 14.473Q-26.926 14.688-27.133 14.711L-29.430 14.711Q-29.621 14.688-29.680 14.473M-22.145 13.223L-24.586 13.223Q-24.532 13.500-24.334 13.723Q-24.137 13.945-23.860 14.068Q-23.582 14.191-23.297 14.191Q-22.825 14.191-22.602 13.902Q-22.594 13.891-22.537 13.785Q-22.481 13.680-22.432 13.637Q-22.383 13.594-22.289 13.582L-22.145 13.582Q-21.953 13.602-21.895 13.816L-21.895 13.871Q-21.961 14.172-22.192 14.369Q-22.422 14.566-22.735 14.658Q-23.047 14.750-23.352 14.750Q-23.836 14.750-24.276 14.522Q-24.715 14.293-24.983 13.893Q-25.250 13.492-25.250 13L-25.250 12.941Q-25.250 12.473-25.004 12.070Q-24.758 11.668-24.350 11.434Q-23.942 11.199-23.473 11.199Q-22.969 11.199-22.616 11.422Q-22.262 11.645-22.078 12.033Q-21.895 12.422-21.895 12.926L-21.895 12.984Q-21.953 13.199-22.145 13.223M-24.578 12.672L-22.551 12.672Q-22.598 12.262-22.836 12.010Q-23.075 11.758-23.473 11.758Q-23.868 11.758-24.174 12.020Q-24.481 12.281-24.578 12.672M-20.825 14.512L-20.825 13.598Q-20.797 13.391-20.586 13.367L-20.418 13.367Q-20.254 13.391-20.196 13.551Q-19.993 14.191-19.266 14.191Q-19.059 14.191-18.830 14.156Q-18.602 14.121-18.434 14.006Q-18.266 13.891-18.266 13.688Q-18.266 13.477-18.489 13.363Q-18.711 13.250-18.985 13.207L-19.684 13.094Q-20.825 12.883-20.825 12.160Q-20.825 11.871-20.680 11.682Q-20.536 11.492-20.295 11.385Q-20.055 11.277-19.799 11.238Q-19.543 11.199-19.266 11.199Q-19.016 11.199-18.823 11.229Q-18.629 11.258-18.465 11.336Q-18.387 11.219-18.258 11.199L-18.180 11.199Q-18.082 11.211-18.020 11.274Q-17.957 11.336-17.946 11.430L-17.946 12.137Q-17.957 12.231-18.020 12.297Q-18.082 12.363-18.180 12.375L-18.348 12.375Q-18.442 12.363-18.508 12.297Q-18.575 12.231-18.586 12.137Q-18.586 11.758-19.282 11.758Q-19.629 11.758-19.948 11.840Q-20.266 11.922-20.266 12.168Q-20.266 12.434-19.594 12.543L-18.891 12.664Q-18.407 12.746-18.057 12.994Q-17.707 13.242-17.707 13.688Q-17.707 14.078-17.944 14.320Q-18.180 14.563-18.530 14.656Q-18.879 14.750-19.266 14.750Q-19.844 14.750-20.243 14.496Q-20.313 14.621-20.362 14.678Q-20.411 14.734-20.516 14.750L-20.586 14.750Q-20.801 14.727-20.825 14.512\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M133.966-22.278h22.607\"\u002F>\u003Cpath stroke=\"none\" d=\"m159.173-22.278-4.16-2.08 1.56 2.08-1.56 2.08\"\u002F>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"m199.244-9.274 34.847 26.48\"\u002F>\u003Cpath stroke=\"none\" d=\"m236.161 18.78-2.054-4.174-.016 2.6-2.5.712\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(261.116 -24.052)\">\u003Cpath d=\"M-41.633 14.488L-42.098 11.824L-42.258 11.824Q-42.465 11.801-42.504 11.582L-42.504 11.496Q-42.465 11.285-42.258 11.262L-41.090 11.262Q-40.879 11.285-40.840 11.496L-40.840 11.582Q-40.879 11.801-41.090 11.824L-41.563 11.824Q-41.450 12.469-41.371 12.914Q-41.293 13.359-41.243 13.678Q-41.192 13.996-41.192 14.070Q-41.184 13.899-40.899 12.902Q-40.860 12.789-40.762 12.715Q-40.664 12.641-40.543 12.641L-40.465 12.641Q-40.344 12.641-40.246 12.715Q-40.149 12.789-40.114 12.902Q-40.004 13.285-39.922 13.609Q-39.840 13.934-39.840 14.070Q-39.828 13.781-39.481 11.824L-39.953 11.824Q-40.161 11.801-40.200 11.582L-40.200 11.496Q-40.161 11.285-39.953 11.262L-38.778 11.262Q-38.586 11.285-38.528 11.496L-38.528 11.582Q-38.582 11.801-38.778 11.824L-38.946 11.824L-39.411 14.488Q-39.434 14.606-39.522 14.678Q-39.610 14.750-39.731 14.750L-39.871 14.750Q-39.993 14.750-40.088 14.678Q-40.184 14.606-40.227 14.488Q-40.274 14.309-40.346 14.053Q-40.418 13.797-40.461 13.588Q-40.504 13.379-40.504 13.277Q-40.512 13.559-40.793 14.488Q-40.821 14.598-40.918 14.674Q-41.016 14.750-41.137 14.750L-41.313 14.750Q-41.434 14.750-41.522 14.678Q-41.610 14.606-41.633 14.488M-34.883 13.223L-37.325 13.223Q-37.270 13.500-37.073 13.723Q-36.875 13.945-36.598 14.068Q-36.321 14.191-36.036 14.191Q-35.563 14.191-35.340 13.902Q-35.332 13.891-35.276 13.785Q-35.219 13.680-35.170 13.637Q-35.121 13.594-35.028 13.582L-34.883 13.582Q-34.692 13.602-34.633 13.816L-34.633 13.871Q-34.700 14.172-34.930 14.369Q-35.161 14.566-35.473 14.658Q-35.786 14.750-36.090 14.750Q-36.575 14.750-37.014 14.522Q-37.453 14.293-37.721 13.893Q-37.989 13.492-37.989 13L-37.989 12.941Q-37.989 12.473-37.743 12.070Q-37.496 11.668-37.088 11.434Q-36.680 11.199-36.211 11.199Q-35.707 11.199-35.354 11.422Q-35 11.645-34.817 12.033Q-34.633 12.422-34.633 12.926L-34.633 12.984Q-34.692 13.199-34.883 13.223M-37.317 12.672L-35.289 12.672Q-35.336 12.262-35.575 12.010Q-35.813 11.758-36.211 11.758Q-36.606 11.758-36.912 12.020Q-37.219 12.281-37.317 12.672M-33.532 14.473L-33.532 14.383Q-33.481 14.176-33.286 14.152L-32.246 14.152L-32.246 11.824L-33.219 11.824Q-33.418 11.801-33.469 11.582L-33.469 11.496Q-33.418 11.285-33.219 11.262L-31.852 11.262Q-31.657 11.281-31.606 11.496L-31.606 14.152L-30.692 14.152Q-30.496 14.176-30.446 14.383L-30.446 14.473Q-30.496 14.688-30.692 14.711L-33.286 14.711Q-33.481 14.688-33.532 14.473M-32.500 10.285L-32.500 10.231Q-32.500 10.059-32.364 9.938Q-32.227 9.816-32.051 9.816Q-31.879 9.816-31.743 9.938Q-31.606 10.059-31.606 10.231L-31.606 10.285Q-31.606 10.461-31.743 10.582Q-31.879 10.703-32.051 10.703Q-32.227 10.703-32.364 10.582Q-32.500 10.461-32.500 10.285M-29.664 15.359Q-29.664 15.059-29.516 14.797Q-29.368 14.535-29.118 14.375Q-29.301 14.121-29.301 13.801Q-29.301 13.516-29.149 13.254Q-29.399 12.930-29.399 12.512Q-29.399 12.152-29.207 11.856Q-29.016 11.559-28.698 11.391Q-28.379 11.223-28.016 11.223Q-27.809 11.223-27.606 11.283Q-27.403 11.344-27.239 11.445Q-26.856 11.184-26.383 11.184Q-26.145 11.184-25.963 11.315Q-25.782 11.445-25.782 11.672Q-25.782 11.820-25.883 11.926Q-25.985 12.031-26.141 12.031Q-26.274 12.031-26.370 11.953Q-26.465 11.875-26.496 11.750Q-26.641 11.758-26.840 11.848Q-26.637 12.160-26.637 12.512Q-26.637 12.793-26.748 13.025Q-26.860 13.258-27.055 13.436Q-27.250 13.613-27.502 13.711Q-27.754 13.809-28.016 13.809Q-28.403 13.809-28.735 13.621Q-28.746 13.621-28.756 13.693Q-28.766 13.766-28.766 13.801Q-28.766 13.922-28.700 14.029Q-28.633 14.137-28.512 14.176Q-28.496 14.172-28.483 14.170Q-28.469 14.168-28.446 14.168Q-28.430 14.168-28.399 14.176Q-28.368 14.184-28.360 14.184L-27.766 14.184Q-26.985 14.184-26.444 14.426Q-25.903 14.668-25.903 15.359Q-25.903 15.660-26.082 15.887Q-26.262 16.113-26.559 16.258Q-26.856 16.402-27.176 16.469Q-27.496 16.535-27.782 16.535Q-28.172 16.535-28.614 16.412Q-29.055 16.289-29.360 16.022Q-29.664 15.754-29.664 15.359M-29.125 15.352Q-29.125 15.570-28.885 15.711Q-28.645 15.852-28.321 15.918Q-27.996 15.984-27.782 15.984Q-27.567 15.984-27.243 15.918Q-26.918 15.852-26.678 15.711Q-26.438 15.570-26.438 15.352Q-26.438 15.063-26.662 14.924Q-26.887 14.785-27.168 14.752Q-27.450 14.719-27.797 14.719L-28.414 14.719Q-28.598 14.719-28.760 14.799Q-28.922 14.879-29.024 15.025Q-29.125 15.172-29.125 15.352M-28.016 13.254Q-27.715 13.254-27.496 13.035Q-27.278 12.816-27.278 12.512Q-27.278 12.356-27.332 12.225Q-27.387 12.094-27.493 11.988Q-27.598 11.883-27.729 11.828Q-27.860 11.774-28.016 11.774Q-28.321 11.774-28.539 11.992Q-28.758 12.211-28.758 12.512Q-28.758 12.809-28.536 13.031Q-28.313 13.254-28.016 13.254M-25.586 14.473L-25.586 14.383Q-25.543 14.176-25.336 14.152L-24.914 14.152L-24.914 10.383L-25.336 10.383Q-25.543 10.359-25.586 10.145L-25.586 10.055Q-25.543 9.848-25.336 9.824L-24.520 9.824Q-24.325 9.848-24.274 10.055L-24.274 11.606Q-23.813 11.223-23.215 11.223Q-22.868 11.223-22.629 11.363Q-22.391 11.504-22.276 11.762Q-22.161 12.020-22.161 12.375L-22.161 14.152L-21.735 14.152Q-21.528 14.176-21.489 14.383L-21.489 14.473Q-21.528 14.688-21.735 14.711L-23.129 14.711Q-23.325 14.688-23.375 14.473L-23.375 14.383Q-23.325 14.172-23.129 14.152L-22.801 14.152L-22.801 12.406Q-22.801 12.098-22.891 11.940Q-22.981 11.781-23.274 11.781Q-23.543 11.781-23.772 11.912Q-24 12.043-24.137 12.272Q-24.274 12.500-24.274 12.766L-24.274 14.152L-23.848 14.152Q-23.641 14.176-23.602 14.383L-23.602 14.473Q-23.641 14.688-23.848 14.711L-25.336 14.711Q-25.543 14.688-25.586 14.473M-20.211 13.606L-20.211 11.824L-20.961 11.824Q-21.161 11.801-21.211 11.582L-21.211 11.496Q-21.161 11.285-20.961 11.262L-20.211 11.262L-20.211 10.512Q-20.161 10.305-19.961 10.277L-19.817 10.277Q-19.621 10.305-19.571 10.512L-19.571 11.262L-18.211 11.262Q-18.020 11.281-17.961 11.496L-17.961 11.582Q-18.016 11.801-18.211 11.824L-19.571 11.824L-19.571 13.574Q-19.571 14.191-18.996 14.191Q-18.746 14.191-18.582 14.006Q-18.418 13.820-18.418 13.574Q-18.418 13.481-18.346 13.410Q-18.274 13.340-18.172 13.328L-18.028 13.328Q-17.828 13.352-17.778 13.559L-17.778 13.606Q-17.778 13.930-17.961 14.193Q-18.145 14.457-18.438 14.604Q-18.731 14.750-19.051 14.750Q-19.563 14.750-19.887 14.440Q-20.211 14.129-20.211 13.606M-16.578 14.512L-16.578 13.598Q-16.551 13.391-16.340 13.367L-16.172 13.367Q-16.008 13.391-15.950 13.551Q-15.746 14.191-15.020 14.191Q-14.813 14.191-14.584 14.156Q-14.356 14.121-14.188 14.006Q-14.020 13.891-14.020 13.688Q-14.020 13.477-14.243 13.363Q-14.465 13.250-14.739 13.207L-15.438 13.094Q-16.578 12.883-16.578 12.160Q-16.578 11.871-16.434 11.682Q-16.289 11.492-16.049 11.385Q-15.809 11.277-15.553 11.238Q-15.297 11.199-15.020 11.199Q-14.770 11.199-14.577 11.229Q-14.383 11.258-14.219 11.336Q-14.141 11.219-14.012 11.199L-13.934 11.199Q-13.836 11.211-13.774 11.274Q-13.711 11.336-13.700 11.430L-13.700 12.137Q-13.711 12.231-13.774 12.297Q-13.836 12.363-13.934 12.375L-14.102 12.375Q-14.196 12.363-14.262 12.297Q-14.328 12.231-14.340 12.137Q-14.340 11.758-15.036 11.758Q-15.383 11.758-15.702 11.840Q-16.020 11.922-16.020 12.168Q-16.020 12.434-15.348 12.543L-14.645 12.664Q-14.161 12.746-13.811 12.994Q-13.461 13.242-13.461 13.688Q-13.461 14.078-13.698 14.320Q-13.934 14.563-14.284 14.656Q-14.633 14.750-15.020 14.750Q-15.598 14.750-15.996 14.496Q-16.067 14.621-16.116 14.678Q-16.164 14.734-16.270 14.750L-16.340 14.750Q-16.555 14.727-16.578 14.512\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"m-19.679 84.288 248.854-47.88\"\u002F>\u003Cpath stroke=\"none\" d=\"m231.728 35.916-4.478-1.256 1.925 1.747-1.14 2.337\"\u002F>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M274.807 31.783h26.875\"\u002F>\u003Cpath stroke=\"none\" d=\"m304.282 31.783-4.16-2.08 1.56 2.08-1.56 2.08\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fsvg>\u003Cfigcaption class=\"tikz-cap\">Scaled dot-product attention as a dataflow: queries and keys give scaled scores, a softmax turns them into weights, and multiplying by \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6833em;\">\u003C\u002Fspan>\u003Cspan class=\"mord mathnormal\" style=\"margin-right:0.2222em;\">V\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan> blends the values.\u003C\u002Ffigcaption>","\u003Csvg style=\"width:100%;max-width:340.611px;height:auto\" xmlns=\"http:\u002F\u002Fwww.w3.org\u002F2000\u002Fsvg\" viewBox=\"-75 -75 255.458 111.090\">\u003Cg stroke=\"currentColor\" style=\"stroke-miterlimit:10;stroke-width:.4\">\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M-59.494 26.376H134.23\"\u002F>\u003Cpath stroke=\"none\" d=\"m136.83 26.376-4.16-2.08 1.56 2.08-1.56 2.08\"\u002F>\u003Cg stroke=\"none\" font-family=\"cmr7\" font-size=\"7\">\u003Cg transform=\"translate(191.521 1.75)\">\u003Cpath d=\"M-49.047 26.376L-50.630 26.376L-50.630 26.096Q-50.401 26.096-50.252 26.062Q-50.104 26.027-50.104 25.887L-50.104 22.268Q-50.104 21.998-50.211 21.936Q-50.319 21.875-50.630 21.875L-50.630 21.594L-49.550 21.519L-49.550 24.807L-48.565 24.038Q-48.360 23.901-48.360 23.751Q-48.360 23.707-48.401 23.672Q-48.442 23.638-48.487 23.638L-48.487 23.358L-47.123 23.358L-47.123 23.638Q-47.612 23.638-48.131 24.038L-48.688 24.472L-47.711 25.696Q-47.509 25.942-47.376 26.019Q-47.243 26.096-46.956 26.096L-46.956 26.376L-48.388 26.376L-48.388 26.096Q-48.200 26.096-48.200 25.983Q-48.200 25.887-48.354 25.696L-49.088 24.787L-49.570 25.166L-49.570 25.887Q-49.570 26.024-49.422 26.060Q-49.273 26.096-49.047 26.096\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(191.521 1.75)\">\u003Cpath d=\"M-46.699 24.841Q-46.699 24.520-46.574 24.231Q-46.449 23.942-46.223 23.719Q-45.998 23.495-45.702 23.375Q-45.407 23.255-45.089 23.255Q-44.761 23.255-44.499 23.355Q-44.238 23.454-44.062 23.636Q-43.886 23.819-43.792 24.077Q-43.698 24.335-43.698 24.667Q-43.698 24.759-43.780 24.780L-46.035 24.780L-46.035 24.841Q-46.035 25.429-45.752 25.812Q-45.468 26.195-44.901 26.195Q-44.579 26.195-44.311 26.002Q-44.043 25.809-43.954 25.494Q-43.947 25.453-43.872 25.439L-43.780 25.439Q-43.698 25.463-43.698 25.535Q-43.698 25.542-43.704 25.569Q-43.817 25.966-44.188 26.205Q-44.559 26.444-44.983 26.444Q-45.420 26.444-45.820 26.236Q-46.220 26.027-46.459 25.660Q-46.699 25.293-46.699 24.841M-46.029 24.571L-44.214 24.571Q-44.214 24.294-44.311 24.042Q-44.409 23.789-44.607 23.633Q-44.805 23.478-45.089 23.478Q-45.366 23.478-45.579 23.636Q-45.793 23.795-45.911 24.050Q-46.029 24.305-46.029 24.571M-42.775 27.511Q-42.645 27.579-42.508 27.579Q-42.337 27.579-42.187 27.490Q-42.036 27.401-41.925 27.256Q-41.814 27.111-41.736 26.943L-41.472 26.376L-42.641 23.850Q-42.717 23.703-42.847 23.671Q-42.976 23.638-43.209 23.638L-43.209 23.358L-41.688 23.358L-41.688 23.638Q-42.036 23.638-42.036 23.785Q-42.033 23.806-42.031 23.823Q-42.030 23.840-42.030 23.850L-41.172 25.709L-40.399 24.038Q-40.365 23.970-40.365 23.891Q-40.365 23.778-40.449 23.708Q-40.533 23.638-40.645 23.638L-40.645 23.358L-39.449 23.358L-39.449 23.638Q-39.668 23.638-39.840 23.742Q-40.013 23.847-40.105 24.038L-41.442 26.943Q-41.613 27.313-41.883 27.559Q-42.153 27.805-42.508 27.805Q-42.778 27.805-42.997 27.639Q-43.216 27.473-43.216 27.210Q-43.216 27.073-43.123 26.984Q-43.031 26.896-42.891 26.896Q-42.754 26.896-42.665 26.984Q-42.576 27.073-42.576 27.210Q-42.576 27.313-42.629 27.391Q-42.682 27.470-42.775 27.511\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(191.521 1.75)\">\u003Cpath d=\"M-34.589 26.376L-36.141 26.376L-36.141 26.096Q-35.915 26.096-35.766 26.062Q-35.618 26.027-35.618 25.887L-35.618 24.038Q-35.618 23.850-35.666 23.766Q-35.713 23.683-35.811 23.664Q-35.908 23.645-36.120 23.645L-36.120 23.365L-35.064 23.290L-35.064 25.887Q-35.064 26.027-34.932 26.062Q-34.801 26.096-34.589 26.096L-34.589 26.376M-35.860 22.069Q-35.860 21.898-35.737 21.779Q-35.614 21.659-35.443 21.659Q-35.276 21.659-35.153 21.779Q-35.030 21.898-35.030 22.069Q-35.030 22.244-35.153 22.367Q-35.276 22.490-35.443 22.490Q-35.614 22.490-35.737 22.367Q-35.860 22.244-35.860 22.069M-32.261 26.376L-33.895 26.376L-33.895 26.096Q-33.666 26.096-33.517 26.062Q-33.369 26.027-33.369 25.887L-33.369 24.038Q-33.369 23.768-33.476 23.707Q-33.584 23.645-33.895 23.645L-33.895 23.365L-32.835 23.290L-32.835 23.939Q-32.665 23.631-32.360 23.460Q-32.056 23.290-31.711 23.290Q-31.205 23.290-30.921 23.513Q-30.638 23.737-30.638 24.233L-30.638 25.887Q-30.638 26.024-30.489 26.060Q-30.340 26.096-30.115 26.096L-30.115 26.376L-31.745 26.376L-31.745 26.096Q-31.516 26.096-31.367 26.062Q-31.219 26.027-31.219 25.887L-31.219 24.247Q-31.219 23.912-31.338 23.712Q-31.458 23.512-31.772 23.512Q-32.042 23.512-32.277 23.648Q-32.511 23.785-32.649 24.019Q-32.788 24.253-32.788 24.527L-32.788 25.887Q-32.788 26.024-32.637 26.060Q-32.487 26.096-32.261 26.096L-32.261 26.376M-29.527 24.865Q-29.527 24.527-29.387 24.236Q-29.247 23.946-29.002 23.732Q-28.758 23.519-28.454 23.404Q-28.149 23.290-27.825 23.290Q-27.555 23.290-27.291 23.389Q-27.028 23.488-26.837 23.666L-26.837 22.268Q-26.837 21.998-26.945 21.936Q-27.052 21.875-27.363 21.875L-27.363 21.594L-26.287 21.519L-26.287 25.703Q-26.287 25.891-26.232 25.974Q-26.177 26.058-26.076 26.077Q-25.976 26.096-25.760 26.096L-25.760 26.376L-26.868 26.444L-26.868 26.027Q-27.285 26.444-27.910 26.444Q-28.341 26.444-28.713 26.232Q-29.086 26.021-29.306 25.660Q-29.527 25.299-29.527 24.865M-27.852 26.222Q-27.644 26.222-27.457 26.150Q-27.271 26.079-27.117 25.942Q-26.963 25.805-26.868 25.627L-26.868 24.018Q-26.953 23.871-27.098 23.751Q-27.244 23.631-27.413 23.572Q-27.582 23.512-27.763 23.512Q-28.324 23.512-28.592 23.901Q-28.860 24.291-28.860 24.872Q-28.860 25.443-28.626 25.833Q-28.392 26.222-27.852 26.222M-25.152 24.841Q-25.152 24.520-25.027 24.231Q-24.902 23.942-24.677 23.719Q-24.451 23.495-24.155 23.375Q-23.860 23.255-23.542 23.255Q-23.214 23.255-22.952 23.355Q-22.691 23.454-22.515 23.636Q-22.339 23.819-22.245 24.077Q-22.151 24.335-22.151 24.667Q-22.151 24.759-22.233 24.780L-24.489 24.780L-24.489 24.841Q-24.489 25.429-24.205 25.812Q-23.921 26.195-23.354 26.195Q-23.033 26.195-22.764 26.002Q-22.496 25.809-22.407 25.494Q-22.400 25.453-22.325 25.439L-22.233 25.439Q-22.151 25.463-22.151 25.535Q-22.151 25.542-22.158 25.569Q-22.270 25.966-22.641 26.205Q-23.012 26.444-23.436 26.444Q-23.874 26.444-24.273 26.236Q-24.673 26.027-24.913 25.660Q-25.152 25.293-25.152 24.841M-24.482 24.571L-22.667 24.571Q-22.667 24.294-22.764 24.042Q-22.862 23.789-23.060 23.633Q-23.258 23.478-23.542 23.478Q-23.819 23.478-24.032 23.636Q-24.246 23.795-24.364 24.050Q-24.482 24.305-24.482 24.571M-20.380 26.376L-21.703 26.376L-21.703 26.096Q-21.143 26.096-20.763 25.696L-20.049 24.899L-20.961 23.850Q-21.098 23.703-21.247 23.671Q-21.395 23.638-21.662 23.638L-21.662 23.358L-20.162 23.358L-20.162 23.638Q-20.353 23.638-20.353 23.772Q-20.353 23.802-20.322 23.850L-19.728 24.534L-19.287 24.038Q-19.174 23.908-19.174 23.792Q-19.174 23.730-19.211 23.684Q-19.249 23.638-19.307 23.638L-19.307 23.358L-17.991 23.358L-17.991 23.638Q-18.552 23.638-18.931 24.038L-19.553 24.739L-18.559 25.887Q-18.459 25.986-18.359 26.031Q-18.258 26.075-18.147 26.085Q-18.036 26.096-17.858 26.096L-17.858 26.376L-19.352 26.376L-19.352 26.096Q-19.287 26.096-19.227 26.062Q-19.167 26.027-19.167 25.962Q-19.167 25.915-19.198 25.887L-19.874 25.101L-20.408 25.696Q-20.520 25.826-20.520 25.942Q-20.520 26.007-20.479 26.051Q-20.438 26.096-20.380 26.096\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M-50.958 32.067v-88.449\"\u002F>\u003Cpath stroke=\"none\" d=\"m-50.958-58.982-2.08 4.16 2.08-1.56 2.08 1.56\"\u002F>\u003Cg stroke=\"none\" font-family=\"cmr7\" font-size=\"7\">\u003Cg transform=\"translate(-11.313 -90.452)\">\u003Cpath d=\"M-49.256 26.349L-50.237 23.850Q-50.298 23.707-50.416 23.672Q-50.534 23.638-50.750 23.638L-50.750 23.358L-49.270 23.358L-49.270 23.638Q-49.649 23.638-49.649 23.799Q-49.649 23.809-49.635 23.850L-48.921 25.682L-48.248 23.977Q-48.278 23.905-48.278 23.877Q-48.278 23.850-48.306 23.850Q-48.367 23.703-48.485 23.671Q-48.603 23.638-48.815 23.638L-48.815 23.358L-47.417 23.358L-47.417 23.638Q-47.793 23.638-47.793 23.799Q-47.793 23.830-47.786 23.850L-47.031 25.788L-46.344 24.038Q-46.323 23.987-46.323 23.932Q-46.323 23.792-46.436 23.715Q-46.549 23.638-46.689 23.638L-46.689 23.358L-45.469 23.358L-45.469 23.638Q-45.674 23.638-45.829 23.744Q-45.985 23.850-46.057 24.038L-46.962 26.349Q-46.997 26.444-47.109 26.444L-47.178 26.444Q-47.287 26.444-47.325 26.349L-48.107 24.346L-48.894 26.349Q-48.928 26.444-49.041 26.444L-49.109 26.444Q-49.218 26.444-49.256 26.349\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-11.313 -90.452)\">\u003Cpath d=\"M-45.192 24.841Q-45.192 24.520-45.067 24.231Q-44.942 23.942-44.716 23.719Q-44.491 23.495-44.195 23.375Q-43.900 23.255-43.582 23.255Q-43.254 23.255-42.992 23.355Q-42.731 23.454-42.555 23.636Q-42.379 23.819-42.285 24.077Q-42.191 24.335-42.191 24.667Q-42.191 24.759-42.273 24.780L-44.528 24.780L-44.528 24.841Q-44.528 25.429-44.245 25.812Q-43.961 26.195-43.394 26.195Q-43.072 26.195-42.804 26.002Q-42.536 25.809-42.447 25.494Q-42.440 25.453-42.365 25.439L-42.273 25.439Q-42.191 25.463-42.191 25.535Q-42.191 25.542-42.197 25.569Q-42.310 25.966-42.681 26.205Q-43.052 26.444-43.476 26.444Q-43.913 26.444-44.313 26.236Q-44.713 26.027-44.952 25.660Q-45.192 25.293-45.192 24.841M-44.522 24.571L-42.707 24.571Q-42.707 24.294-42.804 24.042Q-42.902 23.789-43.100 23.633Q-43.298 23.478-43.582 23.478Q-43.859 23.478-44.072 23.636Q-44.286 23.795-44.404 24.050Q-44.522 24.305-44.522 24.571M-39.986 26.376L-41.538 26.376L-41.538 26.096Q-41.312 26.096-41.163 26.062Q-41.015 26.027-41.015 25.887L-41.015 24.038Q-41.015 23.850-41.063 23.766Q-41.111 23.683-41.208 23.664Q-41.305 23.645-41.517 23.645L-41.517 23.365L-40.461 23.290L-40.461 25.887Q-40.461 26.027-40.330 26.062Q-40.198 26.096-39.986 26.096L-39.986 26.376M-41.257 22.069Q-41.257 21.898-41.134 21.779Q-41.011 21.659-40.840 21.659Q-40.673 21.659-40.550 21.779Q-40.427 21.898-40.427 22.069Q-40.427 22.244-40.550 22.367Q-40.673 22.490-40.840 22.490Q-41.011 22.490-41.134 22.367Q-41.257 22.244-41.257 22.069M-39.381 26.909Q-39.381 26.663-39.184 26.479Q-38.988 26.294-38.732 26.215Q-38.868 26.103-38.940 25.942Q-39.012 25.781-39.012 25.600Q-39.012 25.279-38.800 25.033Q-39.135 24.735-39.135 24.325Q-39.135 23.864-38.745 23.577Q-38.356 23.290-37.877 23.290Q-37.405 23.290-37.070 23.536Q-36.896 23.382-36.686 23.300Q-36.476 23.218-36.247 23.218Q-36.083 23.218-35.961 23.325Q-35.840 23.433-35.840 23.597Q-35.840 23.693-35.912 23.765Q-35.984 23.836-36.076 23.836Q-36.175 23.836-36.245 23.763Q-36.315 23.689-36.315 23.590Q-36.315 23.536-36.301 23.505L-36.295 23.491Q-36.288 23.471-36.279 23.460Q-36.271 23.450-36.267 23.443Q-36.623 23.443-36.910 23.666Q-36.623 23.959-36.623 24.325Q-36.623 24.640-36.807 24.872Q-36.992 25.105-37.281 25.233Q-37.569 25.361-37.877 25.361Q-38.079 25.361-38.270 25.311Q-38.462 25.262-38.639 25.152Q-38.732 25.279-38.732 25.422Q-38.732 25.604-38.603 25.739Q-38.475 25.874-38.291 25.874L-37.658 25.874Q-37.211 25.874-36.841 25.945Q-36.472 26.017-36.213 26.246Q-35.953 26.475-35.953 26.909Q-35.953 27.230-36.248 27.432Q-36.544 27.634-36.947 27.723Q-37.351 27.812-37.665 27.812Q-37.983 27.812-38.386 27.723Q-38.790 27.634-39.085 27.432Q-39.381 27.230-39.381 26.909M-38.926 26.909Q-38.926 27.138-38.708 27.287Q-38.489 27.436-38.197 27.504Q-37.904 27.572-37.665 27.572Q-37.501 27.572-37.293 27.536Q-37.084 27.501-36.877 27.420Q-36.671 27.340-36.539 27.212Q-36.407 27.084-36.407 26.909Q-36.407 26.557-36.788 26.463Q-37.170 26.369-37.672 26.369L-38.291 26.369Q-38.530 26.369-38.728 26.520Q-38.926 26.670-38.926 26.909M-37.877 25.122Q-37.211 25.122-37.211 24.325Q-37.211 23.525-37.877 23.525Q-38.547 23.525-38.547 24.325Q-38.547 25.122-37.877 25.122M-33.676 26.376L-35.310 26.376L-35.310 26.096Q-35.081 26.096-34.933 26.062Q-34.784 26.027-34.784 25.887L-34.784 22.268Q-34.784 21.998-34.892 21.936Q-34.999 21.875-35.310 21.875L-35.310 21.594L-34.230 21.519L-34.230 23.905Q-34.124 23.720-33.946 23.578Q-33.769 23.437-33.560 23.363Q-33.352 23.290-33.126 23.290Q-32.620 23.290-32.337 23.513Q-32.053 23.737-32.053 24.233L-32.053 25.887Q-32.053 26.024-31.904 26.060Q-31.756 26.096-31.530 26.096L-31.530 26.376L-33.160 26.376L-33.160 26.096Q-32.931 26.096-32.783 26.062Q-32.634 26.027-32.634 25.887L-32.634 24.247Q-32.634 23.912-32.754 23.712Q-32.873 23.512-33.188 23.512Q-33.458 23.512-33.692 23.648Q-33.926 23.785-34.064 24.019Q-34.203 24.253-34.203 24.527L-34.203 25.887Q-34.203 26.024-34.052 26.060Q-33.902 26.096-33.676 26.096\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-11.313 -90.452)\">\u003Cpath d=\"M-30.617 25.535L-30.617 23.638L-31.256 23.638L-31.256 23.416Q-30.938 23.416-30.721 23.206Q-30.504 22.996-30.404 22.686Q-30.303 22.377-30.303 22.069L-30.036 22.069L-30.036 23.358L-28.959 23.358L-28.959 23.638L-30.036 23.638L-30.036 25.522Q-30.036 25.798-29.932 25.997Q-29.828 26.195-29.568 26.195Q-29.411 26.195-29.305 26.091Q-29.199 25.986-29.149 25.833Q-29.100 25.679-29.100 25.522L-29.100 25.108L-28.833 25.108L-28.833 25.535Q-28.833 25.761-28.932 25.971Q-29.031 26.181-29.216 26.313Q-29.400 26.444-29.629 26.444Q-30.067 26.444-30.342 26.207Q-30.617 25.969-30.617 25.535\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003Cpath fill=\"var(--tk-soft-neutral)\" d=\"M-37.87 26.376V22.11h3.984v4.267ZM-15.108 26.376v-5.69h3.984v5.69ZM7.655 26.376v-72.554h3.983v72.554ZM30.417 26.376v-7.113H34.4v7.113ZM53.18 26.376V22.11h3.983v4.267ZM75.942 26.376v-5.12h3.983v5.12ZM98.703 26.376v-3.414h3.984v3.414Zm3.984-3.414\" style=\"stroke-width:.8\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" stroke=\"var(--tk-accent)\" d=\"M-33.317 26.376V10.727h3.984v15.65ZM-10.555 26.376V5.036h3.983v21.34ZM12.207 26.376v-38.411h3.984v38.411ZM34.97 26.376V-2.076h3.983v28.452ZM57.732 26.376V6.46h3.983v19.916ZM80.494 26.376V12.15h3.984v14.226ZM103.256 26.376v-9.958h3.983v9.958Zm3.983-9.958\" style=\"stroke-width:.8\"\u002F>\u003Cg stroke=\"none\" font-family=\"cmr7\" font-size=\"7\">\u003Cg transform=\"translate(86.046 -66.536)\">\u003Cpath d=\"M-50.069 25.542L-50.069 24.038Q-50.069 23.768-50.177 23.707Q-50.285 23.645-50.596 23.645L-50.596 23.365L-49.488 23.290L-49.488 25.522L-49.488 25.542Q-49.488 25.822-49.437 25.966Q-49.386 26.109-49.244 26.166Q-49.102 26.222-48.815 26.222Q-48.562 26.222-48.357 26.082Q-48.152 25.942-48.036 25.716Q-47.919 25.491-47.919 25.241L-47.919 24.038Q-47.919 23.768-48.027 23.707Q-48.135 23.645-48.446 23.645L-48.446 23.365L-47.338 23.290L-47.338 25.703Q-47.338 25.894-47.285 25.976Q-47.232 26.058-47.132 26.077Q-47.031 26.096-46.815 26.096L-46.815 26.376L-47.892 26.444L-47.892 25.880Q-48.001 26.062-48.147 26.185Q-48.292 26.308-48.478 26.376Q-48.665 26.444-48.866 26.444Q-50.069 26.444-50.069 25.542M-44.546 26.376L-46.180 26.376L-46.180 26.096Q-45.951 26.096-45.802 26.062Q-45.653 26.027-45.653 25.887L-45.653 24.038Q-45.653 23.768-45.761 23.707Q-45.869 23.645-46.180 23.645L-46.180 23.365L-45.120 23.290L-45.120 23.939Q-44.949 23.631-44.645 23.460Q-44.341 23.290-43.996 23.290Q-43.490 23.290-43.206 23.513Q-42.922 23.737-42.922 24.233L-42.922 25.887Q-42.922 26.024-42.774 26.060Q-42.625 26.096-42.399 26.096L-42.399 26.376L-44.030 26.376L-44.030 26.096Q-43.801 26.096-43.652 26.062Q-43.503 26.027-43.503 25.887L-43.503 24.247Q-43.503 23.912-43.623 23.712Q-43.743 23.512-44.057 23.512Q-44.327 23.512-44.561 23.648Q-44.795 23.785-44.934 24.019Q-45.072 24.253-45.072 24.527L-45.072 25.887Q-45.072 26.024-44.922 26.060Q-44.771 26.096-44.546 26.096L-44.546 26.376M-41.812 26.369L-41.812 25.306Q-41.812 25.282-41.784 25.255Q-41.757 25.228-41.733 25.228L-41.624 25.228Q-41.559 25.228-41.545 25.286Q-41.449 25.720-41.203 25.971Q-40.957 26.222-40.543 26.222Q-40.202 26.222-39.949 26.089Q-39.696 25.956-39.696 25.648Q-39.696 25.491-39.790 25.376Q-39.884 25.262-40.022 25.193Q-40.161 25.125-40.328 25.087L-40.909 24.988Q-41.265 24.920-41.538 24.699Q-41.812 24.479-41.812 24.137Q-41.812 23.888-41.700 23.713Q-41.589 23.539-41.403 23.440Q-41.217 23.341-41.001 23.298Q-40.786 23.255-40.543 23.255Q-40.130 23.255-39.850 23.437L-39.634 23.262Q-39.624 23.259-39.617 23.257Q-39.610 23.255-39.600 23.255L-39.549 23.255Q-39.521 23.255-39.498 23.279Q-39.474 23.303-39.474 23.331L-39.474 24.178Q-39.474 24.199-39.498 24.226Q-39.521 24.253-39.549 24.253L-39.662 24.253Q-39.689 24.253-39.715 24.228Q-39.740 24.202-39.740 24.178Q-39.740 23.942-39.846 23.778Q-39.952 23.614-40.135 23.532Q-40.318 23.450-40.550 23.450Q-40.878 23.450-41.135 23.553Q-41.391 23.655-41.391 23.932Q-41.391 24.127-41.208 24.236Q-41.025 24.346-40.796 24.387L-40.222 24.493Q-39.976 24.541-39.762 24.669Q-39.549 24.797-39.412 25Q-39.275 25.204-39.275 25.453Q-39.275 25.966-39.641 26.205Q-40.007 26.444-40.543 26.444Q-41.039 26.444-41.371 26.150L-41.637 26.424Q-41.658 26.444-41.685 26.444L-41.733 26.444Q-41.757 26.444-41.784 26.417Q-41.812 26.390-41.812 26.369M-38.646 24.865Q-38.646 24.537-38.511 24.236Q-38.376 23.936-38.141 23.715Q-37.905 23.495-37.601 23.375Q-37.296 23.255-36.972 23.255Q-36.466 23.255-36.117 23.358Q-35.769 23.460-35.769 23.836Q-35.769 23.983-35.866 24.084Q-35.963 24.185-36.110 24.185Q-36.264 24.185-36.363 24.086Q-36.462 23.987-36.462 23.836Q-36.462 23.648-36.322 23.556Q-36.524 23.505-36.965 23.505Q-37.320 23.505-37.549 23.701Q-37.778 23.898-37.879 24.207Q-37.980 24.517-37.980 24.865Q-37.980 25.214-37.854 25.520Q-37.727 25.826-37.472 26.010Q-37.218 26.195-36.862 26.195Q-36.640 26.195-36.456 26.111Q-36.271 26.027-36.136 25.872Q-36.001 25.716-35.943 25.508Q-35.929 25.453-35.875 25.453L-35.762 25.453Q-35.731 25.453-35.709 25.477Q-35.687 25.501-35.687 25.535L-35.687 25.556Q-35.772 25.843-35.960 26.041Q-36.148 26.239-36.413 26.342Q-36.678 26.444-36.972 26.444Q-37.402 26.444-37.790 26.238Q-38.178 26.031-38.412 25.668Q-38.646 25.306-38.646 24.865M-35.041 25.648Q-35.041 25.316-34.817 25.089Q-34.593 24.862-34.249 24.734Q-33.906 24.605-33.533 24.553Q-33.161 24.500-32.856 24.500L-32.856 24.247Q-32.856 24.042-32.964 23.862Q-33.072 23.683-33.253 23.580Q-33.434 23.478-33.643 23.478Q-34.049 23.478-34.285 23.570Q-34.196 23.607-34.150 23.691Q-34.104 23.775-34.104 23.877Q-34.104 23.973-34.150 24.052Q-34.196 24.130-34.277 24.175Q-34.357 24.219-34.446 24.219Q-34.596 24.219-34.697 24.122Q-34.798 24.024-34.798 23.877Q-34.798 23.255-33.643 23.255Q-33.431 23.255-33.181 23.319Q-32.932 23.382-32.730 23.501Q-32.528 23.621-32.402 23.806Q-32.275 23.990-32.275 24.233L-32.275 25.809Q-32.275 25.925-32.214 26.021Q-32.152 26.116-32.040 26.116Q-31.930 26.116-31.865 26.022Q-31.800 25.928-31.800 25.809L-31.800 25.361L-31.534 25.361L-31.534 25.809Q-31.534 26.079-31.761 26.244Q-31.988 26.410-32.269 26.410Q-32.477 26.410-32.614 26.256Q-32.750 26.103-32.774 25.887Q-32.921 26.154-33.203 26.299Q-33.485 26.444-33.810 26.444Q-34.087 26.444-34.371 26.369Q-34.654 26.294-34.847 26.115Q-35.041 25.935-35.041 25.648M-34.425 25.648Q-34.425 25.822-34.324 25.952Q-34.224 26.082-34.068 26.152Q-33.913 26.222-33.749 26.222Q-33.530 26.222-33.321 26.125Q-33.113 26.027-32.985 25.846Q-32.856 25.665-32.856 25.439L-32.856 24.711Q-33.181 24.711-33.547 24.802Q-33.913 24.893-34.169 25.105Q-34.425 25.316-34.425 25.648M-29.449 26.376L-31.052 26.376L-31.052 26.096Q-30.826 26.096-30.677 26.062Q-30.529 26.027-30.529 25.887L-30.529 22.268Q-30.529 21.998-30.636 21.936Q-30.744 21.875-31.052 21.875L-31.052 21.594L-29.975 21.519L-29.975 25.887Q-29.975 26.024-29.825 26.060Q-29.674 26.096-29.449 26.096L-29.449 26.376M-28.895 24.841Q-28.895 24.520-28.770 24.231Q-28.645 23.942-28.420 23.719Q-28.194 23.495-27.899 23.375Q-27.603 23.255-27.285 23.255Q-26.957 23.255-26.696 23.355Q-26.434 23.454-26.258 23.636Q-26.082 23.819-25.988 24.077Q-25.894 24.335-25.894 24.667Q-25.894 24.759-25.976 24.780L-28.232 24.780L-28.232 24.841Q-28.232 25.429-27.948 25.812Q-27.665 26.195-27.097 26.195Q-26.776 26.195-26.508 26.002Q-26.239 25.809-26.150 25.494Q-26.144 25.453-26.068 25.439L-25.976 25.439Q-25.894 25.463-25.894 25.535Q-25.894 25.542-25.901 25.569Q-26.014 25.966-26.385 26.205Q-26.755 26.444-27.179 26.444Q-27.617 26.444-28.017 26.236Q-28.416 26.027-28.656 25.660Q-28.895 25.293-28.895 24.841M-28.225 24.571L-26.410 24.571Q-26.410 24.294-26.508 24.042Q-26.605 23.789-26.803 23.633Q-27.001 23.478-27.285 23.478Q-27.562 23.478-27.776 23.636Q-27.989 23.795-28.107 24.050Q-28.225 24.305-28.225 24.571M-25.306 24.865Q-25.306 24.527-25.166 24.236Q-25.026 23.946-24.781 23.732Q-24.537 23.519-24.233 23.404Q-23.929 23.290-23.604 23.290Q-23.334 23.290-23.071 23.389Q-22.808 23.488-22.616 23.666L-22.616 22.268Q-22.616 21.998-22.724 21.936Q-22.832 21.875-23.143 21.875L-23.143 21.594L-22.066 21.519L-22.066 25.703Q-22.066 25.891-22.011 25.974Q-21.957 26.058-21.856 26.077Q-21.755 26.096-21.540 26.096L-21.540 26.376L-22.647 26.444L-22.647 26.027Q-23.064 26.444-23.689 26.444Q-24.120 26.444-24.493 26.232Q-24.865 26.021-25.086 25.660Q-25.306 25.299-25.306 24.865M-23.631 26.222Q-23.423 26.222-23.237 26.150Q-23.050 26.079-22.896 25.942Q-22.743 25.805-22.647 25.627L-22.647 24.018Q-22.732 23.871-22.878 23.751Q-23.023 23.631-23.192 23.572Q-23.361 23.512-23.542 23.512Q-24.103 23.512-24.371 23.901Q-24.640 24.291-24.640 24.872Q-24.640 25.443-24.406 25.833Q-24.171 26.222-23.631 26.222\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(86.046 -66.536)\">\u003Cpath d=\"M-16.040 28.126Q-16.590 27.726-16.961 27.171Q-17.332 26.615-17.513 25.969Q-17.694 25.323-17.694 24.626Q-17.694 24.113-17.594 23.618Q-17.493 23.122-17.288 22.671Q-17.083 22.220-16.770 21.828Q-16.457 21.437-16.040 21.133Q-16.030 21.129-16.023 21.128Q-16.016 21.126-16.006 21.126L-15.938 21.126Q-15.903 21.126-15.881 21.150Q-15.859 21.174-15.859 21.211Q-15.859 21.256-15.886 21.273Q-16.235 21.574-16.488 21.958Q-16.741 22.343-16.893 22.784Q-17.045 23.225-17.117 23.681Q-17.189 24.137-17.189 24.626Q-17.189 25.627-16.879 26.514Q-16.570 27.401-15.886 27.986Q-15.859 28.003-15.859 28.047Q-15.859 28.085-15.881 28.109Q-15.903 28.133-15.938 28.133L-16.006 28.133Q-16.013 28.129-16.021 28.128Q-16.030 28.126-16.040 28.126M-15.049 26.369L-15.049 25.306Q-15.049 25.282-15.022 25.255Q-14.994 25.228-14.970 25.228L-14.861 25.228Q-14.796 25.228-14.782 25.286Q-14.687 25.720-14.441 25.971Q-14.194 26.222-13.781 26.222Q-13.439 26.222-13.186 26.089Q-12.933 25.956-12.933 25.648Q-12.933 25.491-13.027 25.376Q-13.121 25.262-13.260 25.193Q-13.398 25.125-13.566 25.087L-14.147 24.988Q-14.502 24.920-14.776 24.699Q-15.049 24.479-15.049 24.137Q-15.049 23.888-14.938 23.713Q-14.827 23.539-14.640 23.440Q-14.454 23.341-14.239 23.298Q-14.024 23.255-13.781 23.255Q-13.367 23.255-13.087 23.437L-12.872 23.262Q-12.861 23.259-12.855 23.257Q-12.848 23.255-12.838 23.255L-12.786 23.255Q-12.759 23.255-12.735 23.279Q-12.711 23.303-12.711 23.331L-12.711 24.178Q-12.711 24.199-12.735 24.226Q-12.759 24.253-12.786 24.253L-12.899 24.253Q-12.926 24.253-12.952 24.228Q-12.978 24.202-12.978 24.178Q-12.978 23.942-13.084 23.778Q-13.190 23.614-13.372 23.532Q-13.555 23.450-13.788 23.450Q-14.116 23.450-14.372 23.553Q-14.629 23.655-14.629 23.932Q-14.629 24.127-14.446 24.236Q-14.263 24.346-14.034 24.387L-13.460 24.493Q-13.213 24.541-13 24.669Q-12.786 24.797-12.650 25Q-12.513 25.204-12.513 25.453Q-12.513 25.966-12.879 26.205Q-13.244 26.444-13.781 26.444Q-14.276 26.444-14.608 26.150L-14.875 26.424Q-14.895 26.444-14.922 26.444L-14.970 26.444Q-14.994 26.444-15.022 26.417Q-15.049 26.390-15.049 26.369M-11.826 25.648Q-11.826 25.316-11.602 25.089Q-11.378 24.862-11.035 24.734Q-10.691 24.605-10.318 24.553Q-9.946 24.500-9.642 24.500L-9.642 24.247Q-9.642 24.042-9.749 23.862Q-9.857 23.683-10.038 23.580Q-10.219 23.478-10.428 23.478Q-10.835 23.478-11.070 23.570Q-10.982 23.607-10.935 23.691Q-10.889 23.775-10.889 23.877Q-10.889 23.973-10.935 24.052Q-10.982 24.130-11.062 24.175Q-11.142 24.219-11.231 24.219Q-11.381 24.219-11.482 24.122Q-11.583 24.024-11.583 23.877Q-11.583 23.255-10.428 23.255Q-10.216 23.255-9.966 23.319Q-9.717 23.382-9.515 23.501Q-9.314 23.621-9.187 23.806Q-9.061 23.990-9.061 24.233L-9.061 25.809Q-9.061 25.925-8.999 26.021Q-8.938 26.116-8.825 26.116Q-8.715 26.116-8.651 26.022Q-8.586 25.928-8.586 25.809L-8.586 25.361L-8.319 25.361L-8.319 25.809Q-8.319 26.079-8.546 26.244Q-8.774 26.410-9.054 26.410Q-9.262 26.410-9.399 26.256Q-9.536 26.103-9.560 25.887Q-9.707 26.154-9.989 26.299Q-10.271 26.444-10.595 26.444Q-10.872 26.444-11.156 26.369Q-11.440 26.294-11.633 26.115Q-11.826 25.935-11.826 25.648M-11.211 25.648Q-11.211 25.822-11.110 25.952Q-11.009 26.082-10.853 26.152Q-10.698 26.222-10.534 26.222Q-10.315 26.222-10.107 26.125Q-9.898 26.027-9.770 25.846Q-9.642 25.665-9.642 25.439L-9.642 24.711Q-9.966 24.711-10.332 24.802Q-10.698 24.893-10.954 25.105Q-11.211 25.316-11.211 25.648M-7.376 25.535L-7.376 23.638L-8.015 23.638L-8.015 23.416Q-7.697 23.416-7.480 23.206Q-7.263 22.996-7.162 22.686Q-7.061 22.377-7.061 22.069L-6.795 22.069L-6.795 23.358L-5.718 23.358L-5.718 23.638L-6.795 23.638L-6.795 25.522Q-6.795 25.798-6.690 25.997Q-6.586 26.195-6.326 26.195Q-6.169 26.195-6.063 26.091Q-5.957 25.986-5.908 25.833Q-5.858 25.679-5.858 25.522L-5.858 25.108L-5.591 25.108L-5.591 25.535Q-5.591 25.761-5.691 25.971Q-5.790 26.181-5.974 26.313Q-6.159 26.444-6.388 26.444Q-6.825 26.444-7.100 26.207Q-7.376 25.969-7.376 25.535M-4.207 25.542L-4.207 24.038Q-4.207 23.768-4.315 23.707Q-4.422 23.645-4.734 23.645L-4.734 23.365L-3.626 23.290L-3.626 25.522L-3.626 25.542Q-3.626 25.822-3.575 25.966Q-3.524 26.109-3.382 26.166Q-3.240 26.222-2.953 26.222Q-2.700 26.222-2.495 26.082Q-2.290 25.942-2.173 25.716Q-2.057 25.491-2.057 25.241L-2.057 24.038Q-2.057 23.768-2.165 23.707Q-2.273 23.645-2.584 23.645L-2.584 23.365L-1.476 23.290L-1.476 25.703Q-1.476 25.894-1.423 25.976Q-1.370 26.058-1.269 26.077Q-1.169 26.096-0.953 26.096L-0.953 26.376L-2.030 26.444L-2.030 25.880Q-2.139 26.062-2.285 26.185Q-2.430 26.308-2.616 26.376Q-2.802 26.444-3.004 26.444Q-4.207 26.444-4.207 25.542M1.385 26.376L-0.352 26.376L-0.352 26.096Q-0.123 26.096 0.026 26.062Q0.175 26.027 0.175 25.887L0.175 24.038Q0.175 23.768 0.067 23.707Q-0.041 23.645-0.352 23.645L-0.352 23.365L0.677 23.290L0.677 23.997Q0.807 23.689 1.050 23.490Q1.292 23.290 1.610 23.290Q1.829 23.290 2 23.414Q2.171 23.539 2.171 23.751Q2.171 23.888 2.072 23.987Q1.973 24.086 1.839 24.086Q1.703 24.086 1.603 23.987Q1.504 23.888 1.504 23.751Q1.504 23.611 1.603 23.512Q1.313 23.512 1.113 23.708Q0.913 23.905 0.821 24.199Q0.728 24.493 0.728 24.773L0.728 25.887Q0.728 26.096 1.385 26.096L1.385 26.376M2.813 25.648Q2.813 25.316 3.037 25.089Q3.261 24.862 3.605 24.734Q3.948 24.605 4.321 24.553Q4.693 24.500 4.997 24.500L4.997 24.247Q4.997 24.042 4.890 23.862Q4.782 23.683 4.601 23.580Q4.420 23.478 4.211 23.478Q3.805 23.478 3.569 23.570Q3.658 23.607 3.704 23.691Q3.750 23.775 3.750 23.877Q3.750 23.973 3.704 24.052Q3.658 24.130 3.577 24.175Q3.497 24.219 3.408 24.219Q3.258 24.219 3.157 24.122Q3.056 24.024 3.056 23.877Q3.056 23.255 4.211 23.255Q4.423 23.255 4.673 23.319Q4.922 23.382 5.124 23.501Q5.326 23.621 5.452 23.806Q5.578 23.990 5.578 24.233L5.578 25.809Q5.578 25.925 5.640 26.021Q5.702 26.116 5.814 26.116Q5.924 26.116 5.989 26.022Q6.054 25.928 6.054 25.809L6.054 25.361L6.320 25.361L6.320 25.809Q6.320 26.079 6.093 26.244Q5.866 26.410 5.585 26.410Q5.377 26.410 5.240 26.256Q5.103 26.103 5.079 25.887Q4.933 26.154 4.651 26.299Q4.369 26.444 4.044 26.444Q3.767 26.444 3.483 26.369Q3.200 26.294 3.006 26.115Q2.813 25.935 2.813 25.648M3.429 25.648Q3.429 25.822 3.529 25.952Q3.630 26.082 3.786 26.152Q3.941 26.222 4.105 26.222Q4.324 26.222 4.533 26.125Q4.741 26.027 4.869 25.846Q4.997 25.665 4.997 25.439L4.997 24.711Q4.673 24.711 4.307 24.802Q3.941 24.893 3.685 25.105Q3.429 25.316 3.429 25.648M7.264 25.535L7.264 23.638L6.624 23.638L6.624 23.416Q6.942 23.416 7.159 23.206Q7.376 22.996 7.477 22.686Q7.578 22.377 7.578 22.069L7.845 22.069L7.845 23.358L8.921 23.358L8.921 23.638L7.845 23.638L7.845 25.522Q7.845 25.798 7.949 25.997Q8.053 26.195 8.313 26.195Q8.470 26.195 8.576 26.091Q8.682 25.986 8.732 25.833Q8.781 25.679 8.781 25.522L8.781 25.108L9.048 25.108L9.048 25.535Q9.048 25.761 8.949 25.971Q8.849 26.181 8.665 26.313Q8.480 26.444 8.251 26.444Q7.814 26.444 7.539 26.207Q7.264 25.969 7.264 25.535M9.817 24.841Q9.817 24.520 9.942 24.231Q10.066 23.942 10.292 23.719Q10.517 23.495 10.813 23.375Q11.109 23.255 11.427 23.255Q11.755 23.255 12.016 23.355Q12.278 23.454 12.454 23.636Q12.630 23.819 12.724 24.077Q12.818 24.335 12.818 24.667Q12.818 24.759 12.736 24.780L10.480 24.780L10.480 24.841Q10.480 25.429 10.764 25.812Q11.047 26.195 11.615 26.195Q11.936 26.195 12.204 26.002Q12.473 25.809 12.561 25.494Q12.568 25.453 12.643 25.439L12.736 25.439Q12.818 25.463 12.818 25.535Q12.818 25.542 12.811 25.569Q12.698 25.966 12.327 26.205Q11.956 26.444 11.533 26.444Q11.095 26.444 10.695 26.236Q10.295 26.027 10.056 25.660Q9.817 25.293 9.817 24.841M10.487 24.571L12.302 24.571Q12.302 24.294 12.204 24.042Q12.107 23.789 11.909 23.633Q11.710 23.478 11.427 23.478Q11.150 23.478 10.936 23.636Q10.723 23.795 10.605 24.050Q10.487 24.305 10.487 24.571M13.406 24.865Q13.406 24.527 13.546 24.236Q13.686 23.946 13.930 23.732Q14.175 23.519 14.479 23.404Q14.783 23.290 15.108 23.290Q15.378 23.290 15.641 23.389Q15.904 23.488 16.096 23.666L16.096 22.268Q16.096 21.998 15.988 21.936Q15.880 21.875 15.569 21.875L15.569 21.594L16.646 21.519L16.646 25.703Q16.646 25.891 16.701 25.974Q16.755 26.058 16.856 26.077Q16.957 26.096 17.172 26.096L17.172 26.376L16.065 26.444L16.065 26.027Q15.648 26.444 15.022 26.444Q14.592 26.444 14.219 26.232Q13.847 26.021 13.626 25.660Q13.406 25.299 13.406 24.865M15.080 26.222Q15.289 26.222 15.475 26.150Q15.662 26.079 15.815 25.942Q15.969 25.805 16.065 25.627L16.065 24.018Q15.979 23.871 15.834 23.751Q15.689 23.631 15.520 23.572Q15.350 23.512 15.169 23.512Q14.609 23.512 14.340 23.901Q14.072 24.291 14.072 24.872Q14.072 25.443 14.306 25.833Q14.540 26.222 15.080 26.222M18.143 28.133L18.075 28.133Q18.040 28.133 18.018 28.107Q17.996 28.082 17.996 28.047Q17.996 28.003 18.027 27.986Q18.382 27.682 18.632 27.292Q18.881 26.902 19.033 26.470Q19.185 26.038 19.256 25.569Q19.326 25.101 19.326 24.626Q19.326 24.147 19.256 23.681Q19.185 23.214 19.032 22.779Q18.878 22.343 18.627 21.955Q18.375 21.567 18.027 21.273Q17.996 21.256 17.996 21.211Q17.996 21.177 18.018 21.152Q18.040 21.126 18.075 21.126L18.143 21.126Q18.153 21.126 18.162 21.128Q18.170 21.129 18.181 21.133Q18.724 21.533 19.097 22.086Q19.469 22.640 19.650 23.286Q19.831 23.932 19.831 24.626Q19.831 25.327 19.650 25.974Q19.469 26.622 19.095 27.176Q18.721 27.730 18.181 28.126Q18.170 28.126 18.162 28.128Q18.153 28.129 18.143 28.133\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\">\u003Cg transform=\"translate(94.582 -41.671)\">\u003Cpath d=\"M-50.644 26.369L-50.644 25.306Q-50.644 25.282-50.616 25.255Q-50.589 25.228-50.565 25.228L-50.456 25.228Q-50.391 25.228-50.377 25.286Q-50.281 25.720-50.035 25.971Q-49.789 26.222-49.375 26.222Q-49.034 26.222-48.781 26.089Q-48.528 25.956-48.528 25.648Q-48.528 25.491-48.622 25.376Q-48.716 25.262-48.854 25.193Q-48.993 25.125-49.160 25.087L-49.741 24.988Q-50.097 24.920-50.370 24.699Q-50.644 24.479-50.644 24.137Q-50.644 23.888-50.532 23.713Q-50.421 23.539-50.235 23.440Q-50.049 23.341-49.833 23.298Q-49.618 23.255-49.375 23.255Q-48.962 23.255-48.682 23.437L-48.466 23.262Q-48.456 23.259-48.449 23.257Q-48.442 23.255-48.432 23.255L-48.381 23.255Q-48.354 23.255-48.330 23.279Q-48.306 23.303-48.306 23.331L-48.306 24.178Q-48.306 24.199-48.330 24.226Q-48.354 24.253-48.381 24.253L-48.494 24.253Q-48.521 24.253-48.547 24.228Q-48.572 24.202-48.572 24.178Q-48.572 23.942-48.678 23.778Q-48.784 23.614-48.967 23.532Q-49.150 23.450-49.382 23.450Q-49.710 23.450-49.967 23.553Q-50.223 23.655-50.223 23.932Q-50.223 24.127-50.040 24.236Q-49.857 24.346-49.628 24.387L-49.054 24.493Q-48.808 24.541-48.594 24.669Q-48.381 24.797-48.244 25Q-48.107 25.204-48.107 25.453Q-48.107 25.966-48.473 26.205Q-48.839 26.444-49.375 26.444Q-49.871 26.444-50.203 26.150L-50.469 26.424Q-50.490 26.444-50.517 26.444L-50.565 26.444Q-50.589 26.444-50.616 26.417Q-50.644 26.390-50.644 26.369M-47.479 24.865Q-47.479 24.537-47.343 24.236Q-47.208 23.936-46.973 23.715Q-46.737 23.495-46.433 23.375Q-46.128 23.255-45.804 23.255Q-45.298 23.255-44.949 23.358Q-44.601 23.460-44.601 23.836Q-44.601 23.983-44.698 24.084Q-44.795 24.185-44.942 24.185Q-45.096 24.185-45.195 24.086Q-45.294 23.987-45.294 23.836Q-45.294 23.648-45.154 23.556Q-45.356 23.505-45.797 23.505Q-46.152 23.505-46.381 23.701Q-46.610 23.898-46.711 24.207Q-46.812 24.517-46.812 24.865Q-46.812 25.214-46.686 25.520Q-46.559 25.826-46.304 26.010Q-46.050 26.195-45.694 26.195Q-45.472 26.195-45.288 26.111Q-45.103 26.027-44.968 25.872Q-44.833 25.716-44.775 25.508Q-44.761 25.453-44.707 25.453L-44.594 25.453Q-44.563 25.453-44.541 25.477Q-44.519 25.501-44.519 25.535L-44.519 25.556Q-44.604 25.843-44.792 26.041Q-44.980 26.239-45.245 26.342Q-45.510 26.444-45.804 26.444Q-46.234 26.444-46.622 26.238Q-47.010 26.031-47.244 25.668Q-47.479 25.306-47.479 24.865M-43.873 25.648Q-43.873 25.316-43.649 25.089Q-43.425 24.862-43.081 24.734Q-42.738 24.605-42.365 24.553Q-41.993 24.500-41.688 24.500L-41.688 24.247Q-41.688 24.042-41.796 23.862Q-41.904 23.683-42.085 23.580Q-42.266 23.478-42.475 23.478Q-42.881 23.478-43.117 23.570Q-43.028 23.607-42.982 23.691Q-42.936 23.775-42.936 23.877Q-42.936 23.973-42.982 24.052Q-43.028 24.130-43.109 24.175Q-43.189 24.219-43.278 24.219Q-43.428 24.219-43.529 24.122Q-43.630 24.024-43.630 23.877Q-43.630 23.255-42.475 23.255Q-42.263 23.255-42.013 23.319Q-41.764 23.382-41.562 23.501Q-41.360 23.621-41.234 23.806Q-41.107 23.990-41.107 24.233L-41.107 25.809Q-41.107 25.925-41.046 26.021Q-40.984 26.116-40.872 26.116Q-40.762 26.116-40.697 26.022Q-40.632 25.928-40.632 25.809L-40.632 25.361L-40.366 25.361L-40.366 25.809Q-40.366 26.079-40.593 26.244Q-40.820 26.410-41.101 26.410Q-41.309 26.410-41.446 26.256Q-41.583 26.103-41.606 25.887Q-41.753 26.154-42.035 26.299Q-42.317 26.444-42.642 26.444Q-42.919 26.444-43.203 26.369Q-43.486 26.294-43.679 26.115Q-43.873 25.935-43.873 25.648M-43.257 25.648Q-43.257 25.822-43.156 25.952Q-43.056 26.082-42.900 26.152Q-42.745 26.222-42.581 26.222Q-42.362 26.222-42.153 26.125Q-41.945 26.027-41.817 25.846Q-41.688 25.665-41.688 25.439L-41.688 24.711Q-42.013 24.711-42.379 24.802Q-42.745 24.893-43.001 25.105Q-43.257 25.316-43.257 25.648M-38.281 26.376L-39.884 26.376L-39.884 26.096Q-39.658 26.096-39.510 26.062Q-39.361 26.027-39.361 25.887L-39.361 22.268Q-39.361 21.998-39.468 21.936Q-39.576 21.875-39.884 21.875L-39.884 21.594L-38.807 21.519L-38.807 25.887Q-38.807 26.024-38.657 26.060Q-38.506 26.096-38.281 26.096L-38.281 26.376M-37.727 24.841Q-37.727 24.520-37.602 24.231Q-37.478 23.942-37.252 23.719Q-37.026 23.495-36.731 23.375Q-36.435 23.255-36.117 23.255Q-35.789 23.255-35.528 23.355Q-35.266 23.454-35.090 23.636Q-34.914 23.819-34.820 24.077Q-34.726 24.335-34.726 24.667Q-34.726 24.759-34.808 24.780L-37.064 24.780L-37.064 24.841Q-37.064 25.429-36.780 25.812Q-36.497 26.195-35.929 26.195Q-35.608 26.195-35.340 26.002Q-35.071 25.809-34.982 25.494Q-34.976 25.453-34.900 25.439L-34.808 25.439Q-34.726 25.463-34.726 25.535Q-34.726 25.542-34.733 25.569Q-34.846 25.966-35.217 26.205Q-35.587 26.444-36.011 26.444Q-36.449 26.444-36.849 26.236Q-37.249 26.027-37.488 25.660Q-37.727 25.293-37.727 24.841M-37.057 24.571L-35.242 24.571Q-35.242 24.294-35.340 24.042Q-35.437 23.789-35.635 23.633Q-35.833 23.478-36.117 23.478Q-36.394 23.478-36.608 23.636Q-36.821 23.795-36.939 24.050Q-37.057 24.305-37.057 24.571M-34.138 24.865Q-34.138 24.527-33.998 24.236Q-33.858 23.946-33.614 23.732Q-33.369 23.519-33.065 23.404Q-32.761 23.290-32.436 23.290Q-32.166 23.290-31.903 23.389Q-31.640 23.488-31.448 23.666L-31.448 22.268Q-31.448 21.998-31.556 21.936Q-31.664 21.875-31.975 21.875L-31.975 21.594L-30.898 21.519L-30.898 25.703Q-30.898 25.891-30.843 25.974Q-30.789 26.058-30.688 26.077Q-30.587 26.096-30.372 26.096L-30.372 26.376L-31.479 26.444L-31.479 26.027Q-31.896 26.444-32.521 26.444Q-32.952 26.444-33.325 26.232Q-33.697 26.021-33.918 25.660Q-34.138 25.299-34.138 24.865M-32.463 26.222Q-32.255 26.222-32.069 26.150Q-31.882 26.079-31.729 25.942Q-31.575 25.805-31.479 25.627L-31.479 24.018Q-31.564 23.871-31.710 23.751Q-31.855 23.631-32.024 23.572Q-32.193 23.512-32.375 23.512Q-32.935 23.512-33.203 23.901Q-33.472 24.291-33.472 24.872Q-33.472 25.443-33.238 25.833Q-33.003 26.222-32.463 26.222\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fsvg>\u003Cfigcaption class=\"tikz-cap\">Without the \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:1.1072em;vertical-align:-0.25em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">1\u002F\u003C\u002Fspan>\u003Cspan class=\"mord sqrt\">\u003Cspan class=\"vlist-t vlist-t2\">\u003Cspan class=\"vlist-r\">\u003Cspan class=\"vlist\" style=\"height:0.8572em;\">\u003Cspan class=\"svg-align\" style=\"top:-3em;\">\u003Cspan class=\"pstrut\" style=\"height:3em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\" style=\"padding-left:0.833em;\">\u003Cspan class=\"mord\">\u003Cspan class=\"mord mathnormal\">d\u003C\u002Fspan>\u003Cspan class=\"msupsub\">\u003Cspan class=\"vlist-t vlist-t2\">\u003Cspan class=\"vlist-r\">\u003Cspan class=\"vlist\" style=\"height:0.3361em;\">\u003Cspan style=\"top:-2.55em;margin-left:0em;margin-right:0.05em;\">\u003Cspan class=\"pstrut\" style=\"height:2.7em;\">\u003C\u002Fspan>\u003Cspan class=\"sizing reset-size6 size3 mtight\">\u003Cspan class=\"mord mathnormal mtight\" style=\"margin-right:0.0315em;\">k\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"vlist-s\">​\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"vlist-r\">\u003Cspan class=\"vlist\" style=\"height:0.15em;\">\u003Cspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan style=\"top:-2.8172em;\">\u003Cspan class=\"pstrut\" style=\"height:3em;\">\u003C\u002Fspan>\u003Cspan class=\"hide-tail\" style=\"min-width:0.853em;height:1.08em;\">\u003Csvg xmlns=\"http:\u002F\u002Fwww.w3.org\u002F2000\u002Fsvg\" width=\"400em\" height=\"1.08em\" viewBox=\"0 0 400000 1080\" preserveAspectRatio=\"xMinYMin slice\">\u003Cpath d=\"M95,702c-2.7,0,-7.17,-2.7,-13.5,-8c-5.8,-5.3,-9.5,-10,-9.5,-14c0,-2,0.3,-3.3,1,-4c1.3,-2.7,23.83,-20.7,67.5,-54c44.2,-33.3,65.8,-50.3,66.5,-51c1.3,-1.3,3,-2,5,-2c4.7,0,8.7,3.3,12,10s173,378,173,378c0.7,0,35.3,-71,104,-213c68.7,-142,137.5,-285,206.5,-429c69,-144,104.5,-217.7,106.5,-221l0 -0c5.3,-9.3,12,-14,20,-14H400000v40H845.2724s-225.272,467,-225.272,467s-235,486,-235,486c-2.7,4.7,-9,7,-19,7c-6,0,-10,-1,-12,-3s-194,-422,-194,-422s-65,47,-65,47zM834 80h400000v40h-400000z\"\u002F>\u003C\u002Fsvg>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"vlist-s\">​\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"vlist-r\">\u003Cspan class=\"vlist\" style=\"height:0.1828em;\">\u003Cspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan> correction the softmax saturates: large-variance scores (black) collapse onto one key, while scaled scores (blue) stay usable.\u003C\u002Ffigcaption>","\u003Csvg style=\"width:100%;max-width:322.845px;height:auto\" xmlns=\"http:\u002F\u002Fwww.w3.org\u002F2000\u002Fsvg\" viewBox=\"-75 -75 242.134 157.708\">\u003Cg stroke=\"currentColor\" style=\"stroke-miterlimit:10;stroke-width:.4\">\u003Cg transform=\"rotate(-42.001 -185.117 4.793)scale(.99999)\">\u003Cpath d=\"M-24.092 79.875L-24.092 77.978L-24.731 77.978L-24.731 77.756Q-24.413 77.756-24.196 77.546Q-23.979 77.336-23.879 77.026Q-23.778 76.717-23.778 76.409L-23.511 76.409L-23.511 77.698L-22.434 77.698L-22.434 77.978L-23.511 77.978L-23.511 79.862Q-23.511 80.138-23.407 80.337Q-23.303 80.535-23.043 80.535Q-22.886 80.535-22.780 80.431Q-22.674 80.326-22.624 80.173Q-22.575 80.019-22.575 79.862L-22.575 79.448L-22.308 79.448L-22.308 79.875Q-22.308 80.101-22.407 80.311Q-22.506 80.521-22.691 80.653Q-22.875 80.784-23.104 80.784Q-23.542 80.784-23.817 80.547Q-24.092 80.309-24.092 79.875M-19.816 80.716L-21.450 80.716L-21.450 80.436Q-21.221 80.436-21.072 80.402Q-20.924 80.367-20.924 80.227L-20.924 76.608Q-20.924 76.338-21.031 76.276Q-21.139 76.215-21.450 76.215L-21.450 75.934L-20.370 75.859L-20.370 78.245Q-20.264 78.060-20.086 77.918Q-19.909 77.777-19.700 77.703Q-19.492 77.630-19.266 77.630Q-18.760 77.630-18.476 77.853Q-18.193 78.077-18.193 78.573L-18.193 80.227Q-18.193 80.364-18.044 80.400Q-17.895 80.436-17.670 80.436L-17.670 80.716L-19.300 80.716L-19.300 80.436Q-19.071 80.436-18.923 80.402Q-18.774 80.367-18.774 80.227L-18.774 78.587Q-18.774 78.252-18.893 78.052Q-19.013 77.852-19.328 77.852Q-19.598 77.852-19.832 77.988Q-20.066 78.125-20.204 78.359Q-20.343 78.593-20.343 78.867L-20.343 80.227Q-20.343 80.364-20.192 80.400Q-20.042 80.436-19.816 80.436L-19.816 80.716M-17.123 79.181Q-17.123 78.860-16.998 78.571Q-16.873 78.282-16.648 78.059Q-16.422 77.835-16.127 77.715Q-15.831 77.595-15.513 77.595Q-15.185 77.595-14.923 77.695Q-14.662 77.794-14.486 77.976Q-14.310 78.159-14.216 78.417Q-14.122 78.675-14.122 79.007Q-14.122 79.099-14.204 79.120L-16.460 79.120L-16.460 79.181Q-16.460 79.769-16.176 80.152Q-15.892 80.535-15.325 80.535Q-15.004 80.535-14.735 80.342Q-14.467 80.149-14.378 79.834Q-14.371 79.793-14.296 79.779L-14.204 79.779Q-14.122 79.803-14.122 79.875Q-14.122 79.882-14.129 79.909Q-14.242 80.306-14.612 80.545Q-14.983 80.784-15.407 80.784Q-15.845 80.784-16.245 80.576Q-16.644 80.367-16.884 80Q-17.123 79.633-17.123 79.181M-16.453 78.911L-14.638 78.911Q-14.638 78.634-14.735 78.382Q-14.833 78.129-15.031 77.973Q-15.229 77.818-15.513 77.818Q-15.790 77.818-16.004 77.976Q-16.217 78.135-16.335 78.390Q-16.453 78.645-16.453 78.911\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"rotate(-42.001 -171.602 -30.415)scale(.99999)\">\u003Cpath d=\"M-24.560 79.988Q-24.560 79.656-24.337 79.429Q-24.113 79.202-23.769 79.074Q-23.426 78.945-23.053 78.893Q-22.681 78.840-22.376 78.840L-22.376 78.587Q-22.376 78.382-22.484 78.202Q-22.592 78.023-22.773 77.920Q-22.954 77.818-23.162 77.818Q-23.569 77.818-23.805 77.910Q-23.716 77.947-23.670 78.031Q-23.624 78.115-23.624 78.217Q-23.624 78.313-23.670 78.392Q-23.716 78.470-23.797 78.515Q-23.877 78.559-23.966 78.559Q-24.116 78.559-24.217 78.462Q-24.318 78.364-24.318 78.217Q-24.318 77.595-23.162 77.595Q-22.951 77.595-22.701 77.659Q-22.452 77.722-22.250 77.841Q-22.048 77.961-21.922 78.146Q-21.795 78.330-21.795 78.573L-21.795 80.149Q-21.795 80.265-21.734 80.361Q-21.672 80.456-21.559 80.456Q-21.450 80.456-21.385 80.362Q-21.320 80.268-21.320 80.149L-21.320 79.701L-21.054 79.701L-21.054 80.149Q-21.054 80.419-21.281 80.584Q-21.508 80.750-21.788 80.750Q-21.997 80.750-22.134 80.596Q-22.270 80.443-22.294 80.227Q-22.441 80.494-22.723 80.639Q-23.005 80.784-23.330 80.784Q-23.607 80.784-23.891 80.709Q-24.174 80.634-24.367 80.455Q-24.560 80.275-24.560 79.988M-23.945 79.988Q-23.945 80.162-23.844 80.292Q-23.744 80.422-23.588 80.492Q-23.433 80.562-23.268 80.562Q-23.050 80.562-22.841 80.465Q-22.633 80.367-22.505 80.186Q-22.376 80.005-22.376 79.779L-22.376 79.051Q-22.701 79.051-23.067 79.142Q-23.433 79.233-23.689 79.445Q-23.945 79.656-23.945 79.988M-18.955 80.716L-20.589 80.716L-20.589 80.436Q-20.360 80.436-20.211 80.402Q-20.062 80.367-20.062 80.227L-20.062 78.378Q-20.062 78.108-20.170 78.047Q-20.278 77.985-20.589 77.985L-20.589 77.705L-19.529 77.630L-19.529 78.279Q-19.358 77.971-19.054 77.800Q-18.750 77.630-18.405 77.630Q-17.899 77.630-17.615 77.853Q-17.331 78.077-17.331 78.573L-17.331 80.227Q-17.331 80.364-17.183 80.400Q-17.034 80.436-16.808 80.436L-16.808 80.716L-18.439 80.716L-18.439 80.436Q-18.210 80.436-18.061 80.402Q-17.912 80.367-17.912 80.227L-17.912 78.587Q-17.912 78.252-18.032 78.052Q-18.152 77.852-18.466 77.852Q-18.736 77.852-18.970 77.988Q-19.204 78.125-19.343 78.359Q-19.481 78.593-19.481 78.867L-19.481 80.227Q-19.481 80.364-19.331 80.400Q-19.181 80.436-18.955 80.436L-18.955 80.716M-14.604 80.716L-16.156 80.716L-16.156 80.436Q-15.930 80.436-15.781 80.402Q-15.633 80.367-15.633 80.227L-15.633 78.378Q-15.633 78.190-15.681 78.106Q-15.728 78.023-15.826 78.004Q-15.923 77.985-16.135 77.985L-16.135 77.705L-15.079 77.630L-15.079 80.227Q-15.079 80.367-14.947 80.402Q-14.816 80.436-14.604 80.436L-14.604 80.716M-15.875 76.409Q-15.875 76.238-15.752 76.119Q-15.629 75.999-15.458 75.999Q-15.291 75.999-15.168 76.119Q-15.045 76.238-15.045 76.409Q-15.045 76.584-15.168 76.707Q-15.291 76.830-15.458 76.830Q-15.629 76.830-15.752 76.707Q-15.875 76.584-15.875 76.409M-12.276 80.716L-13.910 80.716L-13.910 80.436Q-13.681 80.436-13.532 80.402Q-13.384 80.367-13.384 80.227L-13.384 78.378Q-13.384 78.108-13.491 78.047Q-13.599 77.985-13.910 77.985L-13.910 77.705L-12.850 77.630L-12.850 78.279Q-12.680 77.971-12.375 77.800Q-12.071 77.630-11.726 77.630Q-11.326 77.630-11.049 77.770Q-10.772 77.910-10.687 78.258Q-10.519 77.965-10.220 77.797Q-9.921 77.630-9.576 77.630Q-9.070 77.630-8.787 77.853Q-8.503 78.077-8.503 78.573L-8.503 80.227Q-8.503 80.364-8.354 80.400Q-8.205 80.436-7.980 80.436L-7.980 80.716L-9.610 80.716L-9.610 80.436Q-9.385 80.436-9.234 80.400Q-9.084 80.364-9.084 80.227L-9.084 78.587Q-9.084 78.252-9.204 78.052Q-9.323 77.852-9.638 77.852Q-9.908 77.852-10.142 77.988Q-10.376 78.125-10.514 78.359Q-10.653 78.593-10.653 78.867L-10.653 80.227Q-10.653 80.364-10.504 80.400Q-10.355 80.436-10.130 80.436L-10.130 80.716L-11.760 80.716L-11.760 80.436Q-11.531 80.436-11.382 80.402Q-11.234 80.367-11.234 80.227L-11.234 78.587Q-11.234 78.252-11.353 78.052Q-11.473 77.852-11.787 77.852Q-12.058 77.852-12.292 77.988Q-12.526 78.125-12.664 78.359Q-12.803 78.593-12.803 78.867L-12.803 80.227Q-12.803 80.364-12.652 80.400Q-12.502 80.436-12.276 80.436L-12.276 80.716M-7.334 79.988Q-7.334 79.656-7.110 79.429Q-6.886 79.202-6.543 79.074Q-6.199 78.945-5.827 78.893Q-5.454 78.840-5.150 78.840L-5.150 78.587Q-5.150 78.382-5.257 78.202Q-5.365 78.023-5.546 77.920Q-5.727 77.818-5.936 77.818Q-6.343 77.818-6.579 77.910Q-6.490 77.947-6.443 78.031Q-6.397 78.115-6.397 78.217Q-6.397 78.313-6.443 78.392Q-6.490 78.470-6.570 78.515Q-6.650 78.559-6.739 78.559Q-6.890 78.559-6.990 78.462Q-7.091 78.364-7.091 78.217Q-7.091 77.595-5.936 77.595Q-5.724 77.595-5.475 77.659Q-5.225 77.722-5.023 77.841Q-4.822 77.961-4.695 78.146Q-4.569 78.330-4.569 78.573L-4.569 80.149Q-4.569 80.265-4.507 80.361Q-4.446 80.456-4.333 80.456Q-4.224 80.456-4.159 80.362Q-4.094 80.268-4.094 80.149L-4.094 79.701L-3.827 79.701L-3.827 80.149Q-3.827 80.419-4.054 80.584Q-4.282 80.750-4.562 80.750Q-4.770 80.750-4.907 80.596Q-5.044 80.443-5.068 80.227Q-5.215 80.494-5.497 80.639Q-5.779 80.784-6.103 80.784Q-6.380 80.784-6.664 80.709Q-6.948 80.634-7.141 80.455Q-7.334 80.275-7.334 79.988M-6.719 79.988Q-6.719 80.162-6.618 80.292Q-6.517 80.422-6.361 80.492Q-6.206 80.562-6.042 80.562Q-5.823 80.562-5.615 80.465Q-5.406 80.367-5.278 80.186Q-5.150 80.005-5.150 79.779L-5.150 79.051Q-5.475 79.051-5.840 79.142Q-6.206 79.233-6.462 79.445Q-6.719 79.656-6.719 79.988M-1.742 80.716L-3.345 80.716L-3.345 80.436Q-3.120 80.436-2.971 80.402Q-2.822 80.367-2.822 80.227L-2.822 76.608Q-2.822 76.338-2.930 76.276Q-3.037 76.215-3.345 76.215L-3.345 75.934L-2.268 75.859L-2.268 80.227Q-2.268 80.364-2.118 80.400Q-1.968 80.436-1.742 80.436\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"rotate(-42.001 -158.086 -65.622)scale(.99999)\">\u003Cpath d=\"M-24.619 79.205Q-24.619 78.877-24.484 78.576Q-24.349 78.276-24.113 78.055Q-23.877 77.835-23.573 77.715Q-23.268 77.595-22.944 77.595Q-22.438 77.595-22.089 77.698Q-21.741 77.800-21.741 78.176Q-21.741 78.323-21.838 78.424Q-21.935 78.525-22.082 78.525Q-22.236 78.525-22.335 78.426Q-22.434 78.327-22.434 78.176Q-22.434 77.988-22.294 77.896Q-22.496 77.845-22.937 77.845Q-23.292 77.845-23.521 78.041Q-23.750 78.238-23.851 78.547Q-23.952 78.857-23.952 79.205Q-23.952 79.554-23.826 79.860Q-23.699 80.166-23.444 80.350Q-23.190 80.535-22.834 80.535Q-22.612 80.535-22.428 80.451Q-22.243 80.367-22.108 80.212Q-21.973 80.056-21.915 79.848Q-21.901 79.793-21.847 79.793L-21.734 79.793Q-21.703 79.793-21.681 79.817Q-21.659 79.841-21.659 79.875L-21.659 79.896Q-21.744 80.183-21.932 80.381Q-22.120 80.579-22.385 80.682Q-22.650 80.784-22.944 80.784Q-23.374 80.784-23.762 80.578Q-24.150 80.371-24.384 80.008Q-24.619 79.646-24.619 79.205M-19.321 80.716L-21.057 80.716L-21.057 80.436Q-20.828 80.436-20.679 80.402Q-20.531 80.367-20.531 80.227L-20.531 78.378Q-20.531 78.108-20.638 78.047Q-20.746 77.985-21.057 77.985L-21.057 77.705L-20.028 77.630L-20.028 78.337Q-19.898 78.029-19.656 77.830Q-19.413 77.630-19.095 77.630Q-18.876 77.630-18.705 77.754Q-18.535 77.879-18.535 78.091Q-18.535 78.228-18.634 78.327Q-18.733 78.426-18.866 78.426Q-19.003 78.426-19.102 78.327Q-19.201 78.228-19.201 78.091Q-19.201 77.951-19.102 77.852Q-19.392 77.852-19.592 78.048Q-19.792 78.245-19.885 78.539Q-19.977 78.833-19.977 79.113L-19.977 80.227Q-19.977 80.436-19.321 80.436L-19.321 80.716M-17.991 79.233Q-17.991 78.891-17.856 78.592Q-17.721 78.293-17.482 78.069Q-17.243 77.845-16.925 77.720Q-16.607 77.595-16.275 77.595Q-15.831 77.595-15.431 77.811Q-15.031 78.026-14.797 78.404Q-14.563 78.781-14.563 79.233Q-14.563 79.574-14.705 79.858Q-14.847 80.142-15.091 80.349Q-15.335 80.555-15.645 80.670Q-15.954 80.784-16.275 80.784Q-16.706 80.784-17.108 80.583Q-17.509 80.381-17.750 80.029Q-17.991 79.677-17.991 79.233M-16.275 80.535Q-15.674 80.535-15.450 80.157Q-15.226 79.779-15.226 79.147Q-15.226 78.535-15.460 78.176Q-15.694 77.818-16.275 77.818Q-17.328 77.818-17.328 79.147Q-17.328 79.779-17.102 80.157Q-16.877 80.535-16.275 80.535M-13.968 80.709L-13.968 79.646Q-13.968 79.622-13.941 79.595Q-13.913 79.568-13.890 79.568L-13.780 79.568Q-13.715 79.568-13.702 79.626Q-13.606 80.060-13.360 80.311Q-13.114 80.562-12.700 80.562Q-12.358 80.562-12.105 80.429Q-11.852 80.296-11.852 79.988Q-11.852 79.831-11.946 79.716Q-12.040 79.602-12.179 79.533Q-12.317 79.465-12.485 79.427L-13.066 79.328Q-13.421 79.260-13.695 79.039Q-13.968 78.819-13.968 78.477Q-13.968 78.228-13.857 78.053Q-13.746 77.879-13.560 77.780Q-13.373 77.681-13.158 77.638Q-12.943 77.595-12.700 77.595Q-12.287 77.595-12.006 77.777L-11.791 77.602Q-11.781 77.599-11.774 77.597Q-11.767 77.595-11.757 77.595L-11.705 77.595Q-11.678 77.595-11.654 77.619Q-11.630 77.643-11.630 77.671L-11.630 78.518Q-11.630 78.539-11.654 78.566Q-11.678 78.593-11.705 78.593L-11.818 78.593Q-11.846 78.593-11.871 78.568Q-11.897 78.542-11.897 78.518Q-11.897 78.282-12.003 78.118Q-12.109 77.954-12.292 77.872Q-12.475 77.790-12.707 77.790Q-13.035 77.790-13.291 77.893Q-13.548 77.995-13.548 78.272Q-13.548 78.467-13.365 78.576Q-13.182 78.686-12.953 78.727L-12.379 78.833Q-12.133 78.881-11.919 79.009Q-11.705 79.137-11.569 79.340Q-11.432 79.544-11.432 79.793Q-11.432 80.306-11.798 80.545Q-12.163 80.784-12.700 80.784Q-13.196 80.784-13.527 80.490L-13.794 80.764Q-13.814 80.784-13.842 80.784L-13.890 80.784Q-13.913 80.784-13.941 80.757Q-13.968 80.730-13.968 80.709M-10.803 80.709L-10.803 79.646Q-10.803 79.622-10.776 79.595Q-10.748 79.568-10.725 79.568L-10.615 79.568Q-10.550 79.568-10.537 79.626Q-10.441 80.060-10.195 80.311Q-9.949 80.562-9.535 80.562Q-9.193 80.562-8.940 80.429Q-8.687 80.296-8.687 79.988Q-8.687 79.831-8.781 79.716Q-8.875 79.602-9.014 79.533Q-9.152 79.465-9.320 79.427L-9.901 79.328Q-10.256 79.260-10.530 79.039Q-10.803 78.819-10.803 78.477Q-10.803 78.228-10.692 78.053Q-10.581 77.879-10.395 77.780Q-10.208 77.681-9.993 77.638Q-9.778 77.595-9.535 77.595Q-9.121 77.595-8.841 77.777L-8.626 77.602Q-8.616 77.599-8.609 77.597Q-8.602 77.595-8.592 77.595L-8.540 77.595Q-8.513 77.595-8.489 77.619Q-8.465 77.643-8.465 77.671L-8.465 78.518Q-8.465 78.539-8.489 78.566Q-8.513 78.593-8.540 78.593L-8.653 78.593Q-8.681 78.593-8.706 78.568Q-8.732 78.542-8.732 78.518Q-8.732 78.282-8.838 78.118Q-8.944 77.954-9.127 77.872Q-9.309 77.790-9.542 77.790Q-9.870 77.790-10.126 77.893Q-10.383 77.995-10.383 78.272Q-10.383 78.467-10.200 78.576Q-10.017 78.686-9.788 78.727L-9.214 78.833Q-8.968 78.881-8.754 79.009Q-8.540 79.137-8.404 79.340Q-8.267 79.544-8.267 79.793Q-8.267 80.306-8.633 80.545Q-8.998 80.784-9.535 80.784Q-10.031 80.784-10.362 80.490L-10.629 80.764Q-10.649 80.784-10.677 80.784L-10.725 80.784Q-10.748 80.784-10.776 80.757Q-10.803 80.730-10.803 80.709M-7.679 79.181Q-7.679 78.860-7.554 78.571Q-7.430 78.282-7.204 78.059Q-6.978 77.835-6.683 77.715Q-6.387 77.595-6.069 77.595Q-5.741 77.595-5.480 77.695Q-5.218 77.794-5.042 77.976Q-4.866 78.159-4.772 78.417Q-4.678 78.675-4.678 79.007Q-4.678 79.099-4.760 79.120L-7.016 79.120L-7.016 79.181Q-7.016 79.769-6.732 80.152Q-6.449 80.535-5.881 80.535Q-5.560 80.535-5.292 80.342Q-5.023 80.149-4.934 79.834Q-4.928 79.793-4.852 79.779L-4.760 79.779Q-4.678 79.803-4.678 79.875Q-4.678 79.882-4.685 79.909Q-4.798 80.306-5.169 80.545Q-5.539 80.784-5.963 80.784Q-6.401 80.784-6.801 80.576Q-7.201 80.367-7.440 80Q-7.679 79.633-7.679 79.181M-7.009 78.911L-5.194 78.911Q-5.194 78.634-5.292 78.382Q-5.389 78.129-5.587 77.973Q-5.786 77.818-6.069 77.818Q-6.346 77.818-6.560 77.976Q-6.773 78.135-6.891 78.390Q-7.009 78.645-7.009 78.911M-4.090 79.205Q-4.090 78.867-3.950 78.576Q-3.810 78.286-3.566 78.072Q-3.321 77.859-3.017 77.744Q-2.713 77.630-2.388 77.630Q-2.118 77.630-1.855 77.729Q-1.592 77.828-1.400 78.006L-1.400 76.608Q-1.400 76.338-1.508 76.276Q-1.616 76.215-1.927 76.215L-1.927 75.934L-0.850 75.859L-0.850 80.043Q-0.850 80.231-0.795 80.314Q-0.741 80.398-0.640 80.417Q-0.539 80.436-0.324 80.436L-0.324 80.716L-1.431 80.784L-1.431 80.367Q-1.848 80.784-2.474 80.784Q-2.904 80.784-3.277 80.572Q-3.649 80.361-3.870 80Q-4.090 79.639-4.090 79.205M-2.415 80.562Q-2.207 80.562-2.021 80.490Q-1.834 80.419-1.681 80.282Q-1.527 80.145-1.431 79.967L-1.431 78.358Q-1.516 78.211-1.662 78.091Q-1.807 77.971-1.976 77.912Q-2.145 77.852-2.327 77.852Q-2.887 77.852-3.155 78.241Q-3.424 78.631-3.424 79.212Q-3.424 79.783-3.190 80.173Q-2.955 80.562-2.415 80.562\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg stroke=\"none\" font-family=\"cmr7\" font-size=\"7\">\u003Cg transform=\"rotate(-42.001 -144.571 -100.83)scale(.99999)\">\u003Cpath d=\"M-23.812 80.716L-24.079 80.716L-24.079 76.608Q-24.079 76.338-24.186 76.276Q-24.294 76.215-24.605 76.215L-24.605 75.934L-23.525 75.859L-23.525 78.029Q-23.316 77.838-23.031 77.734Q-22.745 77.630-22.448 77.630Q-22.130 77.630-21.833 77.751Q-21.536 77.872-21.313 78.088Q-21.091 78.303-20.965 78.588Q-20.838 78.874-20.838 79.205Q-20.838 79.650-21.078 80.014Q-21.317 80.378-21.710 80.581Q-22.103 80.784-22.547 80.784Q-22.742 80.784-22.932 80.728Q-23.121 80.672-23.282 80.567Q-23.443 80.463-23.583 80.302L-23.812 80.716M-23.497 78.371L-23.497 79.988Q-23.361 80.248-23.120 80.405Q-22.879 80.562-22.602 80.562Q-22.308 80.562-22.096 80.455Q-21.884 80.347-21.751 80.155Q-21.618 79.964-21.559 79.725Q-21.501 79.486-21.501 79.205Q-21.501 78.846-21.595 78.542Q-21.689 78.238-21.917 78.045Q-22.144 77.852-22.510 77.852Q-22.810 77.852-23.077 77.988Q-23.344 78.125-23.497 78.371\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"rotate(-42.001 -144.571 -100.83)scale(.99999)\">\u003Cpath d=\"M-20.028 79.181Q-20.028 78.860-19.903 78.571Q-19.778 78.282-19.552 78.059Q-19.327 77.835-19.031 77.715Q-18.736 77.595-18.418 77.595Q-18.090 77.595-17.828 77.695Q-17.567 77.794-17.391 77.976Q-17.215 78.159-17.121 78.417Q-17.027 78.675-17.027 79.007Q-17.027 79.099-17.109 79.120L-19.364 79.120L-19.364 79.181Q-19.364 79.769-19.081 80.152Q-18.797 80.535-18.230 80.535Q-17.908 80.535-17.640 80.342Q-17.372 80.149-17.283 79.834Q-17.276 79.793-17.201 79.779L-17.109 79.779Q-17.027 79.803-17.027 79.875Q-17.027 79.882-17.033 79.909Q-17.146 80.306-17.517 80.545Q-17.888 80.784-18.312 80.784Q-18.749 80.784-19.149 80.576Q-19.549 80.367-19.788 80Q-20.028 79.633-20.028 79.181M-19.358 78.911L-17.543 78.911Q-17.543 78.634-17.640 78.382Q-17.738 78.129-17.936 77.973Q-18.134 77.818-18.418 77.818Q-18.695 77.818-18.908 77.976Q-19.122 78.135-19.240 78.390Q-19.358 78.645-19.358 78.911M-16.439 79.205Q-16.439 78.877-16.304 78.576Q-16.169 78.276-15.933 78.055Q-15.697 77.835-15.393 77.715Q-15.089 77.595-14.764 77.595Q-14.258 77.595-13.909 77.698Q-13.561 77.800-13.561 78.176Q-13.561 78.323-13.658 78.424Q-13.756 78.525-13.903 78.525Q-14.056 78.525-14.155 78.426Q-14.255 78.327-14.255 78.176Q-14.255 77.988-14.114 77.896Q-14.316 77.845-14.757 77.845Q-15.113 77.845-15.342 78.041Q-15.571 78.238-15.671 78.547Q-15.772 78.857-15.772 79.205Q-15.772 79.554-15.646 79.860Q-15.519 80.166-15.265 80.350Q-15.010 80.535-14.655 80.535Q-14.432 80.535-14.248 80.451Q-14.063 80.367-13.928 80.212Q-13.793 80.056-13.735 79.848Q-13.721 79.793-13.667 79.793L-13.554 79.793Q-13.523 79.793-13.501 79.817Q-13.479 79.841-13.479 79.875L-13.479 79.896Q-13.564 80.183-13.752 80.381Q-13.940 80.579-14.205 80.682Q-14.470 80.784-14.764 80.784Q-15.195 80.784-15.582 80.578Q-15.970 80.371-16.205 80.008Q-16.439 79.646-16.439 79.205M-12.833 79.988Q-12.833 79.656-12.609 79.429Q-12.385 79.202-12.041 79.074Q-11.698 78.945-11.325 78.893Q-10.953 78.840-10.649 78.840L-10.649 78.587Q-10.649 78.382-10.756 78.202Q-10.864 78.023-11.045 77.920Q-11.226 77.818-11.435 77.818Q-11.842 77.818-12.077 77.910Q-11.988 77.947-11.942 78.031Q-11.896 78.115-11.896 78.217Q-11.896 78.313-11.942 78.392Q-11.988 78.470-12.069 78.515Q-12.149 78.559-12.238 78.559Q-12.388 78.559-12.489 78.462Q-12.590 78.364-12.590 78.217Q-12.590 77.595-11.435 77.595Q-11.223 77.595-10.973 77.659Q-10.724 77.722-10.522 77.841Q-10.321 77.961-10.194 78.146Q-10.068 78.330-10.068 78.573L-10.068 80.149Q-10.068 80.265-10.006 80.361Q-9.945 80.456-9.832 80.456Q-9.722 80.456-9.657 80.362Q-9.593 80.268-9.593 80.149L-9.593 79.701L-9.326 79.701L-9.326 80.149Q-9.326 80.419-9.553 80.584Q-9.780 80.750-10.061 80.750Q-10.269 80.750-10.406 80.596Q-10.543 80.443-10.567 80.227Q-10.714 80.494-10.996 80.639Q-11.278 80.784-11.602 80.784Q-11.879 80.784-12.163 80.709Q-12.447 80.634-12.640 80.455Q-12.833 80.275-12.833 79.988M-12.218 79.988Q-12.218 80.162-12.117 80.292Q-12.016 80.422-11.860 80.492Q-11.705 80.562-11.541 80.562Q-11.322 80.562-11.113 80.465Q-10.905 80.367-10.777 80.186Q-10.649 80.005-10.649 79.779L-10.649 79.051Q-10.973 79.051-11.339 79.142Q-11.705 79.233-11.961 79.445Q-12.218 79.656-12.218 79.988M-8.335 79.882L-8.335 78.378Q-8.335 78.108-8.442 78.047Q-8.550 77.985-8.861 77.985L-8.861 77.705L-7.754 77.630L-7.754 79.862L-7.754 79.882Q-7.754 80.162-7.702 80.306Q-7.651 80.449-7.509 80.506Q-7.367 80.562-7.080 80.562Q-6.827 80.562-6.622 80.422Q-6.417 80.282-6.301 80.056Q-6.185 79.831-6.185 79.581L-6.185 78.378Q-6.185 78.108-6.292 78.047Q-6.400 77.985-6.711 77.985L-6.711 77.705L-5.604 77.630L-5.604 80.043Q-5.604 80.234-5.551 80.316Q-5.498 80.398-5.397 80.417Q-5.296 80.436-5.081 80.436L-5.081 80.716L-6.157 80.784L-6.157 80.220Q-6.267 80.402-6.412 80.525Q-6.557 80.648-6.744 80.716Q-6.930 80.784-7.132 80.784Q-8.335 80.784-8.335 79.882M-4.493 80.709L-4.493 79.646Q-4.493 79.622-4.466 79.595Q-4.438 79.568-4.414 79.568L-4.305 79.568Q-4.240 79.568-4.226 79.626Q-4.131 80.060-3.884 80.311Q-3.638 80.562-3.225 80.562Q-2.883 80.562-2.630 80.429Q-2.377 80.296-2.377 79.988Q-2.377 79.831-2.471 79.716Q-2.565 79.602-2.704 79.533Q-2.842 79.465-3.009 79.427L-3.591 79.328Q-3.946 79.260-4.219 79.039Q-4.493 78.819-4.493 78.477Q-4.493 78.228-4.382 78.053Q-4.271 77.879-4.084 77.780Q-3.898 77.681-3.683 77.638Q-3.468 77.595-3.225 77.595Q-2.811 77.595-2.531 77.777L-2.316 77.602Q-2.305 77.599-2.299 77.597Q-2.292 77.595-2.281 77.595L-2.230 77.595Q-2.203 77.595-2.179 77.619Q-2.155 77.643-2.155 77.671L-2.155 78.518Q-2.155 78.539-2.179 78.566Q-2.203 78.593-2.230 78.593L-2.343 78.593Q-2.370 78.593-2.396 78.568Q-2.422 78.542-2.422 78.518Q-2.422 78.282-2.528 78.118Q-2.634 77.954-2.816 77.872Q-2.999 77.790-3.232 77.790Q-3.560 77.790-3.816 77.893Q-4.072 77.995-4.072 78.272Q-4.072 78.467-3.890 78.576Q-3.707 78.686-3.478 78.727L-2.904 78.833Q-2.657 78.881-2.444 79.009Q-2.230 79.137-2.093 79.340Q-1.957 79.544-1.957 79.793Q-1.957 80.306-2.322 80.545Q-2.688 80.784-3.225 80.784Q-3.720 80.784-4.052 80.490L-4.319 80.764Q-4.339 80.784-4.366 80.784L-4.414 80.784Q-4.438 80.784-4.466 80.757Q-4.493 80.730-4.493 80.709M-1.369 79.181Q-1.369 78.860-1.244 78.571Q-1.119 78.282-0.894 78.059Q-0.668 77.835-0.373 77.715Q-0.077 77.595 0.241 77.595Q0.569 77.595 0.831 77.695Q1.092 77.794 1.268 77.976Q1.444 78.159 1.538 78.417Q1.632 78.675 1.632 79.007Q1.632 79.099 1.550 79.120L-0.706 79.120L-0.706 79.181Q-0.706 79.769-0.422 80.152Q-0.138 80.535 0.429 80.535Q0.750 80.535 1.019 80.342Q1.287 80.149 1.376 79.834Q1.383 79.793 1.458 79.779L1.550 79.779Q1.632 79.803 1.632 79.875Q1.632 79.882 1.625 79.909Q1.512 80.306 1.142 80.545Q0.771 80.784 0.347 80.784Q-0.091 80.784-0.490 80.576Q-0.890 80.367-1.130 80Q-1.369 79.633-1.369 79.181M-0.699 78.911L1.116 78.911Q1.116 78.634 1.019 78.382Q0.921 78.129 0.723 77.973Q0.525 77.818 0.241 77.818Q-0.036 77.818-0.249 77.976Q-0.463 78.135-0.581 78.390Q-0.699 78.645-0.699 78.911\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg transform=\"rotate(-42.001 -131.156 -135.998)scale(.99999)\">\u003Cpath d=\"M-23.002 80.716L-24.554 80.716L-24.554 80.436Q-24.328 80.436-24.179 80.402Q-24.031 80.367-24.031 80.227L-24.031 78.378Q-24.031 78.190-24.079 78.106Q-24.126 78.023-24.224 78.004Q-24.321 77.985-24.533 77.985L-24.533 77.705L-23.477 77.630L-23.477 80.227Q-23.477 80.367-23.345 80.402Q-23.214 80.436-23.002 80.436L-23.002 80.716M-24.273 76.409Q-24.273 76.238-24.150 76.119Q-24.027 75.999-23.856 75.999Q-23.689 75.999-23.566 76.119Q-23.443 76.238-23.443 76.409Q-23.443 76.584-23.566 76.707Q-23.689 76.830-23.856 76.830Q-24.027 76.830-24.150 76.707Q-24.273 76.584-24.273 76.409M-21.829 79.875L-21.829 77.978L-22.469 77.978L-22.469 77.756Q-22.151 77.756-21.934 77.546Q-21.717 77.336-21.616 77.026Q-21.515 76.717-21.515 76.409L-21.248 76.409L-21.248 77.698L-20.172 77.698L-20.172 77.978L-21.248 77.978L-21.248 79.862Q-21.248 80.138-21.144 80.337Q-21.040 80.535-20.780 80.535Q-20.623 80.535-20.517 80.431Q-20.411 80.326-20.361 80.173Q-20.312 80.019-20.312 79.862L-20.312 79.448L-20.045 79.448L-20.045 79.875Q-20.045 80.101-20.144 80.311Q-20.244 80.521-20.428 80.653Q-20.613 80.784-20.842 80.784Q-21.279 80.784-21.554 80.547Q-21.829 80.309-21.829 79.875\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg stroke=\"none\" font-family=\"cmr7\" font-size=\"7\">\u003Cg transform=\"rotate(-42.001 -118.744 -170.782)scale(.99999)\">\u003Cpath d=\"M-23.231 80.689L-24.212 78.190Q-24.273 78.047-24.391 78.012Q-24.509 77.978-24.725 77.978L-24.725 77.698L-23.245 77.698L-23.245 77.978Q-23.624 77.978-23.624 78.139Q-23.624 78.149-23.610 78.190L-22.896 80.022L-22.223 78.317Q-22.253 78.245-22.253 78.217Q-22.253 78.190-22.281 78.190Q-22.342 78.043-22.460 78.011Q-22.578 77.978-22.790 77.978L-22.790 77.698L-21.392 77.698L-21.392 77.978Q-21.768 77.978-21.768 78.139Q-21.768 78.170-21.761 78.190L-21.006 80.128L-20.319 78.378Q-20.298 78.327-20.298 78.272Q-20.298 78.132-20.411 78.055Q-20.524 77.978-20.664 77.978L-20.664 77.698L-19.444 77.698L-19.444 77.978Q-19.649 77.978-19.804 78.084Q-19.960 78.190-20.032 78.378L-20.937 80.689Q-20.972 80.784-21.084 80.784L-21.153 80.784Q-21.262 80.784-21.300 80.689L-22.082 78.686L-22.869 80.689Q-22.903 80.784-23.016 80.784L-23.084 80.784Q-23.193 80.784-23.231 80.689\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"rotate(-42.001 -118.744 -170.782)scale(.99999)\">\u003Cpath d=\"M-19.067 79.988Q-19.067 79.656-18.844 79.429Q-18.620 79.202-18.276 79.074Q-17.933 78.945-17.560 78.893Q-17.188 78.840-16.883 78.840L-16.883 78.587Q-16.883 78.382-16.991 78.202Q-17.099 78.023-17.280 77.920Q-17.461 77.818-17.669 77.818Q-18.076 77.818-18.312 77.910Q-18.223 77.947-18.177 78.031Q-18.131 78.115-18.131 78.217Q-18.131 78.313-18.177 78.392Q-18.223 78.470-18.304 78.515Q-18.384 78.559-18.473 78.559Q-18.623 78.559-18.724 78.462Q-18.825 78.364-18.825 78.217Q-18.825 77.595-17.669 77.595Q-17.458 77.595-17.208 77.659Q-16.959 77.722-16.757 77.841Q-16.555 77.961-16.429 78.146Q-16.302 78.330-16.302 78.573L-16.302 80.149Q-16.302 80.265-16.241 80.361Q-16.179 80.456-16.066 80.456Q-15.957 80.456-15.892 80.362Q-15.827 80.268-15.827 80.149L-15.827 79.701L-15.561 79.701L-15.561 80.149Q-15.561 80.419-15.788 80.584Q-16.015 80.750-16.295 80.750Q-16.504 80.750-16.641 80.596Q-16.777 80.443-16.801 80.227Q-16.948 80.494-17.230 80.639Q-17.512 80.784-17.837 80.784Q-18.114 80.784-18.398 80.709Q-18.681 80.634-18.874 80.455Q-19.067 80.275-19.067 79.988M-18.452 79.988Q-18.452 80.162-18.351 80.292Q-18.251 80.422-18.095 80.492Q-17.940 80.562-17.775 80.562Q-17.557 80.562-17.348 80.465Q-17.140 80.367-17.012 80.186Q-16.883 80.005-16.883 79.779L-16.883 79.051Q-17.208 79.051-17.574 79.142Q-17.940 79.233-18.196 79.445Q-18.452 79.656-18.452 79.988M-15.144 80.709L-15.144 79.646Q-15.144 79.622-15.116 79.595Q-15.089 79.568-15.065 79.568L-14.956 79.568Q-14.891 79.568-14.877 79.626Q-14.781 80.060-14.535 80.311Q-14.289 80.562-13.876 80.562Q-13.534 80.562-13.281 80.429Q-13.028 80.296-13.028 79.988Q-13.028 79.831-13.122 79.716Q-13.216 79.602-13.354 79.533Q-13.493 79.465-13.660 79.427L-14.241 79.328Q-14.597 79.260-14.870 79.039Q-15.144 78.819-15.144 78.477Q-15.144 78.228-15.033 78.053Q-14.921 77.879-14.735 77.780Q-14.549 77.681-14.334 77.638Q-14.118 77.595-13.876 77.595Q-13.462 77.595-13.182 77.777L-12.966 77.602Q-12.956 77.599-12.949 77.597Q-12.942 77.595-12.932 77.595L-12.881 77.595Q-12.854 77.595-12.830 77.619Q-12.806 77.643-12.806 77.671L-12.806 78.518Q-12.806 78.539-12.830 78.566Q-12.854 78.593-12.881 78.593L-12.994 78.593Q-13.021 78.593-13.047 78.568Q-13.072 78.542-13.072 78.518Q-13.072 78.282-13.178 78.118Q-13.284 77.954-13.467 77.872Q-13.650 77.790-13.882 77.790Q-14.211 77.790-14.467 77.893Q-14.723 77.995-14.723 78.272Q-14.723 78.467-14.540 78.576Q-14.357 78.686-14.128 78.727L-13.554 78.833Q-13.308 78.881-13.095 79.009Q-12.881 79.137-12.744 79.340Q-12.607 79.544-12.607 79.793Q-12.607 80.306-12.973 80.545Q-13.339 80.784-13.876 80.784Q-14.371 80.784-14.703 80.490L-14.969 80.764Q-14.990 80.784-15.017 80.784L-15.065 80.784Q-15.089 80.784-15.116 80.757Q-15.144 80.730-15.144 80.709\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg transform=\"rotate(-42.001 -104.026 -206.451)scale(.99999)\">\u003Cpath d=\"M-24.092 79.875L-24.092 77.978L-24.731 77.978L-24.731 77.756Q-24.413 77.756-24.196 77.546Q-23.979 77.336-23.879 77.026Q-23.778 76.717-23.778 76.409L-23.511 76.409L-23.511 77.698L-22.434 77.698L-22.434 77.978L-23.511 77.978L-23.511 79.862Q-23.511 80.138-23.407 80.337Q-23.303 80.535-23.043 80.535Q-22.886 80.535-22.780 80.431Q-22.674 80.326-22.624 80.173Q-22.575 80.019-22.575 79.862L-22.575 79.448L-22.308 79.448L-22.308 79.875Q-22.308 80.101-22.407 80.311Q-22.506 80.521-22.691 80.653Q-22.875 80.784-23.104 80.784Q-23.542 80.784-23.817 80.547Q-24.092 80.309-24.092 79.875M-19.881 80.716L-21.433 80.716L-21.433 80.436Q-21.207 80.436-21.059 80.402Q-20.910 80.367-20.910 80.227L-20.910 78.378Q-20.910 78.190-20.958 78.106Q-21.006 78.023-21.103 78.004Q-21.201 77.985-21.412 77.985L-21.412 77.705L-20.356 77.630L-20.356 80.227Q-20.356 80.367-20.225 80.402Q-20.093 80.436-19.881 80.436L-19.881 80.716M-21.153 76.409Q-21.153 76.238-21.030 76.119Q-20.907 75.999-20.736 75.999Q-20.568 75.999-20.445 76.119Q-20.322 76.238-20.322 76.409Q-20.322 76.584-20.445 76.707Q-20.568 76.830-20.736 76.830Q-20.907 76.830-21.030 76.707Q-21.153 76.584-21.153 76.409M-17.485 80.716L-19.222 80.716L-19.222 80.436Q-18.993 80.436-18.844 80.402Q-18.695 80.367-18.695 80.227L-18.695 78.378Q-18.695 78.108-18.803 78.047Q-18.911 77.985-19.222 77.985L-19.222 77.705L-18.193 77.630L-18.193 78.337Q-18.063 78.029-17.820 77.830Q-17.578 77.630-17.260 77.630Q-17.041 77.630-16.870 77.754Q-16.699 77.879-16.699 78.091Q-16.699 78.228-16.798 78.327Q-16.897 78.426-17.031 78.426Q-17.167 78.426-17.266 78.327Q-17.366 78.228-17.366 78.091Q-17.366 77.951-17.266 77.852Q-17.557 77.852-17.757 78.048Q-17.957 78.245-18.049 78.539Q-18.141 78.833-18.141 79.113L-18.141 80.227Q-18.141 80.436-17.485 80.436L-17.485 80.716M-16.156 79.181Q-16.156 78.860-16.031 78.571Q-15.906 78.282-15.681 78.059Q-15.455 77.835-15.159 77.715Q-14.864 77.595-14.546 77.595Q-14.218 77.595-13.956 77.695Q-13.695 77.794-13.519 77.976Q-13.343 78.159-13.249 78.417Q-13.155 78.675-13.155 79.007Q-13.155 79.099-13.237 79.120L-15.493 79.120L-15.493 79.181Q-15.493 79.769-15.209 80.152Q-14.925 80.535-14.358 80.535Q-14.037 80.535-13.768 80.342Q-13.500 80.149-13.411 79.834Q-13.404 79.793-13.329 79.779L-13.237 79.779Q-13.155 79.803-13.155 79.875Q-13.155 79.882-13.162 79.909Q-13.274 80.306-13.645 80.545Q-14.016 80.784-14.440 80.784Q-14.877 80.784-15.277 80.576Q-15.677 80.367-15.916 80Q-16.156 79.633-16.156 79.181M-15.486 78.911L-13.671 78.911Q-13.671 78.634-13.768 78.382Q-13.866 78.129-14.064 77.973Q-14.262 77.818-14.546 77.818Q-14.823 77.818-15.036 77.976Q-15.250 78.135-15.368 78.390Q-15.486 78.645-15.486 78.911M-12.567 79.205Q-12.567 78.867-12.427 78.576Q-12.287 78.286-12.042 78.072Q-11.798 77.859-11.494 77.744Q-11.189 77.630-10.865 77.630Q-10.595 77.630-10.331 77.729Q-10.068 77.828-9.877 78.006L-9.877 76.608Q-9.877 76.338-9.985 76.276Q-10.092 76.215-10.403 76.215L-10.403 75.934L-9.327 75.859L-9.327 80.043Q-9.327 80.231-9.272 80.314Q-9.217 80.398-9.116 80.417Q-9.016 80.436-8.800 80.436L-8.800 80.716L-9.908 80.784L-9.908 80.367Q-10.325 80.784-10.950 80.784Q-11.381 80.784-11.753 80.572Q-12.126 80.361-12.346 80Q-12.567 79.639-12.567 79.205M-10.892 80.562Q-10.683 80.562-10.497 80.490Q-10.311 80.419-10.157 80.282Q-10.003 80.145-9.908 79.967L-9.908 78.358Q-9.993 78.211-10.138 78.091Q-10.284 77.971-10.453 77.912Q-10.622 77.852-10.803 77.852Q-11.364 77.852-11.632 78.241Q-11.900 78.631-11.900 79.212Q-11.900 79.783-11.666 80.173Q-11.432 80.562-10.892 80.562\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-20.337 -110.455)\">\u003Cpath d=\"M-23.734 79.611L-23.734 77.829L-24.484 77.829Q-24.683 77.806-24.734 77.587L-24.734 77.501Q-24.683 77.290-24.484 77.267L-23.734 77.267L-23.734 76.517Q-23.683 76.310-23.484 76.282L-23.339 76.282Q-23.144 76.310-23.093 76.517L-23.093 77.267L-21.734 77.267Q-21.542 77.286-21.484 77.501L-21.484 77.587Q-21.538 77.806-21.734 77.829L-23.093 77.829L-23.093 79.579Q-23.093 80.196-22.519 80.196Q-22.269 80.196-22.105 80.011Q-21.941 79.825-21.941 79.579Q-21.941 79.486-21.869 79.415Q-21.796 79.345-21.695 79.333L-21.550 79.333Q-21.351 79.357-21.300 79.564L-21.300 79.611Q-21.300 79.935-21.484 80.198Q-21.667 80.462-21.960 80.609Q-22.253 80.755-22.574 80.755Q-23.085 80.755-23.410 80.445Q-23.734 80.134-23.734 79.611M-20.617 80.478L-20.617 80.388Q-20.574 80.181-20.367 80.157L-19.945 80.157L-19.945 76.388L-20.367 76.388Q-20.574 76.364-20.617 76.150L-20.617 76.060Q-20.574 75.853-20.367 75.829L-19.550 75.829Q-19.355 75.853-19.304 76.060L-19.304 77.611Q-18.843 77.228-18.245 77.228Q-17.898 77.228-17.660 77.368Q-17.421 77.509-17.306 77.767Q-17.191 78.025-17.191 78.380L-17.191 80.157L-16.765 80.157Q-16.558 80.181-16.519 80.388L-16.519 80.478Q-16.558 80.693-16.765 80.716L-18.160 80.716Q-18.355 80.693-18.406 80.478L-18.406 80.388Q-18.355 80.177-18.160 80.157L-17.831 80.157L-17.831 78.411Q-17.831 78.103-17.921 77.945Q-18.011 77.786-18.304 77.786Q-18.574 77.786-18.802 77.917Q-19.031 78.048-19.167 78.277Q-19.304 78.505-19.304 78.771L-19.304 80.157L-18.878 80.157Q-18.671 80.181-18.632 80.388L-18.632 80.478Q-18.671 80.693-18.878 80.716L-20.367 80.716Q-20.574 80.693-20.617 80.478M-12.929 79.228L-15.370 79.228Q-15.316 79.505-15.119 79.728Q-14.921 79.950-14.644 80.073Q-14.367 80.196-14.081 80.196Q-13.609 80.196-13.386 79.907Q-13.378 79.896-13.322 79.790Q-13.265 79.685-13.216 79.642Q-13.167 79.599-13.074 79.587L-12.929 79.587Q-12.738 79.607-12.679 79.821L-12.679 79.876Q-12.745 80.177-12.976 80.374Q-13.206 80.571-13.519 80.663Q-13.831 80.755-14.136 80.755Q-14.620 80.755-15.060 80.527Q-15.499 80.298-15.767 79.898Q-16.035 79.497-16.035 79.005L-16.035 78.946Q-16.035 78.478-15.788 78.075Q-15.542 77.673-15.134 77.439Q-14.726 77.204-14.257 77.204Q-13.753 77.204-13.400 77.427Q-13.046 77.650-12.863 78.038Q-12.679 78.427-12.679 78.931L-12.679 78.989Q-12.738 79.204-12.929 79.228M-15.363 78.677L-13.335 78.677Q-13.382 78.267-13.620 78.015Q-13.859 77.763-14.257 77.763Q-14.652 77.763-14.958 78.025Q-15.265 78.286-15.363 78.677\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-33.088 -92.814)\">\u003Cpath d=\"M-24.527 79.603Q-24.527 79.157-24.113 78.900Q-23.699 78.642-23.158 78.542Q-22.617 78.443-22.109 78.435Q-22.109 78.220-22.244 78.068Q-22.378 77.915-22.585 77.839Q-22.792 77.763-23.003 77.763Q-23.347 77.763-23.507 77.786L-23.507 77.845Q-23.507 78.013-23.626 78.128Q-23.745 78.243-23.910 78.243Q-24.085 78.243-24.201 78.120Q-24.316 77.997-24.316 77.829Q-24.316 77.423-23.935 77.314Q-23.554 77.204-22.995 77.204Q-22.726 77.204-22.458 77.282Q-22.191 77.361-21.966 77.511Q-21.742 77.661-21.605 77.882Q-21.468 78.103-21.468 78.380L-21.468 80.099Q-21.468 80.157-20.941 80.157Q-20.745 80.177-20.695 80.388L-20.695 80.478Q-20.745 80.693-20.941 80.716L-21.085 80.716Q-21.429 80.716-21.658 80.669Q-21.886 80.622-22.031 80.435Q-22.492 80.755-23.199 80.755Q-23.535 80.755-23.839 80.614Q-24.144 80.474-24.335 80.212Q-24.527 79.950-24.527 79.603M-23.886 79.611Q-23.886 79.884-23.644 80.040Q-23.402 80.196-23.117 80.196Q-22.898 80.196-22.665 80.138Q-22.433 80.079-22.271 79.941Q-22.109 79.802-22.109 79.579L-22.109 78.989Q-22.390 78.989-22.806 79.046Q-23.222 79.103-23.554 79.241Q-23.886 79.380-23.886 79.611M-20.617 80.478L-20.617 80.388Q-20.574 80.181-20.367 80.157L-19.945 80.157L-19.945 77.829L-20.367 77.829Q-20.574 77.806-20.617 77.587L-20.617 77.501Q-20.570 77.290-20.367 77.267L-19.550 77.267Q-19.355 77.290-19.304 77.501L-19.304 77.587L-19.312 77.611Q-19.085 77.431-18.812 77.329Q-18.538 77.228-18.245 77.228Q-17.898 77.228-17.660 77.368Q-17.421 77.509-17.306 77.767Q-17.191 78.025-17.191 78.380L-17.191 80.157L-16.765 80.157Q-16.558 80.181-16.519 80.388L-16.519 80.478Q-16.558 80.693-16.765 80.716L-18.160 80.716Q-18.355 80.693-18.406 80.478L-18.406 80.388Q-18.355 80.177-18.160 80.157L-17.831 80.157L-17.831 78.411Q-17.831 78.103-17.921 77.945Q-18.011 77.786-18.304 77.786Q-18.574 77.786-18.802 77.917Q-19.031 78.048-19.167 78.277Q-19.304 78.505-19.304 78.771L-19.304 80.157L-18.878 80.157Q-18.671 80.181-18.632 80.388L-18.632 80.478Q-18.671 80.693-18.878 80.716L-20.367 80.716Q-20.574 80.693-20.617 80.478M-15.824 80.478L-15.824 80.388Q-15.773 80.181-15.578 80.157L-14.538 80.157L-14.538 77.829L-15.511 77.829Q-15.710 77.806-15.761 77.587L-15.761 77.501Q-15.710 77.290-15.511 77.267L-14.144 77.267Q-13.949 77.286-13.898 77.501L-13.898 80.157L-12.984 80.157Q-12.788 80.181-12.738 80.388L-12.738 80.478Q-12.788 80.693-12.984 80.716L-15.578 80.716Q-15.773 80.693-15.824 80.478M-14.792 76.290L-14.792 76.236Q-14.792 76.064-14.656 75.943Q-14.519 75.821-14.343 75.821Q-14.171 75.821-14.035 75.943Q-13.898 76.064-13.898 76.236L-13.898 76.290Q-13.898 76.466-14.035 76.587Q-14.171 76.708-14.343 76.708Q-14.519 76.708-14.656 76.587Q-14.792 76.466-14.792 76.290M-12.249 80.478L-12.249 80.388Q-12.199 80.177-12.003 80.157L-11.804 80.157L-11.804 77.829L-12.003 77.829Q-12.210 77.806-12.249 77.587L-12.249 77.501Q-12.199 77.286-12.003 77.267L-11.523 77.267Q-11.331 77.290-11.273 77.501Q-10.953 77.228-10.538 77.228Q-10.347 77.228-10.179 77.337Q-10.011 77.446-9.937 77.618Q-9.761 77.427-9.538 77.327Q-9.316 77.228-9.074 77.228Q-8.656 77.228-8.501 77.570Q-8.347 77.911-8.347 78.380L-8.347 80.157L-8.148 80.157Q-7.937 80.181-7.898 80.388L-7.898 80.478Q-7.949 80.693-8.148 80.716L-8.964 80.716Q-9.160 80.693-9.210 80.478L-9.210 80.388Q-9.160 80.181-8.964 80.157L-8.874 80.157L-8.874 78.411Q-8.874 77.786-9.124 77.786Q-9.456 77.786-9.634 78.097Q-9.812 78.407-9.812 78.771L-9.812 80.157L-9.609 80.157Q-9.402 80.181-9.363 80.388L-9.363 80.478Q-9.413 80.693-9.609 80.716L-10.425 80.716Q-10.624 80.693-10.675 80.478L-10.675 80.388Q-10.624 80.181-10.425 80.157L-10.339 80.157L-10.339 78.411Q-10.339 77.786-10.585 77.786Q-10.917 77.786-11.095 78.099Q-11.273 78.411-11.273 78.771L-11.273 80.157L-11.074 80.157Q-10.867 80.181-10.828 80.388L-10.828 80.478Q-10.878 80.696-11.074 80.716L-12.003 80.716Q-12.210 80.693-12.249 80.478M-7.542 79.603Q-7.542 79.157-7.128 78.900Q-6.714 78.642-6.173 78.542Q-5.632 78.443-5.124 78.435Q-5.124 78.220-5.259 78.068Q-5.394 77.915-5.601 77.839Q-5.808 77.763-6.019 77.763Q-6.363 77.763-6.523 77.786L-6.523 77.845Q-6.523 78.013-6.642 78.128Q-6.761 78.243-6.925 78.243Q-7.101 78.243-7.216 78.120Q-7.331 77.997-7.331 77.829Q-7.331 77.423-6.951 77.314Q-6.570 77.204-6.011 77.204Q-5.742 77.204-5.474 77.282Q-5.206 77.361-4.982 77.511Q-4.757 77.661-4.620 77.882Q-4.484 78.103-4.484 78.380L-4.484 80.099Q-4.484 80.157-3.956 80.157Q-3.761 80.177-3.710 80.388L-3.710 80.478Q-3.761 80.693-3.956 80.716L-4.101 80.716Q-4.445 80.716-4.673 80.669Q-4.902 80.622-5.046 80.435Q-5.507 80.755-6.214 80.755Q-6.550 80.755-6.855 80.614Q-7.160 80.474-7.351 80.212Q-7.542 79.950-7.542 79.603M-6.902 79.611Q-6.902 79.884-6.660 80.040Q-6.417 80.196-6.132 80.196Q-5.913 80.196-5.681 80.138Q-5.449 80.079-5.287 79.941Q-5.124 79.802-5.124 79.579L-5.124 78.989Q-5.406 78.989-5.822 79.046Q-6.238 79.103-6.570 79.241Q-6.902 79.380-6.902 79.611M-3.253 80.478L-3.253 80.388Q-3.203 80.181-3.007 80.157L-1.902 80.157L-1.902 76.388L-3.007 76.388Q-3.203 76.364-3.253 76.150L-3.253 76.060Q-3.203 75.853-3.007 75.829L-1.511 75.829Q-1.320 75.853-1.261 76.060L-1.261 80.157L-0.160 80.157Q0.040 80.181 0.090 80.388L0.090 80.478Q0.040 80.693-0.160 80.716L-3.007 80.716Q-3.203 80.693-3.253 80.478\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-37.338 -75.174)\">\u003Cpath d=\"M-24.374 78.989Q-24.374 78.509-24.130 78.095Q-23.886 77.681-23.470 77.443Q-23.054 77.204-22.574 77.204Q-22.019 77.204-21.640 77.314Q-21.261 77.423-21.261 77.829Q-21.261 77.997-21.374 78.120Q-21.488 78.243-21.660 78.243Q-21.831 78.243-21.951 78.128Q-22.070 78.013-22.070 77.845L-22.070 77.786Q-22.230 77.763-22.566 77.763Q-22.894 77.763-23.162 77.933Q-23.429 78.103-23.581 78.386Q-23.734 78.669-23.734 78.989Q-23.734 79.310-23.562 79.591Q-23.390 79.872-23.105 80.034Q-22.820 80.196-22.492 80.196Q-22.179 80.196-22.052 80.093Q-21.925 79.989-21.808 79.800Q-21.691 79.611-21.574 79.595L-21.406 79.595Q-21.300 79.607-21.236 79.675Q-21.171 79.743-21.171 79.845Q-21.171 79.892-21.191 79.931Q-21.300 80.224-21.503 80.403Q-21.706 80.583-21.982 80.669Q-22.257 80.755-22.574 80.755Q-23.058 80.755-23.474 80.517Q-23.890 80.278-24.132 79.876Q-24.374 79.474-24.374 78.989M-20.464 80.478L-20.464 80.388Q-20.406 80.181-20.214 80.157L-19.503 80.157L-19.503 77.829L-20.214 77.829Q-20.410 77.806-20.464 77.587L-20.464 77.501Q-20.406 77.290-20.214 77.267L-19.113 77.267Q-18.913 77.286-18.863 77.501L-18.863 77.829Q-18.601 77.544-18.245 77.386Q-17.890 77.228-17.503 77.228Q-17.210 77.228-16.976 77.362Q-16.742 77.497-16.742 77.763Q-16.742 77.931-16.851 78.048Q-16.960 78.165-17.128 78.165Q-17.281 78.165-17.396 78.054Q-17.511 77.943-17.511 77.786Q-17.886 77.786-18.201 77.987Q-18.515 78.189-18.689 78.523Q-18.863 78.857-18.863 79.236L-18.863 80.157L-17.917 80.157Q-17.710 80.181-17.671 80.388L-17.671 80.478Q-17.710 80.693-17.917 80.716L-20.214 80.716Q-20.406 80.693-20.464 80.478M-14.320 80.755Q-14.792 80.755-15.177 80.511Q-15.562 80.267-15.785 79.857Q-16.007 79.446-16.007 78.989Q-16.007 78.646-15.882 78.323Q-15.757 78.001-15.527 77.747Q-15.296 77.493-14.990 77.349Q-14.683 77.204-14.320 77.204Q-13.956 77.204-13.644 77.351Q-13.331 77.497-13.109 77.743Q-12.886 77.989-12.759 78.310Q-12.632 78.630-12.632 78.989Q-12.632 79.446-12.857 79.859Q-13.081 80.271-13.466 80.513Q-13.851 80.755-14.320 80.755M-14.320 80.196Q-13.855 80.196-13.564 79.802Q-13.273 79.407-13.273 78.923Q-13.273 78.630-13.408 78.362Q-13.542 78.095-13.783 77.929Q-14.023 77.763-14.320 77.763Q-14.624 77.763-14.863 77.929Q-15.101 78.095-15.236 78.362Q-15.370 78.630-15.370 78.923Q-15.370 79.403-15.078 79.800Q-14.785 80.196-14.320 80.196M-11.609 80.517L-11.609 79.603Q-11.581 79.396-11.370 79.372L-11.203 79.372Q-11.038 79.396-10.980 79.556Q-10.777 80.196-10.050 80.196Q-9.843 80.196-9.615 80.161Q-9.386 80.126-9.218 80.011Q-9.050 79.896-9.050 79.693Q-9.050 79.482-9.273 79.368Q-9.495 79.255-9.769 79.212L-10.468 79.099Q-11.609 78.888-11.609 78.165Q-11.609 77.876-11.464 77.687Q-11.320 77.497-11.079 77.390Q-10.839 77.282-10.583 77.243Q-10.328 77.204-10.050 77.204Q-9.800 77.204-9.607 77.234Q-9.413 77.263-9.249 77.341Q-9.171 77.224-9.042 77.204L-8.964 77.204Q-8.867 77.216-8.804 77.278Q-8.742 77.341-8.730 77.435L-8.730 78.142Q-8.742 78.236-8.804 78.302Q-8.867 78.368-8.964 78.380L-9.132 78.380Q-9.226 78.368-9.292 78.302Q-9.359 78.236-9.370 78.142Q-9.370 77.763-10.066 77.763Q-10.413 77.763-10.732 77.845Q-11.050 77.927-11.050 78.173Q-11.050 78.439-10.378 78.548L-9.675 78.669Q-9.191 78.751-8.841 78.999Q-8.492 79.247-8.492 79.693Q-8.492 80.083-8.728 80.325Q-8.964 80.568-9.314 80.661Q-9.663 80.755-10.050 80.755Q-10.628 80.755-11.027 80.501Q-11.097 80.626-11.146 80.683Q-11.195 80.739-11.300 80.755L-11.370 80.755Q-11.585 80.732-11.609 80.517M-7.363 80.517L-7.363 79.603Q-7.335 79.396-7.124 79.372L-6.956 79.372Q-6.792 79.396-6.734 79.556Q-6.531 80.196-5.804 80.196Q-5.597 80.196-5.369 80.161Q-5.140 80.126-4.972 80.011Q-4.804 79.896-4.804 79.693Q-4.804 79.482-5.027 79.368Q-5.249 79.255-5.523 79.212L-6.222 79.099Q-7.363 78.888-7.363 78.165Q-7.363 77.876-7.218 77.687Q-7.074 77.497-6.833 77.390Q-6.593 77.282-6.337 77.243Q-6.081 77.204-5.804 77.204Q-5.554 77.204-5.361 77.234Q-5.167 77.263-5.003 77.341Q-4.925 77.224-4.796 77.204L-4.718 77.204Q-4.620 77.216-4.558 77.278Q-4.495 77.341-4.484 77.435L-4.484 78.142Q-4.495 78.236-4.558 78.302Q-4.620 78.368-4.718 78.380L-4.886 78.380Q-4.980 78.368-5.046 78.302Q-5.113 78.236-5.124 78.142Q-5.124 77.763-5.820 77.763Q-6.167 77.763-6.486 77.845Q-6.804 77.927-6.804 78.173Q-6.804 78.439-6.132 78.548L-5.429 78.669Q-4.945 78.751-4.595 78.999Q-4.245 79.247-4.245 79.693Q-4.245 80.083-4.482 80.325Q-4.718 80.568-5.068 80.661Q-5.417 80.755-5.804 80.755Q-6.382 80.755-6.781 80.501Q-6.851 80.626-6.900 80.683Q-6.949 80.739-7.054 80.755L-7.124 80.755Q-7.339 80.732-7.363 80.517M-0.191 79.228L-2.632 79.228Q-2.578 79.505-2.380 79.728Q-2.183 79.950-1.906 80.073Q-1.628 80.196-1.343 80.196Q-0.870 80.196-0.648 79.907Q-0.640 79.896-0.583 79.790Q-0.527 79.685-0.478 79.642Q-0.429 79.599-0.335 79.587L-0.191 79.587Q0.001 79.607 0.059 79.821L0.059 79.876Q-0.007 80.177-0.238 80.374Q-0.468 80.571-0.781 80.663Q-1.093 80.755-1.398 80.755Q-1.882 80.755-2.322 80.527Q-2.761 80.298-3.029 79.898Q-3.296 79.497-3.296 79.005L-3.296 78.946Q-3.296 78.478-3.050 78.075Q-2.804 77.673-2.396 77.439Q-1.988 77.204-1.519 77.204Q-1.015 77.204-0.662 77.427Q-0.308 77.650-0.124 78.038Q0.059 78.427 0.059 78.931L0.059 78.989Q0.001 79.204-0.191 79.228M-2.624 78.677L-0.597 78.677Q-0.644 78.267-0.882 78.015Q-1.120 77.763-1.519 77.763Q-1.913 77.763-2.220 78.025Q-2.527 78.286-2.624 78.677M2.407 80.755Q1.942 80.755 1.577 80.505Q1.212 80.255 1.006 79.851Q0.801 79.446 0.801 78.989Q0.801 78.646 0.926 78.325Q1.051 78.005 1.284 77.755Q1.516 77.505 1.821 77.366Q2.126 77.228 2.481 77.228Q2.993 77.228 3.399 77.548L3.399 76.388L2.977 76.388Q2.766 76.364 2.727 76.150L2.727 76.060Q2.766 75.853 2.977 75.829L3.790 75.829Q3.989 75.853 4.040 76.060L4.040 80.157L4.465 80.157Q4.672 80.181 4.712 80.388L4.712 80.478Q4.672 80.693 4.465 80.716L3.649 80.716Q3.450 80.693 3.399 80.478L3.399 80.349Q3.204 80.540 2.942 80.648Q2.680 80.755 2.407 80.755M2.446 80.196Q2.805 80.196 3.057 79.927Q3.309 79.657 3.399 79.282L3.399 78.450Q3.340 78.263 3.213 78.112Q3.087 77.962 2.909 77.874Q2.731 77.786 2.536 77.786Q2.227 77.786 1.977 77.956Q1.727 78.126 1.583 78.411Q1.438 78.696 1.438 78.997Q1.438 79.446 1.723 79.821Q2.008 80.196 2.446 80.196\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-37.338 -57.533)\">\u003Cpath d=\"M-24.191 80.478L-24.191 76.388L-24.613 76.388Q-24.820 76.364-24.863 76.150L-24.863 76.060Q-24.820 75.853-24.613 75.829L-23.796 75.829Q-23.601 75.853-23.550 76.060L-23.550 77.571Q-23.339 77.403-23.076 77.316Q-22.812 77.228-22.542 77.228Q-22.203 77.228-21.906 77.372Q-21.609 77.517-21.394 77.769Q-21.179 78.021-21.064 78.333Q-20.949 78.646-20.949 78.989Q-20.949 79.454-21.175 79.864Q-21.402 80.275-21.788 80.515Q-22.175 80.755-22.644 80.755Q-23.163 80.755-23.550 80.388L-23.550 80.478Q-23.601 80.696-23.796 80.716L-23.941 80.716Q-24.132 80.693-24.191 80.478M-22.687 80.196Q-22.378 80.196-22.128 80.027Q-21.878 79.857-21.734 79.575Q-21.589 79.294-21.589 78.989Q-21.589 78.700-21.714 78.421Q-21.839 78.142-22.072 77.964Q-22.304 77.786-22.605 77.786Q-22.925 77.786-23.187 77.972Q-23.449 78.157-23.550 78.458L-23.550 79.310Q-23.460 79.677-23.240 79.937Q-23.019 80.196-22.687 80.196M-17.175 79.228L-19.617 79.228Q-19.562 79.505-19.365 79.728Q-19.167 79.950-18.890 80.073Q-18.613 80.196-18.328 80.196Q-17.855 80.196-17.632 79.907Q-17.624 79.896-17.568 79.790Q-17.511 79.685-17.462 79.642Q-17.413 79.599-17.320 79.587L-17.175 79.587Q-16.984 79.607-16.925 79.821L-16.925 79.876Q-16.992 80.177-17.222 80.374Q-17.453 80.571-17.765 80.663Q-18.078 80.755-18.382 80.755Q-18.867 80.755-19.306 80.527Q-19.745 80.298-20.013 79.898Q-20.281 79.497-20.281 79.005L-20.281 78.946Q-20.281 78.478-20.035 78.075Q-19.788 77.673-19.380 77.439Q-18.972 77.204-18.503 77.204Q-17.999 77.204-17.646 77.427Q-17.292 77.650-17.109 78.038Q-16.925 78.427-16.925 78.931L-16.925 78.989Q-16.984 79.204-17.175 79.228M-19.609 78.677L-17.581 78.677Q-17.628 78.267-17.867 78.015Q-18.105 77.763-18.503 77.763Q-18.898 77.763-19.204 78.025Q-19.511 78.286-19.609 78.677M-15.882 78.989Q-15.882 78.509-15.638 78.095Q-15.394 77.681-14.978 77.443Q-14.562 77.204-14.081 77.204Q-13.527 77.204-13.148 77.314Q-12.769 77.423-12.769 77.829Q-12.769 77.997-12.882 78.120Q-12.995 78.243-13.167 78.243Q-13.339 78.243-13.458 78.128Q-13.578 78.013-13.578 77.845L-13.578 77.786Q-13.738 77.763-14.074 77.763Q-14.402 77.763-14.669 77.933Q-14.937 78.103-15.089 78.386Q-15.242 78.669-15.242 78.989Q-15.242 79.310-15.070 79.591Q-14.898 79.872-14.613 80.034Q-14.328 80.196-13.999 80.196Q-13.687 80.196-13.560 80.093Q-13.433 79.989-13.316 79.800Q-13.199 79.611-13.081 79.595L-12.913 79.595Q-12.808 79.607-12.744 79.675Q-12.679 79.743-12.679 79.845Q-12.679 79.892-12.699 79.931Q-12.808 80.224-13.011 80.403Q-13.214 80.583-13.490 80.669Q-13.765 80.755-14.081 80.755Q-14.566 80.755-14.982 80.517Q-15.398 80.278-15.640 79.876Q-15.882 79.474-15.882 78.989M-11.788 79.603Q-11.788 79.157-11.374 78.900Q-10.960 78.642-10.419 78.542Q-9.878 78.443-9.370 78.435Q-9.370 78.220-9.505 78.068Q-9.640 77.915-9.847 77.839Q-10.054 77.763-10.265 77.763Q-10.609 77.763-10.769 77.786L-10.769 77.845Q-10.769 78.013-10.888 78.128Q-11.007 78.243-11.171 78.243Q-11.347 78.243-11.462 78.120Q-11.578 77.997-11.578 77.829Q-11.578 77.423-11.197 77.314Q-10.816 77.204-10.257 77.204Q-9.988 77.204-9.720 77.282Q-9.453 77.361-9.228 77.511Q-9.003 77.661-8.867 77.882Q-8.730 78.103-8.730 78.380L-8.730 80.099Q-8.730 80.157-8.203 80.157Q-8.007 80.177-7.956 80.388L-7.956 80.478Q-8.007 80.693-8.203 80.716L-8.347 80.716Q-8.691 80.716-8.919 80.669Q-9.148 80.622-9.292 80.435Q-9.753 80.755-10.460 80.755Q-10.796 80.755-11.101 80.614Q-11.406 80.474-11.597 80.212Q-11.788 79.950-11.788 79.603M-11.148 79.611Q-11.148 79.884-10.906 80.040Q-10.663 80.196-10.378 80.196Q-10.160 80.196-9.927 80.138Q-9.695 80.079-9.533 79.941Q-9.370 79.802-9.370 79.579L-9.370 78.989Q-9.652 78.989-10.068 79.046Q-10.484 79.103-10.816 79.241Q-11.148 79.380-11.148 79.611M-7.206 79.861L-7.206 77.829L-7.628 77.829Q-7.835 77.806-7.878 77.587L-7.878 77.501Q-7.831 77.290-7.628 77.267L-6.812 77.267Q-6.617 77.290-6.566 77.501L-6.566 79.829Q-6.566 80.064-6.396 80.130Q-6.226 80.196-5.941 80.196Q-5.734 80.196-5.538 80.120Q-5.343 80.044-5.218 79.894Q-5.093 79.743-5.093 79.532L-5.093 77.829L-5.515 77.829Q-5.726 77.806-5.765 77.587L-5.765 77.501Q-5.726 77.290-5.515 77.267L-4.703 77.267Q-4.503 77.290-4.453 77.501L-4.453 80.157L-4.027 80.157Q-3.820 80.181-3.781 80.388L-3.781 80.478Q-3.820 80.693-4.027 80.716L-4.843 80.716Q-5.042 80.693-5.093 80.493Q-5.495 80.755-6.003 80.755Q-6.238 80.755-6.453 80.714Q-6.667 80.673-6.833 80.571Q-6.999 80.470-7.103 80.292Q-7.206 80.114-7.206 79.861M-3.117 80.517L-3.117 79.603Q-3.089 79.396-2.878 79.372L-2.710 79.372Q-2.546 79.396-2.488 79.556Q-2.285 80.196-1.558 80.196Q-1.351 80.196-1.122 80.161Q-0.894 80.126-0.726 80.011Q-0.558 79.896-0.558 79.693Q-0.558 79.482-0.781 79.368Q-1.003 79.255-1.277 79.212L-1.976 79.099Q-3.117 78.888-3.117 78.165Q-3.117 77.876-2.972 77.687Q-2.828 77.497-2.587 77.390Q-2.347 77.282-2.091 77.243Q-1.835 77.204-1.558 77.204Q-1.308 77.204-1.115 77.234Q-0.921 77.263-0.757 77.341Q-0.679 77.224-0.550 77.204L-0.472 77.204Q-0.374 77.216-0.312 77.278Q-0.249 77.341-0.238 77.435L-0.238 78.142Q-0.249 78.236-0.312 78.302Q-0.374 78.368-0.472 78.380L-0.640 78.380Q-0.734 78.368-0.800 78.302Q-0.867 78.236-0.878 78.142Q-0.878 77.763-1.574 77.763Q-1.921 77.763-2.240 77.845Q-2.558 77.927-2.558 78.173Q-2.558 78.439-1.886 78.548L-1.183 78.669Q-0.699 78.751-0.349 78.999Q0.001 79.247 0.001 79.693Q0.001 80.083-0.236 80.325Q-0.472 80.568-0.822 80.661Q-1.171 80.755-1.558 80.755Q-2.136 80.755-2.535 80.501Q-2.605 80.626-2.654 80.683Q-2.703 80.739-2.808 80.755L-2.878 80.755Q-3.093 80.732-3.117 80.517M4.055 79.228L1.614 79.228Q1.669 79.505 1.866 79.728Q2.063 79.950 2.340 80.073Q2.618 80.196 2.903 80.196Q3.376 80.196 3.598 79.907Q3.606 79.896 3.663 79.790Q3.719 79.685 3.768 79.642Q3.817 79.599 3.911 79.587L4.055 79.587Q4.247 79.607 4.305 79.821L4.305 79.876Q4.239 80.177 4.008 80.374Q3.778 80.571 3.465 80.663Q3.153 80.755 2.848 80.755Q2.364 80.755 1.924 80.527Q1.485 80.298 1.217 79.898Q0.950 79.497 0.950 79.005L0.950 78.946Q0.950 78.478 1.196 78.075Q1.442 77.673 1.850 77.439Q2.258 77.204 2.727 77.204Q3.231 77.204 3.585 77.427Q3.938 77.650 4.122 78.038Q4.305 78.427 4.305 78.931L4.305 78.989Q4.247 79.204 4.055 79.228M1.622 78.677L3.649 78.677Q3.602 78.267 3.364 78.015Q3.126 77.763 2.727 77.763Q2.333 77.763 2.026 78.025Q1.719 78.286 1.622 78.677\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-16.087 -39.893)\">\u003Cpath d=\"M-24.316 80.478L-24.316 80.388Q-24.265 80.181-24.070 80.157L-23.031 80.157L-23.031 77.829L-24.003 77.829Q-24.203 77.806-24.253 77.587L-24.253 77.501Q-24.203 77.290-24.003 77.267L-22.636 77.267Q-22.441 77.286-22.390 77.501L-22.390 80.157L-21.476 80.157Q-21.281 80.181-21.230 80.388L-21.230 80.478Q-21.281 80.693-21.476 80.716L-24.070 80.716Q-24.265 80.693-24.316 80.478M-23.285 76.290L-23.285 76.236Q-23.285 76.064-23.148 75.943Q-23.011 75.821-22.835 75.821Q-22.663 75.821-22.527 75.943Q-22.390 76.064-22.390 76.236L-22.390 76.290Q-22.390 76.466-22.527 76.587Q-22.663 76.708-22.835 76.708Q-23.011 76.708-23.148 76.587Q-23.285 76.466-23.285 76.290M-19.488 79.611L-19.488 77.829L-20.238 77.829Q-20.437 77.806-20.488 77.587L-20.488 77.501Q-20.437 77.290-20.238 77.267L-19.488 77.267L-19.488 76.517Q-19.437 76.310-19.238 76.282L-19.093 76.282Q-18.898 76.310-18.847 76.517L-18.847 77.267L-17.488 77.267Q-17.296 77.286-17.238 77.501L-17.238 77.587Q-17.292 77.806-17.488 77.829L-18.847 77.829L-18.847 79.579Q-18.847 80.196-18.273 80.196Q-18.023 80.196-17.859 80.011Q-17.695 79.825-17.695 79.579Q-17.695 79.486-17.622 79.415Q-17.550 79.345-17.449 79.333L-17.304 79.333Q-17.105 79.357-17.054 79.564L-17.054 79.611Q-17.054 79.935-17.238 80.198Q-17.421 80.462-17.714 80.609Q-18.007 80.755-18.328 80.755Q-18.839 80.755-19.163 80.445Q-19.488 80.134-19.488 79.611\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-20.337 -22.974)\">\u003Cpath d=\"M-23.925 80.493L-24.390 77.829L-24.550 77.829Q-24.757 77.806-24.796 77.587L-24.796 77.501Q-24.757 77.290-24.550 77.267L-23.382 77.267Q-23.171 77.290-23.132 77.501L-23.132 77.587Q-23.171 77.806-23.382 77.829L-23.855 77.829Q-23.742 78.474-23.663 78.919Q-23.585 79.364-23.535 79.683Q-23.484 80.001-23.484 80.075Q-23.476 79.903-23.191 78.907Q-23.152 78.794-23.054 78.720Q-22.956 78.646-22.835 78.646L-22.757 78.646Q-22.636 78.646-22.538 78.720Q-22.441 78.794-22.406 78.907Q-22.296 79.290-22.214 79.614Q-22.132 79.939-22.132 80.075Q-22.120 79.786-21.773 77.829L-22.245 77.829Q-22.453 77.806-22.492 77.587L-22.492 77.501Q-22.453 77.290-22.245 77.267L-21.070 77.267Q-20.878 77.290-20.820 77.501L-20.820 77.587Q-20.874 77.806-21.070 77.829L-21.238 77.829L-21.703 80.493Q-21.726 80.611-21.814 80.683Q-21.902 80.755-22.023 80.755L-22.163 80.755Q-22.285 80.755-22.380 80.683Q-22.476 80.611-22.519 80.493Q-22.566 80.314-22.638 80.058Q-22.710 79.802-22.753 79.593Q-22.796 79.384-22.796 79.282Q-22.804 79.564-23.085 80.493Q-23.113 80.603-23.210 80.679Q-23.308 80.755-23.429 80.755L-23.605 80.755Q-23.726 80.755-23.814 80.683Q-23.902 80.611-23.925 80.493M-20.281 79.603Q-20.281 79.157-19.867 78.900Q-19.453 78.642-18.912 78.542Q-18.370 78.443-17.863 78.435Q-17.863 78.220-17.997 78.068Q-18.132 77.915-18.339 77.839Q-18.546 77.763-18.757 77.763Q-19.101 77.763-19.261 77.786L-19.261 77.845Q-19.261 78.013-19.380 78.128Q-19.499 78.243-19.663 78.243Q-19.839 78.243-19.954 78.120Q-20.070 77.997-20.070 77.829Q-20.070 77.423-19.689 77.314Q-19.308 77.204-18.749 77.204Q-18.480 77.204-18.212 77.282Q-17.945 77.361-17.720 77.511Q-17.495 77.661-17.359 77.882Q-17.222 78.103-17.222 78.380L-17.222 80.099Q-17.222 80.157-16.695 80.157Q-16.499 80.177-16.449 80.388L-16.449 80.478Q-16.499 80.693-16.695 80.716L-16.839 80.716Q-17.183 80.716-17.412 80.669Q-17.640 80.622-17.785 80.435Q-18.245 80.755-18.953 80.755Q-19.288 80.755-19.593 80.614Q-19.898 80.474-20.089 80.212Q-20.281 79.950-20.281 79.603M-19.640 79.611Q-19.640 79.884-19.398 80.040Q-19.156 80.196-18.870 80.196Q-18.652 80.196-18.419 80.138Q-18.187 80.079-18.025 79.941Q-17.863 79.802-17.863 79.579L-17.863 78.989Q-18.144 78.989-18.560 79.046Q-18.976 79.103-19.308 79.241Q-19.640 79.380-19.640 79.611M-15.855 80.517L-15.855 79.603Q-15.828 79.396-15.617 79.372L-15.449 79.372Q-15.285 79.396-15.226 79.556Q-15.023 80.196-14.296 80.196Q-14.089 80.196-13.861 80.161Q-13.632 80.126-13.464 80.011Q-13.296 79.896-13.296 79.693Q-13.296 79.482-13.519 79.368Q-13.742 79.255-14.015 79.212L-14.714 79.099Q-15.855 78.888-15.855 78.165Q-15.855 77.876-15.710 77.687Q-15.566 77.497-15.326 77.390Q-15.085 77.282-14.829 77.243Q-14.574 77.204-14.296 77.204Q-14.046 77.204-13.853 77.234Q-13.660 77.263-13.495 77.341Q-13.417 77.224-13.288 77.204L-13.210 77.204Q-13.113 77.216-13.050 77.278Q-12.988 77.341-12.976 77.435L-12.976 78.142Q-12.988 78.236-13.050 78.302Q-13.113 78.368-13.210 78.380L-13.378 78.380Q-13.472 78.368-13.538 78.302Q-13.605 78.236-13.617 78.142Q-13.617 77.763-14.312 77.763Q-14.660 77.763-14.978 77.845Q-15.296 77.927-15.296 78.173Q-15.296 78.439-14.624 78.548L-13.921 78.669Q-13.437 78.751-13.087 78.999Q-12.738 79.247-12.738 79.693Q-12.738 80.083-12.974 80.325Q-13.210 80.568-13.560 80.661Q-13.910 80.755-14.296 80.755Q-14.874 80.755-15.273 80.501Q-15.343 80.626-15.392 80.683Q-15.441 80.739-15.546 80.755L-15.617 80.755Q-15.831 80.732-15.855 80.517\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-28.838 -4.612)\">\u003Cpath d=\"M-23.734 79.611L-23.734 77.829L-24.484 77.829Q-24.683 77.806-24.734 77.587L-24.734 77.501Q-24.683 77.290-24.484 77.267L-23.734 77.267L-23.734 76.517Q-23.683 76.310-23.484 76.282L-23.339 76.282Q-23.144 76.310-23.093 76.517L-23.093 77.267L-21.734 77.267Q-21.542 77.286-21.484 77.501L-21.484 77.587Q-21.538 77.806-21.734 77.829L-23.093 77.829L-23.093 79.579Q-23.093 80.196-22.519 80.196Q-22.269 80.196-22.105 80.011Q-21.941 79.825-21.941 79.579Q-21.941 79.486-21.869 79.415Q-21.796 79.345-21.695 79.333L-21.550 79.333Q-21.351 79.357-21.300 79.564L-21.300 79.611Q-21.300 79.935-21.484 80.198Q-21.667 80.462-21.960 80.609Q-22.253 80.755-22.574 80.755Q-23.085 80.755-23.410 80.445Q-23.734 80.134-23.734 79.611M-20.070 80.478L-20.070 80.388Q-20.019 80.181-19.824 80.157L-18.785 80.157L-18.785 77.829L-19.757 77.829Q-19.956 77.806-20.007 77.587L-20.007 77.501Q-19.956 77.290-19.757 77.267L-18.390 77.267Q-18.195 77.286-18.144 77.501L-18.144 80.157L-17.230 80.157Q-17.035 80.181-16.984 80.388L-16.984 80.478Q-17.035 80.693-17.230 80.716L-19.824 80.716Q-20.019 80.693-20.070 80.478M-19.038 76.290L-19.038 76.236Q-19.038 76.064-18.902 75.943Q-18.765 75.821-18.589 75.821Q-18.417 75.821-18.281 75.943Q-18.144 76.064-18.144 76.236L-18.144 76.290Q-18.144 76.466-18.281 76.587Q-18.417 76.708-18.589 76.708Q-18.765 76.708-18.902 76.587Q-19.038 76.466-19.038 76.290M-16.218 80.478L-16.218 80.388Q-16.160 80.181-15.968 80.157L-15.257 80.157L-15.257 77.829L-15.968 77.829Q-16.163 77.806-16.218 77.587L-16.218 77.501Q-16.160 77.290-15.968 77.267L-14.867 77.267Q-14.667 77.286-14.617 77.501L-14.617 77.829Q-14.355 77.544-13.999 77.386Q-13.644 77.228-13.257 77.228Q-12.964 77.228-12.730 77.362Q-12.495 77.497-12.495 77.763Q-12.495 77.931-12.605 78.048Q-12.714 78.165-12.882 78.165Q-13.035 78.165-13.150 78.054Q-13.265 77.943-13.265 77.786Q-13.640 77.786-13.954 77.987Q-14.269 78.189-14.443 78.523Q-14.617 78.857-14.617 79.236L-14.617 80.157L-13.671 80.157Q-13.464 80.181-13.425 80.388L-13.425 80.478Q-13.464 80.693-13.671 80.716L-15.968 80.716Q-16.160 80.693-16.218 80.478M-8.683 79.228L-11.124 79.228Q-11.070 79.505-10.872 79.728Q-10.675 79.950-10.398 80.073Q-10.120 80.196-9.835 80.196Q-9.363 80.196-9.140 79.907Q-9.132 79.896-9.076 79.790Q-9.019 79.685-8.970 79.642Q-8.921 79.599-8.828 79.587L-8.683 79.587Q-8.492 79.607-8.433 79.821L-8.433 79.876Q-8.499 80.177-8.730 80.374Q-8.960 80.571-9.273 80.663Q-9.585 80.755-9.890 80.755Q-10.374 80.755-10.814 80.527Q-11.253 80.298-11.521 79.898Q-11.788 79.497-11.788 79.005L-11.788 78.946Q-11.788 78.478-11.542 78.075Q-11.296 77.673-10.888 77.439Q-10.480 77.204-10.011 77.204Q-9.507 77.204-9.154 77.427Q-8.800 77.650-8.617 78.038Q-8.433 78.427-8.433 78.931L-8.433 78.989Q-8.492 79.204-8.683 79.228M-11.117 78.677L-9.089 78.677Q-9.136 78.267-9.374 78.015Q-9.613 77.763-10.011 77.763Q-10.406 77.763-10.712 78.025Q-11.019 78.286-11.117 78.677M-6.085 80.755Q-6.550 80.755-6.915 80.505Q-7.281 80.255-7.486 79.851Q-7.691 79.446-7.691 78.989Q-7.691 78.646-7.566 78.325Q-7.441 78.005-7.208 77.755Q-6.976 77.505-6.671 77.366Q-6.367 77.228-6.011 77.228Q-5.499 77.228-5.093 77.548L-5.093 76.388L-5.515 76.388Q-5.726 76.364-5.765 76.150L-5.765 76.060Q-5.726 75.853-5.515 75.829L-4.703 75.829Q-4.503 75.853-4.453 76.060L-4.453 80.157L-4.027 80.157Q-3.820 80.181-3.781 80.388L-3.781 80.478Q-3.820 80.693-4.027 80.716L-4.843 80.716Q-5.042 80.693-5.093 80.478L-5.093 80.349Q-5.288 80.540-5.550 80.648Q-5.812 80.755-6.085 80.755M-6.046 80.196Q-5.687 80.196-5.435 79.927Q-5.183 79.657-5.093 79.282L-5.093 78.450Q-5.152 78.263-5.279 78.112Q-5.406 77.962-5.583 77.874Q-5.761 77.786-5.956 77.786Q-6.265 77.786-6.515 77.956Q-6.765 78.126-6.910 78.411Q-7.054 78.696-7.054 78.997Q-7.054 79.446-6.769 79.821Q-6.484 80.196-6.046 80.196\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cpath fill=\"var(--tk-accent)\" d=\"M-23.04-27.773v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M3.99-27.773v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M31.02-27.773v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M58.05-27.773v-13.76h21.083v13.76ZM85.08-27.773v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M112.11-27.773v-13.76h21.083v13.76ZM139.14-27.773v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M-23.04-10.132v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-accent)\" d=\"M3.99-10.132v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M31.02-10.132v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M58.05-10.132v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M85.08-10.132v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M112.11-10.132v-13.76h21.083v13.76ZM139.14-10.132v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M-23.04 7.508v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M3.99 7.508v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M31.02 7.508v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M58.05 7.508v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M85.08 7.508v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M112.11 7.508v-13.76h21.083v13.76ZM139.14 7.508v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M-23.04 25.149v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M3.99 25.149v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M31.02 25.149v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M58.05 25.149v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M85.08 25.149v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M112.11 25.149v-13.76h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M139.14 25.149v-13.76h21.083v13.76ZM-23.04 42.79V29.03h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-accent)\" d=\"M3.99 42.79V29.03h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M31.02 42.79V29.03h21.083v13.76ZM58.05 42.79V29.03h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M85.08 42.79V29.03h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M112.11 42.79V29.03h21.083v13.76ZM139.14 42.79V29.03h21.083v13.76ZM-23.04 60.43V46.67h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M3.99 60.43V46.67h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M31.02 60.43V46.67h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M58.05 60.43V46.67h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M85.08 60.43V46.67h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M112.11 60.43V46.67h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M139.14 60.43V46.67h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M-23.04 78.07V64.31h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M3.99 78.07V64.31h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M31.02 78.07V64.31h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M58.05 78.07V64.31h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M85.08 78.07V64.31h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" d=\"M112.11 78.07V64.31h21.083v13.76Zm21.083-13.76\"\u002F>\u003Cpath fill=\"var(--tk-accent)\" d=\"M139.14 78.07V64.31h21.083v13.76Zm21.083-13.76\"\u002F>\u003C\u002Fg>\u003C\u002Fsvg>\u003Cfigcaption class=\"tikz-cap\">A self-attention matrix. Each row is a token's query and darker cells carry more weight; here "it" attends strongly to "animal", resolving the reference in one hop.\u003C\u002Ffigcaption>","\u003Csvg style=\"width:100%;max-width:476.460px;height:auto\" xmlns=\"http:\u002F\u002Fwww.w3.org\u002F2000\u002Fsvg\" viewBox=\"-75 -75 357.345 157.199\">\u003Cg stroke=\"currentColor\" style=\"stroke-miterlimit:10;stroke-width:.4\">\u003Cg stroke=\"var(--tk-accent)\">\u003Cpath fill=\"none\" d=\"M-65.403 16.133h48.37V-9.474h-48.37Z\"\u002F>\u003Cg fill=\"var(--tk-accent)\" stroke=\"none\" font-family=\"cmr8\" font-size=\"8\">\u003Cg transform=\"translate(-14.517 1.956)\">\u003Cpath d=\"M-39.121 3.330L-40.899 3.330L-40.899 3.033Q-40.625 3.033-40.457 2.986Q-40.289 2.939-40.289 2.771L-40.289 0.635Q-40.289 0.420-40.346 0.324Q-40.403 0.228-40.516 0.207Q-40.629 0.185-40.875 0.185L-40.875-0.111L-39.676-0.197L-39.676 2.771Q-39.676 2.939-39.530 2.986Q-39.383 3.033-39.121 3.033L-39.121 3.330M-40.563-1.592Q-40.563-1.783-40.428-1.914Q-40.293-2.045-40.098-2.045Q-39.977-2.045-39.873-1.982Q-39.770-1.920-39.707-1.816Q-39.645-1.713-39.645-1.592Q-39.645-1.397-39.776-1.262Q-39.907-1.127-40.098-1.127Q-40.297-1.127-40.430-1.260Q-40.563-1.393-40.563-1.592M-36.692 3.330L-38.547 3.330L-38.547 3.033Q-38.274 3.033-38.106 2.986Q-37.938 2.939-37.938 2.771L-37.938 0.635Q-37.938 0.420-38 0.324Q-38.063 0.228-38.182 0.207Q-38.301 0.185-38.547 0.185L-38.547-0.111L-37.356-0.197L-37.356 0.537Q-37.242 0.322-37.049 0.154Q-36.856-0.014-36.617-0.106Q-36.379-0.197-36.125-0.197Q-34.957-0.197-34.957 0.881L-34.957 2.771Q-34.957 2.939-34.787 2.986Q-34.617 3.033-34.348 3.033L-34.348 3.330L-36.203 3.330L-36.203 3.033Q-35.930 3.033-35.762 2.986Q-35.594 2.939-35.594 2.771L-35.594 0.896Q-35.594 0.514-35.715 0.285Q-35.836 0.057-36.188 0.057Q-36.500 0.057-36.754 0.219Q-37.008 0.381-37.155 0.650Q-37.301 0.920-37.301 1.217L-37.301 2.771Q-37.301 2.939-37.131 2.986Q-36.961 3.033-36.692 3.033L-36.692 3.330M-32.020 4.881L-33.875 4.881L-33.875 4.588Q-33.606 4.588-33.438 4.543Q-33.270 4.498-33.270 4.322L-33.270 0.498Q-33.270 0.291-33.426 0.238Q-33.582 0.185-33.875 0.185L-33.875-0.111L-32.653-0.197L-32.653 0.268Q-32.422 0.045-32.108-0.076Q-31.793-0.197-31.453-0.197Q-30.981-0.197-30.576 0.049Q-30.172 0.295-29.940 0.711Q-29.707 1.127-29.707 1.603Q-29.707 1.978-29.856 2.307Q-30.004 2.635-30.274 2.887Q-30.543 3.139-30.887 3.273Q-31.231 3.408-31.590 3.408Q-31.879 3.408-32.151 3.287Q-32.422 3.166-32.629 2.955L-32.629 4.322Q-32.629 4.498-32.461 4.543Q-32.293 4.588-32.020 4.588L-32.020 4.881M-32.629 0.666L-32.629 2.506Q-32.477 2.795-32.215 2.975Q-31.953 3.154-31.645 3.154Q-31.360 3.154-31.137 3.016Q-30.914 2.877-30.762 2.646Q-30.610 2.416-30.532 2.144Q-30.453 1.873-30.453 1.603Q-30.453 1.271-30.578 0.914Q-30.703 0.557-30.951 0.320Q-31.199 0.084-31.547 0.084Q-31.871 0.084-32.166 0.240Q-32.461 0.396-32.629 0.666M-28.500 2.377L-28.500 0.635Q-28.500 0.420-28.563 0.324Q-28.625 0.228-28.744 0.207Q-28.864 0.185-29.110 0.185L-29.110-0.111L-27.864-0.197L-27.864 2.353L-27.864 2.377Q-27.864 2.689-27.809 2.851Q-27.754 3.014-27.604 3.084Q-27.453 3.154-27.133 3.154Q-26.703 3.154-26.430 2.816Q-26.157 2.478-26.157 2.033L-26.157 0.635Q-26.157 0.420-26.219 0.324Q-26.282 0.228-26.401 0.207Q-26.520 0.185-26.766 0.185L-26.766-0.111L-25.520-0.197L-25.520 2.588Q-25.520 2.799-25.457 2.894Q-25.395 2.990-25.276 3.012Q-25.157 3.033-24.910 3.033L-24.910 3.330L-26.133 3.408L-26.133 2.787Q-26.301 3.076-26.582 3.242Q-26.864 3.408-27.184 3.408Q-28.500 3.408-28.500 2.377M-23.840 2.369L-23.840 0.178L-24.543 0.178L-24.543-0.076Q-24.188-0.076-23.946-0.309Q-23.703-0.541-23.592-0.889Q-23.481-1.236-23.481-1.592L-23.199-1.592L-23.199-0.119L-22.024-0.119L-22.024 0.178L-23.199 0.178L-23.199 2.353Q-23.199 2.674-23.080 2.902Q-22.961 3.131-22.680 3.131Q-22.500 3.131-22.383 3.008Q-22.266 2.885-22.213 2.705Q-22.160 2.525-22.160 2.353L-22.160 1.881L-21.879 1.881L-21.879 2.369Q-21.879 2.623-21.985 2.863Q-22.090 3.103-22.287 3.256Q-22.485 3.408-22.742 3.408Q-23.059 3.408-23.311 3.285Q-23.563 3.162-23.701 2.928Q-23.840 2.693-23.840 2.369\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-14.517 1.956)\">\u003Cpath d=\"M-16.392 3.330L-18.329 3.330L-18.329 3.033Q-17.509 3.033-17.161 2.553L-15.743 0.611L-17.314-1.572Q-17.458-1.748-17.665-1.793Q-17.872-1.838-18.208-1.838L-18.208-2.135L-15.978-2.135L-15.978-1.838Q-16.118-1.838-16.282-1.781Q-16.447-1.725-16.447-1.623Q-16.447-1.592-16.431-1.572L-15.294 0.002L-14.306-1.357Q-14.255-1.412-14.255-1.502Q-14.255-1.658-14.398-1.748Q-14.540-1.838-14.704-1.838L-14.704-2.135L-12.775-2.135L-12.775-1.838Q-13.142-1.838-13.447-1.727Q-13.751-1.615-13.943-1.357L-15.118 0.260L-13.314 2.771Q-13.169 2.943-12.960 2.988Q-12.751 3.033-12.415 3.033L-12.415 3.330L-14.650 3.330L-14.650 3.033Q-14.521 3.033-14.349 2.978Q-14.177 2.924-14.177 2.818Q-14.177 2.803-14.193 2.771L-15.560 0.865L-16.790 2.553Q-16.841 2.615-16.841 2.697Q-16.841 2.846-16.698 2.939Q-16.556 3.033-16.392 3.033\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"M37.027-46.463h48.37V-72.07h-48.37Z\"\u002F>\u003Cg stroke=\"none\" font-family=\"cmr8\" font-size=\"8\">\u003Cg transform=\"translate(90.152 -59.818)\">\u003Cpath d=\"M-39.051 3.330L-40.907 3.330L-40.907 3.033Q-40.633 3.033-40.465 2.986Q-40.297 2.939-40.297 2.771L-40.297-1.389Q-40.297-1.604-40.360-1.699Q-40.422-1.795-40.541-1.816Q-40.660-1.838-40.907-1.838L-40.907-2.135L-39.684-2.221L-39.684 0.482Q-39.559 0.271-39.371 0.121Q-39.184-0.029-38.957-0.113Q-38.731-0.197-38.485-0.197Q-37.317-0.197-37.317 0.881L-37.317 2.771Q-37.317 2.939-37.147 2.986Q-36.977 3.033-36.707 3.033L-36.707 3.330L-38.563 3.330L-38.563 3.033Q-38.289 3.033-38.121 2.986Q-37.953 2.939-37.953 2.771L-37.953 0.896Q-37.953 0.514-38.074 0.285Q-38.196 0.057-38.547 0.057Q-38.860 0.057-39.114 0.219Q-39.367 0.381-39.514 0.650Q-39.660 0.920-39.660 1.217L-39.660 2.771Q-39.660 2.939-39.490 2.986Q-39.321 3.033-39.051 3.033L-39.051 3.330M-36.262 1.576Q-36.262 1.096-36.030 0.680Q-35.797 0.264-35.387 0.014Q-34.977-0.236-34.500-0.236Q-33.770-0.236-33.371 0.205Q-32.973 0.646-32.973 1.377Q-32.973 1.482-33.067 1.506L-35.516 1.506L-35.516 1.576Q-35.516 1.986-35.395 2.342Q-35.274 2.697-35.002 2.914Q-34.731 3.131-34.301 3.131Q-33.938 3.131-33.641 2.902Q-33.344 2.674-33.242 2.322Q-33.235 2.275-33.149 2.260L-33.067 2.260Q-32.973 2.287-32.973 2.369Q-32.973 2.377-32.981 2.408Q-33.043 2.635-33.182 2.818Q-33.321 3.002-33.512 3.135Q-33.703 3.268-33.922 3.338Q-34.141 3.408-34.379 3.408Q-34.750 3.408-35.088 3.271Q-35.426 3.135-35.694 2.883Q-35.961 2.631-36.112 2.291Q-36.262 1.951-36.262 1.576M-35.508 1.268L-33.547 1.268Q-33.547 0.963-33.649 0.672Q-33.750 0.381-33.967 0.199Q-34.184 0.018-34.500 0.018Q-34.801 0.018-35.032 0.205Q-35.262 0.393-35.385 0.684Q-35.508 0.975-35.508 1.268M-32.387 2.498Q-32.387 2.014-31.985 1.719Q-31.582 1.424-31.032 1.305Q-30.481 1.185-29.989 1.185L-29.989 0.896Q-29.989 0.670-30.104 0.463Q-30.219 0.256-30.416 0.137Q-30.614 0.018-30.844 0.018Q-31.270 0.018-31.555 0.123Q-31.485 0.150-31.438 0.205Q-31.391 0.260-31.365 0.330Q-31.340 0.400-31.340 0.475Q-31.340 0.580-31.391 0.672Q-31.442 0.764-31.533 0.814Q-31.625 0.865-31.731 0.865Q-31.836 0.865-31.928 0.814Q-32.020 0.764-32.071 0.672Q-32.121 0.580-32.121 0.475Q-32.121 0.057-31.733-0.090Q-31.344-0.236-30.844-0.236Q-30.512-0.236-30.158-0.106Q-29.805 0.025-29.576 0.279Q-29.348 0.533-29.348 0.881L-29.348 2.682Q-29.348 2.814-29.276 2.924Q-29.203 3.033-29.074 3.033Q-28.949 3.033-28.881 2.928Q-28.813 2.822-28.813 2.682L-28.813 2.170L-28.532 2.170L-28.532 2.682Q-28.532 2.885-28.649 3.043Q-28.766 3.201-28.948 3.285Q-29.129 3.369-29.332 3.369Q-29.563 3.369-29.715 3.197Q-29.867 3.025-29.899 2.795Q-30.059 3.076-30.367 3.242Q-30.676 3.408-31.028 3.408Q-31.539 3.408-31.963 3.185Q-32.387 2.963-32.387 2.498M-31.699 2.498Q-31.699 2.783-31.473 2.969Q-31.246 3.154-30.953 3.154Q-30.707 3.154-30.483 3.037Q-30.258 2.920-30.123 2.717Q-29.989 2.514-29.989 2.260L-29.989 1.428Q-30.254 1.428-30.539 1.482Q-30.824 1.537-31.096 1.666Q-31.367 1.795-31.533 2.002Q-31.699 2.209-31.699 2.498M-26.422 3.408Q-26.903 3.408-27.311 3.164Q-27.719 2.920-27.957 2.506Q-28.196 2.092-28.196 1.603Q-28.196 1.111-27.938 0.695Q-27.680 0.279-27.248 0.041Q-26.817-0.197-26.324-0.197Q-25.703-0.197-25.254 0.240L-25.254-1.389Q-25.254-1.604-25.317-1.699Q-25.379-1.795-25.496-1.816Q-25.614-1.838-25.860-1.838L-25.860-2.135L-24.637-2.221L-24.637 2.588Q-24.637 2.799-24.574 2.894Q-24.512 2.990-24.395 3.012Q-24.278 3.033-24.028 3.033L-24.028 3.330L-25.278 3.408L-25.278 2.924Q-25.742 3.408-26.422 3.408M-26.356 3.154Q-26.016 3.154-25.723 2.963Q-25.430 2.771-25.278 2.475L-25.278 0.643Q-25.426 0.369-25.688 0.213Q-25.949 0.057-26.262 0.057Q-26.887 0.057-27.170 0.504Q-27.453 0.951-27.453 1.611Q-27.453 2.256-27.201 2.705Q-26.949 3.154-26.356 3.154\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(90.152 -59.818)\">\u003Cpath d=\"M-17.319 3.330L-20.112 3.330L-20.112 3.033Q-19.050 3.033-19.050 2.771L-19.050-1.397Q-19.479-1.182-20.159-1.182L-20.159-1.479Q-19.140-1.479-18.624-1.990L-18.479-1.990Q-18.405-1.971-18.386-1.893L-18.386 2.771Q-18.386 3.033-17.319 3.033\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"M37.027 16.133h48.37V-9.474h-48.37Z\"\u002F>\u003Cg stroke=\"none\" font-family=\"cmr8\" font-size=\"8\">\u003Cg transform=\"translate(90.152 2.778)\">\u003Cpath d=\"M-39.051 3.330L-40.907 3.330L-40.907 3.033Q-40.633 3.033-40.465 2.986Q-40.297 2.939-40.297 2.771L-40.297-1.389Q-40.297-1.604-40.360-1.699Q-40.422-1.795-40.541-1.816Q-40.660-1.838-40.907-1.838L-40.907-2.135L-39.684-2.221L-39.684 0.482Q-39.559 0.271-39.371 0.121Q-39.184-0.029-38.957-0.113Q-38.731-0.197-38.485-0.197Q-37.317-0.197-37.317 0.881L-37.317 2.771Q-37.317 2.939-37.147 2.986Q-36.977 3.033-36.707 3.033L-36.707 3.330L-38.563 3.330L-38.563 3.033Q-38.289 3.033-38.121 2.986Q-37.953 2.939-37.953 2.771L-37.953 0.896Q-37.953 0.514-38.074 0.285Q-38.196 0.057-38.547 0.057Q-38.860 0.057-39.114 0.219Q-39.367 0.381-39.514 0.650Q-39.660 0.920-39.660 1.217L-39.660 2.771Q-39.660 2.939-39.490 2.986Q-39.321 3.033-39.051 3.033L-39.051 3.330M-36.262 1.576Q-36.262 1.096-36.030 0.680Q-35.797 0.264-35.387 0.014Q-34.977-0.236-34.500-0.236Q-33.770-0.236-33.371 0.205Q-32.973 0.646-32.973 1.377Q-32.973 1.482-33.067 1.506L-35.516 1.506L-35.516 1.576Q-35.516 1.986-35.395 2.342Q-35.274 2.697-35.002 2.914Q-34.731 3.131-34.301 3.131Q-33.938 3.131-33.641 2.902Q-33.344 2.674-33.242 2.322Q-33.235 2.275-33.149 2.260L-33.067 2.260Q-32.973 2.287-32.973 2.369Q-32.973 2.377-32.981 2.408Q-33.043 2.635-33.182 2.818Q-33.321 3.002-33.512 3.135Q-33.703 3.268-33.922 3.338Q-34.141 3.408-34.379 3.408Q-34.750 3.408-35.088 3.271Q-35.426 3.135-35.694 2.883Q-35.961 2.631-36.112 2.291Q-36.262 1.951-36.262 1.576M-35.508 1.268L-33.547 1.268Q-33.547 0.963-33.649 0.672Q-33.750 0.381-33.967 0.199Q-34.184 0.018-34.500 0.018Q-34.801 0.018-35.032 0.205Q-35.262 0.393-35.385 0.684Q-35.508 0.975-35.508 1.268M-32.387 2.498Q-32.387 2.014-31.985 1.719Q-31.582 1.424-31.032 1.305Q-30.481 1.185-29.989 1.185L-29.989 0.896Q-29.989 0.670-30.104 0.463Q-30.219 0.256-30.416 0.137Q-30.614 0.018-30.844 0.018Q-31.270 0.018-31.555 0.123Q-31.485 0.150-31.438 0.205Q-31.391 0.260-31.365 0.330Q-31.340 0.400-31.340 0.475Q-31.340 0.580-31.391 0.672Q-31.442 0.764-31.533 0.814Q-31.625 0.865-31.731 0.865Q-31.836 0.865-31.928 0.814Q-32.020 0.764-32.071 0.672Q-32.121 0.580-32.121 0.475Q-32.121 0.057-31.733-0.090Q-31.344-0.236-30.844-0.236Q-30.512-0.236-30.158-0.106Q-29.805 0.025-29.576 0.279Q-29.348 0.533-29.348 0.881L-29.348 2.682Q-29.348 2.814-29.276 2.924Q-29.203 3.033-29.074 3.033Q-28.949 3.033-28.881 2.928Q-28.813 2.822-28.813 2.682L-28.813 2.170L-28.532 2.170L-28.532 2.682Q-28.532 2.885-28.649 3.043Q-28.766 3.201-28.948 3.285Q-29.129 3.369-29.332 3.369Q-29.563 3.369-29.715 3.197Q-29.867 3.025-29.899 2.795Q-30.059 3.076-30.367 3.242Q-30.676 3.408-31.028 3.408Q-31.539 3.408-31.963 3.185Q-32.387 2.963-32.387 2.498M-31.699 2.498Q-31.699 2.783-31.473 2.969Q-31.246 3.154-30.953 3.154Q-30.707 3.154-30.483 3.037Q-30.258 2.920-30.123 2.717Q-29.989 2.514-29.989 2.260L-29.989 1.428Q-30.254 1.428-30.539 1.482Q-30.824 1.537-31.096 1.666Q-31.367 1.795-31.533 2.002Q-31.699 2.209-31.699 2.498M-26.422 3.408Q-26.903 3.408-27.311 3.164Q-27.719 2.920-27.957 2.506Q-28.196 2.092-28.196 1.603Q-28.196 1.111-27.938 0.695Q-27.680 0.279-27.248 0.041Q-26.817-0.197-26.324-0.197Q-25.703-0.197-25.254 0.240L-25.254-1.389Q-25.254-1.604-25.317-1.699Q-25.379-1.795-25.496-1.816Q-25.614-1.838-25.860-1.838L-25.860-2.135L-24.637-2.221L-24.637 2.588Q-24.637 2.799-24.574 2.894Q-24.512 2.990-24.395 3.012Q-24.278 3.033-24.028 3.033L-24.028 3.330L-25.278 3.408L-25.278 2.924Q-25.742 3.408-26.422 3.408M-26.356 3.154Q-26.016 3.154-25.723 2.963Q-25.430 2.771-25.278 2.475L-25.278 0.643Q-25.426 0.369-25.688 0.213Q-25.949 0.057-26.262 0.057Q-26.887 0.057-27.170 0.504Q-27.453 0.951-27.453 1.611Q-27.453 2.256-27.201 2.705Q-26.949 3.154-26.356 3.154\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(90.152 2.778)\">\u003Cpath d=\"M-17.327 3.330L-20.487 3.330L-20.487 3.123Q-20.487 3.096-20.464 3.064L-19.112 1.666Q-18.733 1.279-18.485 0.990Q-18.237 0.701-18.063 0.344Q-17.890-0.014-17.890-0.404Q-17.890-0.752-18.022-1.045Q-18.155-1.338-18.409-1.516Q-18.663-1.693-19.018-1.693Q-19.378-1.693-19.669-1.498Q-19.960-1.303-20.104-0.975L-20.050-0.975Q-19.866-0.975-19.741-0.854Q-19.616-0.732-19.616-0.541Q-19.616-0.361-19.741-0.232Q-19.866-0.104-20.050-0.104Q-20.229-0.104-20.358-0.232Q-20.487-0.361-20.487-0.541Q-20.487-0.943-20.267-1.279Q-20.046-1.615-19.681-1.803Q-19.315-1.990-18.913-1.990Q-18.433-1.990-18.017-1.803Q-17.601-1.615-17.349-1.254Q-17.097-0.893-17.097-0.404Q-17.097-0.045-17.251 0.258Q-17.405 0.560-17.657 0.820Q-17.909 1.080-18.259 1.365Q-18.608 1.650-18.776 1.803L-19.706 2.643L-18.991 2.643Q-17.616 2.643-17.577 2.603Q-17.507 2.525-17.464 2.340Q-17.421 2.154-17.378 1.865L-17.097 1.865\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"M37.027 78.73h48.37V53.121h-48.37Z\"\u002F>\u003Cg stroke=\"none\" font-family=\"cmr8\" font-size=\"8\">\u003Cg transform=\"translate(89.916 65.374)\">\u003Cpath d=\"M-39.051 3.330L-40.907 3.330L-40.907 3.033Q-40.633 3.033-40.465 2.986Q-40.297 2.939-40.297 2.771L-40.297-1.389Q-40.297-1.604-40.360-1.699Q-40.422-1.795-40.541-1.816Q-40.660-1.838-40.907-1.838L-40.907-2.135L-39.684-2.221L-39.684 0.482Q-39.559 0.271-39.371 0.121Q-39.184-0.029-38.957-0.113Q-38.731-0.197-38.485-0.197Q-37.317-0.197-37.317 0.881L-37.317 2.771Q-37.317 2.939-37.147 2.986Q-36.977 3.033-36.707 3.033L-36.707 3.330L-38.563 3.330L-38.563 3.033Q-38.289 3.033-38.121 2.986Q-37.953 2.939-37.953 2.771L-37.953 0.896Q-37.953 0.514-38.074 0.285Q-38.196 0.057-38.547 0.057Q-38.860 0.057-39.114 0.219Q-39.367 0.381-39.514 0.650Q-39.660 0.920-39.660 1.217L-39.660 2.771Q-39.660 2.939-39.490 2.986Q-39.321 3.033-39.051 3.033L-39.051 3.330M-36.262 1.576Q-36.262 1.096-36.030 0.680Q-35.797 0.264-35.387 0.014Q-34.977-0.236-34.500-0.236Q-33.770-0.236-33.371 0.205Q-32.973 0.646-32.973 1.377Q-32.973 1.482-33.067 1.506L-35.516 1.506L-35.516 1.576Q-35.516 1.986-35.395 2.342Q-35.274 2.697-35.002 2.914Q-34.731 3.131-34.301 3.131Q-33.938 3.131-33.641 2.902Q-33.344 2.674-33.242 2.322Q-33.235 2.275-33.149 2.260L-33.067 2.260Q-32.973 2.287-32.973 2.369Q-32.973 2.377-32.981 2.408Q-33.043 2.635-33.182 2.818Q-33.321 3.002-33.512 3.135Q-33.703 3.268-33.922 3.338Q-34.141 3.408-34.379 3.408Q-34.750 3.408-35.088 3.271Q-35.426 3.135-35.694 2.883Q-35.961 2.631-36.112 2.291Q-36.262 1.951-36.262 1.576M-35.508 1.268L-33.547 1.268Q-33.547 0.963-33.649 0.672Q-33.750 0.381-33.967 0.199Q-34.184 0.018-34.500 0.018Q-34.801 0.018-35.032 0.205Q-35.262 0.393-35.385 0.684Q-35.508 0.975-35.508 1.268M-32.387 2.498Q-32.387 2.014-31.985 1.719Q-31.582 1.424-31.032 1.305Q-30.481 1.185-29.989 1.185L-29.989 0.896Q-29.989 0.670-30.104 0.463Q-30.219 0.256-30.416 0.137Q-30.614 0.018-30.844 0.018Q-31.270 0.018-31.555 0.123Q-31.485 0.150-31.438 0.205Q-31.391 0.260-31.365 0.330Q-31.340 0.400-31.340 0.475Q-31.340 0.580-31.391 0.672Q-31.442 0.764-31.533 0.814Q-31.625 0.865-31.731 0.865Q-31.836 0.865-31.928 0.814Q-32.020 0.764-32.071 0.672Q-32.121 0.580-32.121 0.475Q-32.121 0.057-31.733-0.090Q-31.344-0.236-30.844-0.236Q-30.512-0.236-30.158-0.106Q-29.805 0.025-29.576 0.279Q-29.348 0.533-29.348 0.881L-29.348 2.682Q-29.348 2.814-29.276 2.924Q-29.203 3.033-29.074 3.033Q-28.949 3.033-28.881 2.928Q-28.813 2.822-28.813 2.682L-28.813 2.170L-28.532 2.170L-28.532 2.682Q-28.532 2.885-28.649 3.043Q-28.766 3.201-28.948 3.285Q-29.129 3.369-29.332 3.369Q-29.563 3.369-29.715 3.197Q-29.867 3.025-29.899 2.795Q-30.059 3.076-30.367 3.242Q-30.676 3.408-31.028 3.408Q-31.539 3.408-31.963 3.185Q-32.387 2.963-32.387 2.498M-31.699 2.498Q-31.699 2.783-31.473 2.969Q-31.246 3.154-30.953 3.154Q-30.707 3.154-30.483 3.037Q-30.258 2.920-30.123 2.717Q-29.989 2.514-29.989 2.260L-29.989 1.428Q-30.254 1.428-30.539 1.482Q-30.824 1.537-31.096 1.666Q-31.367 1.795-31.533 2.002Q-31.699 2.209-31.699 2.498M-26.422 3.408Q-26.903 3.408-27.311 3.164Q-27.719 2.920-27.957 2.506Q-28.196 2.092-28.196 1.603Q-28.196 1.111-27.938 0.695Q-27.680 0.279-27.248 0.041Q-26.817-0.197-26.324-0.197Q-25.703-0.197-25.254 0.240L-25.254-1.389Q-25.254-1.604-25.317-1.699Q-25.379-1.795-25.496-1.816Q-25.614-1.838-25.860-1.838L-25.860-2.135L-24.637-2.221L-24.637 2.588Q-24.637 2.799-24.574 2.894Q-24.512 2.990-24.395 3.012Q-24.278 3.033-24.028 3.033L-24.028 3.330L-25.278 3.408L-25.278 2.924Q-25.742 3.408-26.422 3.408M-26.356 3.154Q-26.016 3.154-25.723 2.963Q-25.430 2.771-25.278 2.475L-25.278 0.643Q-25.426 0.369-25.688 0.213Q-25.949 0.057-26.262 0.057Q-26.887 0.057-27.170 0.504Q-27.453 0.951-27.453 1.611Q-27.453 2.256-27.201 2.705Q-26.949 3.154-26.356 3.154\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(89.916 65.374)\">\u003Cpath d=\"M-18.745 3.330L-20.601 3.330L-20.601 3.033Q-20.327 3.033-20.159 2.986Q-19.991 2.939-19.991 2.771L-19.991-1.389Q-19.991-1.604-20.054-1.699Q-20.116-1.795-20.235-1.816Q-20.354-1.838-20.601-1.838L-20.601-2.135L-19.378-2.221L-19.378 0.482Q-19.253 0.271-19.065 0.121Q-18.878-0.029-18.651-0.113Q-18.425-0.197-18.179-0.197Q-17.011-0.197-17.011 0.881L-17.011 2.771Q-17.011 2.939-16.841 2.986Q-16.671 3.033-16.401 3.033L-16.401 3.330L-18.257 3.330L-18.257 3.033Q-17.983 3.033-17.815 2.986Q-17.647 2.939-17.647 2.771L-17.647 0.896Q-17.647 0.514-17.768 0.285Q-17.890 0.057-18.241 0.057Q-18.554 0.057-18.808 0.219Q-19.061 0.381-19.208 0.650Q-19.354 0.920-19.354 1.217L-19.354 2.771Q-19.354 2.939-19.184 2.986Q-19.015 3.033-18.745 3.033\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"none\" font-family=\"cmr9\" font-size=\"9\">\u003Cg transform=\"translate(101.145 38.784)\">\u003Cpath d=\"M-40.446-5.175Q-40.446-5.303-40.377-5.419Q-40.309-5.536-40.193-5.606Q-40.076-5.676-39.940-5.676Q-39.738-5.676-39.586-5.529Q-39.435-5.382-39.435-5.175Q-39.435-4.969-39.584-4.819Q-39.734-4.670-39.940-4.670Q-40.151-4.670-40.298-4.822Q-40.446-4.973-40.446-5.175\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.270\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(101.145 38.784)\">\u003Cpath d=\"M-40.446-1.175Q-40.446-1.303-40.377-1.419Q-40.309-1.536-40.193-1.606Q-40.076-1.676-39.940-1.676Q-39.738-1.676-39.586-1.529Q-39.435-1.382-39.435-1.175Q-39.435-0.969-39.584-0.819Q-39.734-0.670-39.940-0.670Q-40.151-0.670-40.298-0.822Q-40.446-0.973-40.446-1.175\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.270\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(101.145 38.784)\">\u003Cpath d=\"M-40.446 2.825Q-40.446 2.697-40.377 2.581Q-40.309 2.464-40.193 2.394Q-40.076 2.324-39.940 2.324Q-39.738 2.324-39.586 2.471Q-39.435 2.618-39.435 2.825Q-39.435 3.031-39.584 3.181Q-39.734 3.330-39.940 3.330Q-40.151 3.330-40.298 3.178Q-40.446 3.027-40.446 2.825\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.270\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"var(--tk-accent)\">\u003Cpath fill=\"none\" d=\"M142.302 16.133h48.37V-9.474h-48.37Z\"\u002F>\u003Cg transform=\"translate(195.663 2.46)\">\u003Cpath d=\"M-40.938 1.603Q-40.938 1.107-40.688 0.682Q-40.438 0.256-40.018 0.010Q-39.598-0.236-39.098-0.236Q-38.559-0.236-38.168-0.111Q-37.778 0.014-37.778 0.428Q-37.778 0.533-37.828 0.625Q-37.879 0.717-37.971 0.768Q-38.063 0.818-38.172 0.818Q-38.278 0.818-38.369 0.768Q-38.461 0.717-38.512 0.625Q-38.563 0.533-38.563 0.428Q-38.563 0.205-38.395 0.100Q-38.617 0.041-39.090 0.041Q-39.387 0.041-39.602 0.180Q-39.817 0.318-39.948 0.549Q-40.078 0.779-40.137 1.049Q-40.196 1.318-40.196 1.603Q-40.196 1.998-40.063 2.348Q-39.930 2.697-39.658 2.914Q-39.387 3.131-38.989 3.131Q-38.614 3.131-38.338 2.914Q-38.063 2.697-37.961 2.338Q-37.946 2.275-37.883 2.275L-37.778 2.275Q-37.742 2.275-37.717 2.303Q-37.692 2.330-37.692 2.369L-37.692 2.393Q-37.824 2.873-38.209 3.141Q-38.594 3.408-39.098 3.408Q-39.461 3.408-39.795 3.271Q-40.129 3.135-40.389 2.885Q-40.649 2.635-40.793 2.299Q-40.938 1.963-40.938 1.603M-37.203 1.635Q-37.203 1.131-36.948 0.699Q-36.692 0.268-36.256 0.016Q-35.821-0.236-35.321-0.236Q-34.934-0.236-34.592-0.092Q-34.250 0.053-33.989 0.314Q-33.727 0.576-33.584 0.912Q-33.442 1.248-33.442 1.635Q-33.442 2.127-33.705 2.537Q-33.969 2.947-34.399 3.178Q-34.828 3.408-35.321 3.408Q-35.813 3.408-36.246 3.176Q-36.680 2.943-36.942 2.535Q-37.203 2.127-37.203 1.635M-35.321 3.131Q-34.864 3.131-34.612 2.908Q-34.360 2.685-34.272 2.334Q-34.184 1.982-34.184 1.537Q-34.184 1.107-34.278 0.769Q-34.371 0.432-34.625 0.225Q-34.879 0.018-35.321 0.018Q-35.969 0.018-36.213 0.434Q-36.457 0.850-36.457 1.537Q-36.457 1.982-36.369 2.334Q-36.282 2.685-36.030 2.908Q-35.778 3.131-35.321 3.131M-31.028 3.330L-32.883 3.330L-32.883 3.033Q-32.610 3.033-32.442 2.986Q-32.274 2.939-32.274 2.771L-32.274 0.635Q-32.274 0.420-32.336 0.324Q-32.399 0.228-32.518 0.207Q-32.637 0.185-32.883 0.185L-32.883-0.111L-31.692-0.197L-31.692 0.537Q-31.578 0.322-31.385 0.154Q-31.192-0.014-30.953-0.106Q-30.715-0.197-30.461-0.197Q-29.293-0.197-29.293 0.881L-29.293 2.771Q-29.293 2.939-29.123 2.986Q-28.953 3.033-28.684 3.033L-28.684 3.330L-30.539 3.330L-30.539 3.033Q-30.266 3.033-30.098 2.986Q-29.930 2.939-29.930 2.771L-29.930 0.896Q-29.930 0.514-30.051 0.285Q-30.172 0.057-30.524 0.057Q-30.836 0.057-31.090 0.219Q-31.344 0.381-31.490 0.650Q-31.637 0.920-31.637 1.217L-31.637 2.771Q-31.637 2.939-31.467 2.986Q-31.297 3.033-31.028 3.033L-31.028 3.330M-28.196 1.603Q-28.196 1.107-27.946 0.682Q-27.696 0.256-27.276 0.010Q-26.856-0.236-26.356-0.236Q-25.817-0.236-25.426-0.111Q-25.035 0.014-25.035 0.428Q-25.035 0.533-25.086 0.625Q-25.137 0.717-25.229 0.768Q-25.321 0.818-25.430 0.818Q-25.535 0.818-25.627 0.768Q-25.719 0.717-25.770 0.625Q-25.821 0.533-25.821 0.428Q-25.821 0.205-25.653 0.100Q-25.875 0.041-26.348 0.041Q-26.645 0.041-26.860 0.180Q-27.074 0.318-27.205 0.549Q-27.336 0.779-27.395 1.049Q-27.453 1.318-27.453 1.603Q-27.453 1.998-27.321 2.348Q-27.188 2.697-26.916 2.914Q-26.645 3.131-26.246 3.131Q-25.871 3.131-25.596 2.914Q-25.321 2.697-25.219 2.338Q-25.203 2.275-25.141 2.275L-25.035 2.275Q-25 2.275-24.975 2.303Q-24.949 2.330-24.949 2.369L-24.949 2.393Q-25.082 2.873-25.467 3.141Q-25.852 3.408-26.356 3.408Q-26.719 3.408-27.053 3.271Q-27.387 3.135-27.647 2.885Q-27.907 2.635-28.051 2.299Q-28.196 1.963-28.196 1.603M-24.364 2.498Q-24.364 2.014-23.961 1.719Q-23.559 1.424-23.008 1.305Q-22.457 1.185-21.965 1.185L-21.965 0.896Q-21.965 0.670-22.080 0.463Q-22.196 0.256-22.393 0.137Q-22.590 0.018-22.821 0.018Q-23.246 0.018-23.532 0.123Q-23.461 0.150-23.414 0.205Q-23.367 0.260-23.342 0.330Q-23.317 0.400-23.317 0.475Q-23.317 0.580-23.367 0.672Q-23.418 0.764-23.510 0.814Q-23.602 0.865-23.707 0.865Q-23.813 0.865-23.905 0.814Q-23.996 0.764-24.047 0.672Q-24.098 0.580-24.098 0.475Q-24.098 0.057-23.709-0.090Q-23.321-0.236-22.821-0.236Q-22.489-0.236-22.135-0.106Q-21.782 0.025-21.553 0.279Q-21.324 0.533-21.324 0.881L-21.324 2.682Q-21.324 2.814-21.252 2.924Q-21.180 3.033-21.051 3.033Q-20.926 3.033-20.858 2.928Q-20.789 2.822-20.789 2.682L-20.789 2.170L-20.508 2.170L-20.508 2.682Q-20.508 2.885-20.625 3.043Q-20.742 3.201-20.924 3.285Q-21.106 3.369-21.309 3.369Q-21.539 3.369-21.692 3.197Q-21.844 3.025-21.875 2.795Q-22.035 3.076-22.344 3.242Q-22.653 3.408-23.004 3.408Q-23.516 3.408-23.940 3.185Q-24.364 2.963-24.364 2.498M-23.676 2.498Q-23.676 2.783-23.449 2.969Q-23.223 3.154-22.930 3.154Q-22.684 3.154-22.459 3.037Q-22.235 2.920-22.100 2.717Q-21.965 2.514-21.965 2.260L-21.965 1.428Q-22.231 1.428-22.516 1.482Q-22.801 1.537-23.073 1.666Q-23.344 1.795-23.510 2.002Q-23.676 2.209-23.676 2.498M-19.590 2.369L-19.590 0.178L-20.293 0.178L-20.293-0.076Q-19.938-0.076-19.696-0.309Q-19.453-0.541-19.342-0.889Q-19.231-1.236-19.231-1.592L-18.949-1.592L-18.949-0.119L-17.774-0.119L-17.774 0.178L-18.949 0.178L-18.949 2.353Q-18.949 2.674-18.830 2.902Q-18.711 3.131-18.430 3.131Q-18.250 3.131-18.133 3.008Q-18.016 2.885-17.963 2.705Q-17.910 2.525-17.910 2.353L-17.910 1.881L-17.629 1.881L-17.629 2.369Q-17.629 2.623-17.735 2.863Q-17.840 3.103-18.037 3.256Q-18.235 3.408-18.492 3.408Q-18.809 3.408-19.061 3.285Q-19.313 3.162-19.451 2.928Q-19.590 2.693-19.590 2.369\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"var(--tk-accent)\">\u003Cpath fill=\"none\" d=\"M230.505 16.133h48.37V-9.474h-48.37Z\"\u002F>\u003Cg fill=\"var(--tk-accent)\" stroke=\"none\">\u003Cg transform=\"translate(288.75 2.061)\">\u003Cpath d=\"M-39.676 3.385L-40.149-1.693Q-40.180-1.838-40.653-1.838Q-40.754-1.865-40.754-1.967L-40.723-2.068Q-40.692-2.127-40.633-2.135L-38.715-2.135Q-38.680-2.135-38.649-2.094Q-38.617-2.053-38.617-2.014L-38.645-1.908Q-38.684-1.846-38.739-1.838Q-38.973-1.838-39.166-1.799Q-39.360-1.760-39.379-1.623L-39.012 2.393L-36.899-1.158L-36.946-1.693Q-36.996-1.838-37.457-1.838Q-37.555-1.869-37.555-1.967L-37.532-2.068Q-37.500-2.123-37.434-2.135L-35.524-2.135Q-35.481-2.135-35.453-2.098Q-35.426-2.061-35.426-2.014L-35.449-1.908Q-35.485-1.850-35.539-1.838Q-36.149-1.838-36.188-1.623L-36.180-1.572L-35.813 2.393L-33.578-1.365Q-33.574-1.381-33.549-1.449Q-33.524-1.518-33.524-1.565Q-33.524-1.721-33.666-1.779Q-33.809-1.838-33.996-1.838Q-34.090-1.869-34.090-1.967L-34.067-2.068Q-34.035-2.123-33.973-2.135L-32.434-2.135Q-32.395-2.135-32.367-2.098Q-32.340-2.061-32.340-2.014L-32.364-1.908Q-32.403-1.846-32.457-1.838Q-32.739-1.838-32.944-1.695Q-33.149-1.553-33.332-1.279L-33.348-1.264L-36.106 3.385Q-36.164 3.498-36.289 3.498L-36.348 3.498Q-36.395 3.498-36.436 3.467Q-36.477 3.435-36.477 3.385L-36.860-0.725L-39.297 3.385Q-39.371 3.498-39.485 3.498L-39.547 3.498Q-39.590 3.498-39.633 3.465Q-39.676 3.432-39.676 3.385\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(288.75 2.061)\">\u003Cpath d=\"M-32.701 3.127Q-32.701 2.600-32.445 2.112Q-32.188 1.624-31.756 1.248Q-31.324 0.871-30.800 0.660Q-30.275 0.449-29.760 0.449Q-29.376 0.449-29.046 0.564Q-28.717 0.678-28.462 0.898Q-28.207 1.117-28.063 1.437Q-27.920 1.756-27.920 2.143Q-27.920 2.843-28.353 3.458Q-28.787 4.073-29.472 4.437Q-30.158 4.800-30.846 4.800Q-31.224 4.800-31.557 4.687Q-31.889 4.574-32.149 4.356Q-32.408 4.138-32.554 3.827Q-32.701 3.517-32.701 3.127M-30.788 4.554Q-30.319 4.554-29.906 4.322Q-29.493 4.091-29.193 3.712Q-28.892 3.332-28.724 2.863Q-28.555 2.395-28.555 1.949Q-28.555 1.577-28.711 1.287Q-28.866 0.997-29.155 0.841Q-29.443 0.684-29.821 0.684Q-30.451 0.684-30.959 1.054Q-31.468 1.425-31.749 2.017Q-32.030 2.609-32.030 3.233Q-32.030 3.613-31.884 3.915Q-31.737 4.217-31.456 4.385Q-31.175 4.554-30.788 4.554\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.180\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"m-19.943-9.674 57.677-35.233\"\u002F>\u003Cpath stroke=\"none\" d=\"m39.953-46.263-4.634.394 2.415.962-.247 2.588\"\u002F>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M-16.834 3.33h51.06\"\u002F>\u003Cpath stroke=\"none\" d=\"m36.827 3.33-4.16-2.08 1.56 2.08-1.56 2.08\"\u002F>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"m-19.943 16.333 57.677 35.234\"\u002F>\u003Cpath stroke=\"none\" d=\"m39.953 52.922-2.466-3.943.247 2.588-2.415.961\"\u002F>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"m83.079-46.263 59.323 35.26\"\u002F>\u003Cpath stroke=\"none\" d=\"m144.637-9.674-2.513-3.913.278 2.585-2.403.99\"\u002F>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M85.596 3.33h53.906\"\u002F>\u003Cpath stroke=\"none\" d=\"m142.102 3.33-4.16-2.08 1.56 2.08-1.56 2.08\"\u002F>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"m83.079 52.922 59.323-35.26\"\u002F>\u003Cpath stroke=\"none\" d=\"m144.637 16.333-4.638.338 2.403.99-.278 2.586\"\u002F>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M190.871 3.33h36.834\"\u002F>\u003Cpath stroke=\"none\" d=\"m230.305 3.33-4.16-2.08 1.56 2.08-1.56 2.08\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fsvg>\u003Cfigcaption class=\"tikz-cap\">Multi-head attention. The input is projected into \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6944em;\">\u003C\u002Fspan>\u003Cspan class=\"mord mathnormal\">h\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan> subspaces, each running its own scaled dot-product attention; the outputs are concatenated and mixed by \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.8333em;vertical-align:-0.15em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">\u003Cspan class=\"mord mathnormal\" style=\"margin-right:0.1389em;\">W\u003C\u002Fspan>\u003Cspan class=\"msupsub\">\u003Cspan class=\"vlist-t vlist-t2\">\u003Cspan class=\"vlist-r\">\u003Cspan class=\"vlist\" style=\"height:0.3283em;\">\u003Cspan style=\"top:-2.55em;margin-left:-0.1389em;margin-right:0.05em;\">\u003Cspan class=\"pstrut\" style=\"height:2.7em;\">\u003C\u002Fspan>\u003Cspan class=\"sizing reset-size6 size3 mtight\">\u003Cspan class=\"mord mathnormal mtight\" style=\"margin-right:0.0278em;\">O\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"vlist-s\">​\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"vlist-r\">\u003Cspan class=\"vlist\" style=\"height:0.15em;\">\u003Cspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>.\u003C\u002Ffigcaption>","\u003Csvg style=\"width:100%;max-width:416.053px;height:auto\" xmlns=\"http:\u002F\u002Fwww.w3.org\u002F2000\u002Fsvg\" viewBox=\"-75 -75 312.040 191.343\">\u003Cg stroke=\"currentColor\" style=\"stroke-miterlimit:10;stroke-width:.4\">\u003Cg fill=\"var(--tk-soft-accent)\" stroke=\"var(--tk-accent)\">\u003Cpath d=\"M.972-49.308h96.74V-72.07H.972Z\"\u002F>\u003Cg fill=\"currentColor\" stroke=\"none\" font-family=\"cmtt8\" font-size=\"8\">\u003Cg transform=\"translate(-23.375 -88.604)\">\u003Cpath d=\"M49.580 30.122L49.580 30.032Q49.623 29.825 49.830 29.801L50.084 29.801L51.104 27.751L50.190 26.032L49.916 26.032Q49.721 26.008 49.670 25.794L49.670 25.704Q49.721 25.497 49.916 25.473L50.990 25.473Q51.186 25.497 51.237 25.704L51.237 25.794Q51.186 26.008 50.990 26.032L50.908 26.032L51.420 27.008L51.885 26.032L51.791 26.032Q51.580 26.008 51.541 25.794L51.541 25.704Q51.580 25.497 51.791 25.473L52.854 25.473Q53.045 25.497 53.104 25.704L53.104 25.794Q53.045 26.008 52.854 26.032L52.588 26.032L51.733 27.751L52.815 29.801L53.084 29.801Q53.283 29.825 53.334 30.032L53.334 30.122Q53.283 30.337 53.084 30.360L52.022 30.360Q51.815 30.337 51.776 30.122L51.776 30.032Q51.815 29.825 52.022 29.801L52.112 29.801L51.420 28.415L50.783 29.801L50.893 29.801Q51.092 29.825 51.143 30.032L51.143 30.122Q51.092 30.337 50.893 30.360L49.830 30.360Q49.623 30.337 49.580 30.122\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-23.375 -88.604)\">\u003Cpath d=\"M59.408 29.801Q59.408 29.579 59.574 29.413Q59.740 29.247 59.971 29.247Q60.119 29.247 60.246 29.325Q60.373 29.403 60.447 29.528Q60.522 29.653 60.522 29.801Q60.522 30.028 60.356 30.194Q60.190 30.360 59.971 30.360Q59.744 30.360 59.576 30.192Q59.408 30.024 59.408 29.801M59.408 27.465Q59.408 27.243 59.574 27.077Q59.740 26.911 59.971 26.911Q60.119 26.911 60.246 26.989Q60.373 27.067 60.447 27.192Q60.522 27.317 60.522 27.465Q60.522 27.692 60.356 27.858Q60.190 28.024 59.971 28.024Q59.744 28.024 59.576 27.856Q59.408 27.688 59.408 27.465\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-23.375 -88.604)\">\u003Cpath d=\"M66.412 30.122L66.412 30.032Q66.455 29.825 66.662 29.801L67.084 29.801L67.084 27.473L66.662 27.473Q66.455 27.450 66.412 27.231L66.412 27.145Q66.459 26.934 66.662 26.911L67.479 26.911Q67.674 26.934 67.725 27.145L67.725 27.231L67.717 27.255Q67.944 27.075 68.217 26.973Q68.490 26.872 68.783 26.872Q69.131 26.872 69.369 27.012Q69.608 27.153 69.723 27.411Q69.838 27.669 69.838 28.024L69.838 29.801L70.264 29.801Q70.471 29.825 70.510 30.032L70.510 30.122Q70.471 30.337 70.264 30.360L68.869 30.360Q68.674 30.337 68.623 30.122L68.623 30.032Q68.674 29.821 68.869 29.801L69.197 29.801L69.197 28.055Q69.197 27.747 69.108 27.589Q69.018 27.430 68.725 27.430Q68.455 27.430 68.227 27.561Q67.998 27.692 67.862 27.921Q67.725 28.149 67.725 28.415L67.725 29.801L68.151 29.801Q68.358 29.825 68.397 30.032L68.397 30.122Q68.358 30.337 68.151 30.360L66.662 30.360Q66.455 30.337 66.412 30.122\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-23.375 -88.604)\">\u003Cpath d=\"M75.100 30.122L75.100 30.032Q75.139 29.825 75.346 29.801L75.752 29.801L76.682 28.583L75.811 27.473L75.393 27.473Q75.197 27.454 75.147 27.231L75.147 27.145Q75.197 26.934 75.393 26.911L76.553 26.911Q76.752 26.934 76.803 27.145L76.803 27.231Q76.752 27.450 76.553 27.473L76.444 27.473L76.940 28.130L77.416 27.473L77.299 27.473Q77.100 27.450 77.049 27.231L77.049 27.145Q77.100 26.934 77.299 26.911L78.459 26.911Q78.654 26.930 78.705 27.145L78.705 27.231Q78.654 27.450 78.459 27.473L78.049 27.473L77.201 28.583L78.162 29.801L78.569 29.801Q78.768 29.825 78.819 30.032L78.819 30.122Q78.779 30.337 78.569 30.360L77.416 30.360Q77.209 30.337 77.170 30.122L77.170 30.032Q77.209 29.825 77.416 29.801L77.545 29.801L76.940 28.946L76.354 29.801L76.498 29.801Q76.705 29.825 76.744 30.032L76.744 30.122Q76.705 30.337 76.498 30.360L75.346 30.360Q75.151 30.340 75.100 30.122\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-23.375 -88.604)\">\u003Cpath d=\"M84.480 29.481Q84.597 29.684 84.831 29.782Q85.066 29.880 85.327 29.880Q85.624 29.880 85.898 29.751Q86.171 29.622 86.345 29.385Q86.519 29.149 86.519 28.848Q86.519 28.594 86.400 28.352Q86.281 28.110 86.066 27.964Q85.851 27.817 85.581 27.817Q84.976 27.817 84.640 28.137Q84.562 28.223 84.507 28.270Q84.452 28.317 84.366 28.329L84.265 28.329Q84.066 28.305 84.015 28.087L84.015 25.704Q84.066 25.497 84.265 25.473L86.624 25.473Q86.820 25.493 86.870 25.704L86.870 25.794Q86.820 26.008 86.624 26.032L84.656 26.032L84.656 27.458Q85.062 27.255 85.581 27.255Q86.011 27.255 86.376 27.471Q86.741 27.688 86.950 28.057Q87.159 28.426 87.159 28.848Q87.159 29.321 86.896 29.680Q86.632 30.040 86.208 30.239Q85.784 30.438 85.327 30.438Q85.073 30.438 84.796 30.364Q84.519 30.290 84.284 30.133Q84.050 29.977 83.909 29.743Q83.769 29.508 83.769 29.223Q83.769 29.055 83.884 28.932Q83.999 28.809 84.175 28.809Q84.257 28.809 84.327 28.839Q84.398 28.868 84.454 28.922Q84.511 28.977 84.542 29.053Q84.573 29.130 84.573 29.208Q84.573 29.368 84.480 29.481M88.445 30.122L88.445 30.032Q88.495 29.825 88.695 29.801L89.511 29.801L89.511 26.618Q89.132 26.926 88.679 26.926Q88.448 26.926 88.398 26.696L88.398 26.606Q88.448 26.391 88.644 26.368Q88.972 26.368 89.226 26.130Q89.480 25.891 89.620 25.544Q89.691 25.415 89.847 25.391L89.902 25.391Q90.097 25.411 90.148 25.626L90.148 29.801L90.964 29.801Q91.163 29.825 91.214 30.032L91.214 30.122Q91.163 30.337 90.964 30.360L88.695 30.360Q88.495 30.337 88.445 30.122M92.261 30.122L92.261 30.047Q92.292 29.880 92.394 29.833L93.683 28.766Q94.015 28.489 94.197 28.337Q94.378 28.184 94.573 27.964Q94.769 27.743 94.890 27.493Q95.011 27.243 95.011 26.977Q95.011 26.653 94.835 26.419Q94.659 26.184 94.380 26.069Q94.101 25.954 93.781 25.954Q93.523 25.954 93.294 26.077Q93.066 26.200 92.964 26.415Q93.066 26.547 93.066 26.696Q93.066 26.856 92.947 26.979Q92.827 27.102 92.667 27.102Q92.491 27.102 92.376 26.977Q92.261 26.852 92.261 26.680Q92.261 26.387 92.396 26.145Q92.531 25.903 92.769 25.731Q93.007 25.559 93.275 25.475Q93.542 25.391 93.835 25.391Q94.316 25.391 94.732 25.581Q95.148 25.770 95.400 26.131Q95.652 26.493 95.652 26.977Q95.652 27.321 95.519 27.624Q95.386 27.926 95.161 28.186Q94.937 28.446 94.628 28.708Q94.320 28.969 94.109 29.145L93.300 29.801L95.011 29.801L95.011 29.657Q95.062 29.446 95.261 29.422L95.402 29.422Q95.601 29.442 95.652 29.657L95.652 30.122Q95.601 30.337 95.402 30.360L92.507 30.360Q92.312 30.340 92.261 30.122\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-soft-neutral)\">\u003Cpath d=\"M-68.737 1.907h25.608v-22.762h-25.608Z\"\u002F>\u003Cg transform=\"translate(-109.525 -37.39)\">\u003Cpath d=\"M51.463 30.438Q50.822 30.438 50.436 30.061Q50.049 29.684 49.891 29.112Q49.733 28.540 49.733 27.911Q49.733 27.446 49.893 26.991Q50.053 26.536 50.342 26.176Q50.631 25.817 51.039 25.604Q51.447 25.391 51.936 25.391Q52.209 25.391 52.459 25.489Q52.709 25.587 52.862 25.782Q53.014 25.977 53.014 26.270Q53.014 26.446 52.901 26.567Q52.787 26.688 52.615 26.688Q52.440 26.688 52.322 26.575Q52.205 26.462 52.205 26.290Q52.205 26.169 52.279 26.047Q52.170 25.954 51.936 25.954Q51.518 25.954 51.182 26.194Q50.846 26.434 50.641 26.821Q50.436 27.208 50.389 27.606Q50.627 27.407 50.936 27.299Q51.244 27.192 51.565 27.192Q51.904 27.192 52.203 27.317Q52.502 27.442 52.721 27.661Q52.940 27.880 53.065 28.178Q53.190 28.477 53.190 28.817Q53.190 29.165 53.051 29.465Q52.912 29.766 52.672 29.983Q52.432 30.200 52.117 30.319Q51.803 30.438 51.463 30.438M50.479 28.895Q50.541 29.157 50.670 29.381Q50.799 29.606 50.998 29.743Q51.197 29.880 51.463 29.880Q51.916 29.880 52.233 29.573Q52.549 29.266 52.549 28.817Q52.549 28.536 52.414 28.290Q52.279 28.044 52.043 27.901Q51.807 27.758 51.510 27.758Q51.119 27.758 50.787 27.985Q50.455 28.212 50.455 28.583Q50.455 28.622 50.471 28.700Q50.487 28.778 50.487 28.817Q50.487 28.844 50.485 28.860Q50.483 28.876 50.479 28.895M56.147 29.016L54.076 29.016Q53.881 28.993 53.826 28.774L53.826 28.536Q53.826 28.462 53.869 28.391L55.717 25.520Q55.803 25.391 55.955 25.391L56.412 25.391Q56.522 25.391 56.600 25.469Q56.678 25.547 56.678 25.657L56.678 28.458L57.342 28.458Q57.537 28.481 57.588 28.688L57.588 28.774Q57.537 28.993 57.342 29.016L56.678 29.016L56.678 29.801L57.252 29.801Q57.459 29.825 57.498 30.032L57.498 30.122Q57.459 30.337 57.252 30.360L55.572 30.360Q55.365 30.337 55.322 30.122L55.322 30.032Q55.365 29.825 55.572 29.801L56.147 29.801L56.147 29.016M56.147 25.864L54.475 28.458L56.147 28.458\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-soft-neutral)\">\u003Cpath d=\"M-38.861 1.907h25.607v-22.762H-38.86Z\"\u002F>\u003Cg transform=\"translate(-79.65 -37.39)\">\u003Cpath d=\"M51.463 30.438Q50.822 30.438 50.436 30.061Q50.049 29.684 49.891 29.112Q49.733 28.540 49.733 27.911Q49.733 27.446 49.893 26.991Q50.053 26.536 50.342 26.176Q50.631 25.817 51.039 25.604Q51.447 25.391 51.936 25.391Q52.209 25.391 52.459 25.489Q52.709 25.587 52.862 25.782Q53.014 25.977 53.014 26.270Q53.014 26.446 52.901 26.567Q52.787 26.688 52.615 26.688Q52.440 26.688 52.322 26.575Q52.205 26.462 52.205 26.290Q52.205 26.169 52.279 26.047Q52.170 25.954 51.936 25.954Q51.518 25.954 51.182 26.194Q50.846 26.434 50.641 26.821Q50.436 27.208 50.389 27.606Q50.627 27.407 50.936 27.299Q51.244 27.192 51.565 27.192Q51.904 27.192 52.203 27.317Q52.502 27.442 52.721 27.661Q52.940 27.880 53.065 28.178Q53.190 28.477 53.190 28.817Q53.190 29.165 53.051 29.465Q52.912 29.766 52.672 29.983Q52.432 30.200 52.117 30.319Q51.803 30.438 51.463 30.438M50.479 28.895Q50.541 29.157 50.670 29.381Q50.799 29.606 50.998 29.743Q51.197 29.880 51.463 29.880Q51.916 29.880 52.233 29.573Q52.549 29.266 52.549 28.817Q52.549 28.536 52.414 28.290Q52.279 28.044 52.043 27.901Q51.807 27.758 51.510 27.758Q51.119 27.758 50.787 27.985Q50.455 28.212 50.455 28.583Q50.455 28.622 50.471 28.700Q50.487 28.778 50.487 28.817Q50.487 28.844 50.485 28.860Q50.483 28.876 50.479 28.895M56.147 29.016L54.076 29.016Q53.881 28.993 53.826 28.774L53.826 28.536Q53.826 28.462 53.869 28.391L55.717 25.520Q55.803 25.391 55.955 25.391L56.412 25.391Q56.522 25.391 56.600 25.469Q56.678 25.547 56.678 25.657L56.678 28.458L57.342 28.458Q57.537 28.481 57.588 28.688L57.588 28.774Q57.537 28.993 57.342 29.016L56.678 29.016L56.678 29.801L57.252 29.801Q57.459 29.825 57.498 30.032L57.498 30.122Q57.459 30.337 57.252 30.360L55.572 30.360Q55.365 30.337 55.322 30.122L55.322 30.032Q55.365 29.825 55.572 29.801L56.147 29.801L56.147 29.016M56.147 25.864L54.475 28.458L56.147 28.458\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-soft-neutral)\">\u003Cpath d=\"M-8.986 1.907h25.608v-22.762H-8.986Z\"\u002F>\u003Cg transform=\"translate(-49.774 -37.39)\">\u003Cpath d=\"M51.463 30.438Q50.822 30.438 50.436 30.061Q50.049 29.684 49.891 29.112Q49.733 28.540 49.733 27.911Q49.733 27.446 49.893 26.991Q50.053 26.536 50.342 26.176Q50.631 25.817 51.039 25.604Q51.447 25.391 51.936 25.391Q52.209 25.391 52.459 25.489Q52.709 25.587 52.862 25.782Q53.014 25.977 53.014 26.270Q53.014 26.446 52.901 26.567Q52.787 26.688 52.615 26.688Q52.440 26.688 52.322 26.575Q52.205 26.462 52.205 26.290Q52.205 26.169 52.279 26.047Q52.170 25.954 51.936 25.954Q51.518 25.954 51.182 26.194Q50.846 26.434 50.641 26.821Q50.436 27.208 50.389 27.606Q50.627 27.407 50.936 27.299Q51.244 27.192 51.565 27.192Q51.904 27.192 52.203 27.317Q52.502 27.442 52.721 27.661Q52.940 27.880 53.065 28.178Q53.190 28.477 53.190 28.817Q53.190 29.165 53.051 29.465Q52.912 29.766 52.672 29.983Q52.432 30.200 52.117 30.319Q51.803 30.438 51.463 30.438M50.479 28.895Q50.541 29.157 50.670 29.381Q50.799 29.606 50.998 29.743Q51.197 29.880 51.463 29.880Q51.916 29.880 52.233 29.573Q52.549 29.266 52.549 28.817Q52.549 28.536 52.414 28.290Q52.279 28.044 52.043 27.901Q51.807 27.758 51.510 27.758Q51.119 27.758 50.787 27.985Q50.455 28.212 50.455 28.583Q50.455 28.622 50.471 28.700Q50.487 28.778 50.487 28.817Q50.487 28.844 50.485 28.860Q50.483 28.876 50.479 28.895M56.147 29.016L54.076 29.016Q53.881 28.993 53.826 28.774L53.826 28.536Q53.826 28.462 53.869 28.391L55.717 25.520Q55.803 25.391 55.955 25.391L56.412 25.391Q56.522 25.391 56.600 25.469Q56.678 25.547 56.678 25.657L56.678 28.458L57.342 28.458Q57.537 28.481 57.588 28.688L57.588 28.774Q57.537 28.993 57.342 29.016L56.678 29.016L56.678 29.801L57.252 29.801Q57.459 29.825 57.498 30.032L57.498 30.122Q57.459 30.337 57.252 30.360L55.572 30.360Q55.365 30.337 55.322 30.122L55.322 30.032Q55.365 29.825 55.572 29.801L56.147 29.801L56.147 29.016M56.147 25.864L54.475 28.458L56.147 28.458\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-soft-neutral)\">\u003Cpath d=\"M20.89 1.907h25.607v-22.762H20.89Z\"\u002F>\u003Cg transform=\"translate(-19.899 -37.39)\">\u003Cpath d=\"M51.463 30.438Q50.822 30.438 50.436 30.061Q50.049 29.684 49.891 29.112Q49.733 28.540 49.733 27.911Q49.733 27.446 49.893 26.991Q50.053 26.536 50.342 26.176Q50.631 25.817 51.039 25.604Q51.447 25.391 51.936 25.391Q52.209 25.391 52.459 25.489Q52.709 25.587 52.862 25.782Q53.014 25.977 53.014 26.270Q53.014 26.446 52.901 26.567Q52.787 26.688 52.615 26.688Q52.440 26.688 52.322 26.575Q52.205 26.462 52.205 26.290Q52.205 26.169 52.279 26.047Q52.170 25.954 51.936 25.954Q51.518 25.954 51.182 26.194Q50.846 26.434 50.641 26.821Q50.436 27.208 50.389 27.606Q50.627 27.407 50.936 27.299Q51.244 27.192 51.565 27.192Q51.904 27.192 52.203 27.317Q52.502 27.442 52.721 27.661Q52.940 27.880 53.065 28.178Q53.190 28.477 53.190 28.817Q53.190 29.165 53.051 29.465Q52.912 29.766 52.672 29.983Q52.432 30.200 52.117 30.319Q51.803 30.438 51.463 30.438M50.479 28.895Q50.541 29.157 50.670 29.381Q50.799 29.606 50.998 29.743Q51.197 29.880 51.463 29.880Q51.916 29.880 52.233 29.573Q52.549 29.266 52.549 28.817Q52.549 28.536 52.414 28.290Q52.279 28.044 52.043 27.901Q51.807 27.758 51.510 27.758Q51.119 27.758 50.787 27.985Q50.455 28.212 50.455 28.583Q50.455 28.622 50.471 28.700Q50.487 28.778 50.487 28.817Q50.487 28.844 50.485 28.860Q50.483 28.876 50.479 28.895M56.147 29.016L54.076 29.016Q53.881 28.993 53.826 28.774L53.826 28.536Q53.826 28.462 53.869 28.391L55.717 25.520Q55.803 25.391 55.955 25.391L56.412 25.391Q56.522 25.391 56.600 25.469Q56.678 25.547 56.678 25.657L56.678 28.458L57.342 28.458Q57.537 28.481 57.588 28.688L57.588 28.774Q57.537 28.993 57.342 29.016L56.678 29.016L56.678 29.801L57.252 29.801Q57.459 29.825 57.498 30.032L57.498 30.122Q57.459 30.337 57.252 30.360L55.572 30.360Q55.365 30.337 55.322 30.122L55.322 30.032Q55.365 29.825 55.572 29.801L56.147 29.801L56.147 29.016M56.147 25.864L54.475 28.458L56.147 28.458\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-soft-neutral)\">\u003Cpath d=\"M50.765 1.907h25.608v-22.762H50.765Z\"\u002F>\u003Cg transform=\"translate(9.977 -37.39)\">\u003Cpath d=\"M51.463 30.438Q50.822 30.438 50.436 30.061Q50.049 29.684 49.891 29.112Q49.733 28.540 49.733 27.911Q49.733 27.446 49.893 26.991Q50.053 26.536 50.342 26.176Q50.631 25.817 51.039 25.604Q51.447 25.391 51.936 25.391Q52.209 25.391 52.459 25.489Q52.709 25.587 52.862 25.782Q53.014 25.977 53.014 26.270Q53.014 26.446 52.901 26.567Q52.787 26.688 52.615 26.688Q52.440 26.688 52.322 26.575Q52.205 26.462 52.205 26.290Q52.205 26.169 52.279 26.047Q52.170 25.954 51.936 25.954Q51.518 25.954 51.182 26.194Q50.846 26.434 50.641 26.821Q50.436 27.208 50.389 27.606Q50.627 27.407 50.936 27.299Q51.244 27.192 51.565 27.192Q51.904 27.192 52.203 27.317Q52.502 27.442 52.721 27.661Q52.940 27.880 53.065 28.178Q53.190 28.477 53.190 28.817Q53.190 29.165 53.051 29.465Q52.912 29.766 52.672 29.983Q52.432 30.200 52.117 30.319Q51.803 30.438 51.463 30.438M50.479 28.895Q50.541 29.157 50.670 29.381Q50.799 29.606 50.998 29.743Q51.197 29.880 51.463 29.880Q51.916 29.880 52.233 29.573Q52.549 29.266 52.549 28.817Q52.549 28.536 52.414 28.290Q52.279 28.044 52.043 27.901Q51.807 27.758 51.510 27.758Q51.119 27.758 50.787 27.985Q50.455 28.212 50.455 28.583Q50.455 28.622 50.471 28.700Q50.487 28.778 50.487 28.817Q50.487 28.844 50.485 28.860Q50.483 28.876 50.479 28.895M56.147 29.016L54.076 29.016Q53.881 28.993 53.826 28.774L53.826 28.536Q53.826 28.462 53.869 28.391L55.717 25.520Q55.803 25.391 55.955 25.391L56.412 25.391Q56.522 25.391 56.600 25.469Q56.678 25.547 56.678 25.657L56.678 28.458L57.342 28.458Q57.537 28.481 57.588 28.688L57.588 28.774Q57.537 28.993 57.342 29.016L56.678 29.016L56.678 29.801L57.252 29.801Q57.459 29.825 57.498 30.032L57.498 30.122Q57.459 30.337 57.252 30.360L55.572 30.360Q55.365 30.337 55.322 30.122L55.322 30.032Q55.365 29.825 55.572 29.801L56.147 29.801L56.147 29.016M56.147 25.864L54.475 28.458L56.147 28.458\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-soft-neutral)\">\u003Cpath d=\"M80.64 1.907h25.608v-22.762H80.641Z\"\u002F>\u003Cg transform=\"translate(39.852 -37.39)\">\u003Cpath d=\"M51.463 30.438Q50.822 30.438 50.436 30.061Q50.049 29.684 49.891 29.112Q49.733 28.540 49.733 27.911Q49.733 27.446 49.893 26.991Q50.053 26.536 50.342 26.176Q50.631 25.817 51.039 25.604Q51.447 25.391 51.936 25.391Q52.209 25.391 52.459 25.489Q52.709 25.587 52.862 25.782Q53.014 25.977 53.014 26.270Q53.014 26.446 52.901 26.567Q52.787 26.688 52.615 26.688Q52.440 26.688 52.322 26.575Q52.205 26.462 52.205 26.290Q52.205 26.169 52.279 26.047Q52.170 25.954 51.936 25.954Q51.518 25.954 51.182 26.194Q50.846 26.434 50.641 26.821Q50.436 27.208 50.389 27.606Q50.627 27.407 50.936 27.299Q51.244 27.192 51.565 27.192Q51.904 27.192 52.203 27.317Q52.502 27.442 52.721 27.661Q52.940 27.880 53.065 28.178Q53.190 28.477 53.190 28.817Q53.190 29.165 53.051 29.465Q52.912 29.766 52.672 29.983Q52.432 30.200 52.117 30.319Q51.803 30.438 51.463 30.438M50.479 28.895Q50.541 29.157 50.670 29.381Q50.799 29.606 50.998 29.743Q51.197 29.880 51.463 29.880Q51.916 29.880 52.233 29.573Q52.549 29.266 52.549 28.817Q52.549 28.536 52.414 28.290Q52.279 28.044 52.043 27.901Q51.807 27.758 51.510 27.758Q51.119 27.758 50.787 27.985Q50.455 28.212 50.455 28.583Q50.455 28.622 50.471 28.700Q50.487 28.778 50.487 28.817Q50.487 28.844 50.485 28.860Q50.483 28.876 50.479 28.895M56.147 29.016L54.076 29.016Q53.881 28.993 53.826 28.774L53.826 28.536Q53.826 28.462 53.869 28.391L55.717 25.520Q55.803 25.391 55.955 25.391L56.412 25.391Q56.522 25.391 56.600 25.469Q56.678 25.547 56.678 25.657L56.678 28.458L57.342 28.458Q57.537 28.481 57.588 28.688L57.588 28.774Q57.537 28.993 57.342 29.016L56.678 29.016L56.678 29.801L57.252 29.801Q57.459 29.825 57.498 30.032L57.498 30.122Q57.459 30.337 57.252 30.360L55.572 30.360Q55.365 30.337 55.322 30.122L55.322 30.032Q55.365 29.825 55.572 29.801L56.147 29.801L56.147 29.016M56.147 25.864L54.475 28.458L56.147 28.458\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-soft-neutral)\">\u003Cpath d=\"M110.516 1.907h25.608v-22.762h-25.608Z\"\u002F>\u003Cg transform=\"translate(69.728 -37.39)\">\u003Cpath d=\"M51.463 30.438Q50.822 30.438 50.436 30.061Q50.049 29.684 49.891 29.112Q49.733 28.540 49.733 27.911Q49.733 27.446 49.893 26.991Q50.053 26.536 50.342 26.176Q50.631 25.817 51.039 25.604Q51.447 25.391 51.936 25.391Q52.209 25.391 52.459 25.489Q52.709 25.587 52.862 25.782Q53.014 25.977 53.014 26.270Q53.014 26.446 52.901 26.567Q52.787 26.688 52.615 26.688Q52.440 26.688 52.322 26.575Q52.205 26.462 52.205 26.290Q52.205 26.169 52.279 26.047Q52.170 25.954 51.936 25.954Q51.518 25.954 51.182 26.194Q50.846 26.434 50.641 26.821Q50.436 27.208 50.389 27.606Q50.627 27.407 50.936 27.299Q51.244 27.192 51.565 27.192Q51.904 27.192 52.203 27.317Q52.502 27.442 52.721 27.661Q52.940 27.880 53.065 28.178Q53.190 28.477 53.190 28.817Q53.190 29.165 53.051 29.465Q52.912 29.766 52.672 29.983Q52.432 30.200 52.117 30.319Q51.803 30.438 51.463 30.438M50.479 28.895Q50.541 29.157 50.670 29.381Q50.799 29.606 50.998 29.743Q51.197 29.880 51.463 29.880Q51.916 29.880 52.233 29.573Q52.549 29.266 52.549 28.817Q52.549 28.536 52.414 28.290Q52.279 28.044 52.043 27.901Q51.807 27.758 51.510 27.758Q51.119 27.758 50.787 27.985Q50.455 28.212 50.455 28.583Q50.455 28.622 50.471 28.700Q50.487 28.778 50.487 28.817Q50.487 28.844 50.485 28.860Q50.483 28.876 50.479 28.895M56.147 29.016L54.076 29.016Q53.881 28.993 53.826 28.774L53.826 28.536Q53.826 28.462 53.869 28.391L55.717 25.520Q55.803 25.391 55.955 25.391L56.412 25.391Q56.522 25.391 56.600 25.469Q56.678 25.547 56.678 25.657L56.678 28.458L57.342 28.458Q57.537 28.481 57.588 28.688L57.588 28.774Q57.537 28.993 57.342 29.016L56.678 29.016L56.678 29.801L57.252 29.801Q57.459 29.825 57.498 30.032L57.498 30.122Q57.459 30.337 57.252 30.360L55.572 30.360Q55.365 30.337 55.322 30.122L55.322 30.032Q55.365 29.825 55.572 29.801L56.147 29.801L56.147 29.016M56.147 25.864L54.475 28.458L56.147 28.458\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-soft-neutral)\">\u003Cpath d=\"M140.392 1.907h25.607v-22.762h-25.607Z\"\u002F>\u003Cg transform=\"translate(99.603 -37.39)\">\u003Cpath d=\"M51.463 30.438Q50.822 30.438 50.436 30.061Q50.049 29.684 49.891 29.112Q49.733 28.540 49.733 27.911Q49.733 27.446 49.893 26.991Q50.053 26.536 50.342 26.176Q50.631 25.817 51.039 25.604Q51.447 25.391 51.936 25.391Q52.209 25.391 52.459 25.489Q52.709 25.587 52.862 25.782Q53.014 25.977 53.014 26.270Q53.014 26.446 52.901 26.567Q52.787 26.688 52.615 26.688Q52.440 26.688 52.322 26.575Q52.205 26.462 52.205 26.290Q52.205 26.169 52.279 26.047Q52.170 25.954 51.936 25.954Q51.518 25.954 51.182 26.194Q50.846 26.434 50.641 26.821Q50.436 27.208 50.389 27.606Q50.627 27.407 50.936 27.299Q51.244 27.192 51.565 27.192Q51.904 27.192 52.203 27.317Q52.502 27.442 52.721 27.661Q52.940 27.880 53.065 28.178Q53.190 28.477 53.190 28.817Q53.190 29.165 53.051 29.465Q52.912 29.766 52.672 29.983Q52.432 30.200 52.117 30.319Q51.803 30.438 51.463 30.438M50.479 28.895Q50.541 29.157 50.670 29.381Q50.799 29.606 50.998 29.743Q51.197 29.880 51.463 29.880Q51.916 29.880 52.233 29.573Q52.549 29.266 52.549 28.817Q52.549 28.536 52.414 28.290Q52.279 28.044 52.043 27.901Q51.807 27.758 51.510 27.758Q51.119 27.758 50.787 27.985Q50.455 28.212 50.455 28.583Q50.455 28.622 50.471 28.700Q50.487 28.778 50.487 28.817Q50.487 28.844 50.485 28.860Q50.483 28.876 50.479 28.895M56.147 29.016L54.076 29.016Q53.881 28.993 53.826 28.774L53.826 28.536Q53.826 28.462 53.869 28.391L55.717 25.520Q55.803 25.391 55.955 25.391L56.412 25.391Q56.522 25.391 56.600 25.469Q56.678 25.547 56.678 25.657L56.678 28.458L57.342 28.458Q57.537 28.481 57.588 28.688L57.588 28.774Q57.537 28.993 57.342 29.016L56.678 29.016L56.678 29.801L57.252 29.801Q57.459 29.825 57.498 30.032L57.498 30.122Q57.459 30.337 57.252 30.360L55.572 30.360Q55.365 30.337 55.322 30.122L55.322 30.032Q55.365 29.825 55.572 29.801L56.147 29.801L56.147 29.016M56.147 25.864L54.475 28.458L56.147 28.458\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"none\" font-family=\"cmtt8\" font-size=\"8\">\u003Cg transform=\"translate(117.344 -37.945)\">\u003Cpath d=\"M49.694 28.946Q49.694 28.661 49.850 28.407Q50.006 28.153 50.256 27.979Q50.506 27.805 50.791 27.719Q50.553 27.645 50.326 27.503Q50.100 27.360 49.957 27.151Q49.815 26.942 49.815 26.696Q49.815 26.297 50.061 26.003Q50.307 25.708 50.690 25.549Q51.072 25.391 51.463 25.391Q51.854 25.391 52.237 25.549Q52.619 25.708 52.865 26.006Q53.112 26.305 53.112 26.696Q53.112 26.942 52.969 27.149Q52.826 27.356 52.600 27.501Q52.373 27.645 52.135 27.719Q52.588 27.856 52.908 28.182Q53.229 28.508 53.229 28.946Q53.229 29.383 52.971 29.727Q52.713 30.071 52.303 30.255Q51.893 30.438 51.463 30.438Q51.033 30.438 50.623 30.256Q50.213 30.075 49.953 29.731Q49.694 29.387 49.694 28.946M50.334 28.946Q50.334 29.219 50.500 29.434Q50.666 29.649 50.928 29.764Q51.190 29.880 51.463 29.880Q51.737 29.880 51.996 29.764Q52.256 29.649 52.422 29.432Q52.588 29.215 52.588 28.946Q52.588 28.669 52.422 28.452Q52.256 28.235 51.996 28.118Q51.737 28.001 51.463 28.001Q51.190 28.001 50.928 28.118Q50.666 28.235 50.500 28.450Q50.334 28.665 50.334 28.946M50.455 26.696Q50.455 26.934 50.612 27.102Q50.768 27.270 51.004 27.354Q51.240 27.438 51.463 27.438Q51.682 27.438 51.920 27.354Q52.158 27.270 52.315 27.100Q52.471 26.930 52.471 26.696Q52.471 26.462 52.315 26.292Q52.158 26.122 51.920 26.038Q51.682 25.954 51.463 25.954Q51.240 25.954 51.004 26.038Q50.768 26.122 50.612 26.290Q50.455 26.458 50.455 26.696\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(117.344 -37.945)\">\u003Cpath d=\"M57.912 30.122L57.912 30.032Q57.955 29.825 58.162 29.801L58.584 29.801L58.584 26.032L58.162 26.032Q57.955 26.008 57.912 25.794L57.912 25.704Q57.955 25.497 58.162 25.473L58.979 25.473Q59.174 25.497 59.225 25.704L59.225 27.255Q59.686 26.872 60.283 26.872Q60.631 26.872 60.869 27.012Q61.108 27.153 61.223 27.411Q61.338 27.669 61.338 28.024L61.338 29.801L61.764 29.801Q61.971 29.825 62.010 30.032L62.010 30.122Q61.971 30.337 61.764 30.360L60.369 30.360Q60.174 30.337 60.123 30.122L60.123 30.032Q60.174 29.821 60.369 29.801L60.697 29.801L60.697 28.055Q60.697 27.747 60.608 27.589Q60.518 27.430 60.225 27.430Q59.955 27.430 59.727 27.561Q59.498 27.692 59.362 27.921Q59.225 28.149 59.225 28.415L59.225 29.801L59.651 29.801Q59.858 29.825 59.897 30.032L59.897 30.122Q59.858 30.337 59.651 30.360L58.162 30.360Q57.955 30.337 57.912 30.122M65.600 28.872L63.158 28.872Q63.213 29.149 63.410 29.372Q63.608 29.594 63.885 29.717Q64.162 29.840 64.447 29.840Q64.920 29.840 65.143 29.551Q65.151 29.540 65.207 29.434Q65.264 29.329 65.313 29.286Q65.362 29.243 65.455 29.231L65.600 29.231Q65.791 29.251 65.850 29.465L65.850 29.520Q65.783 29.821 65.553 30.018Q65.322 30.215 65.010 30.307Q64.697 30.399 64.393 30.399Q63.908 30.399 63.469 30.171Q63.029 29.942 62.762 29.542Q62.494 29.141 62.494 28.649L62.494 28.590Q62.494 28.122 62.740 27.719Q62.987 27.317 63.395 27.083Q63.803 26.848 64.272 26.848Q64.776 26.848 65.129 27.071Q65.483 27.294 65.666 27.682Q65.850 28.071 65.850 28.575L65.850 28.633Q65.791 28.848 65.600 28.872M63.166 28.321L65.194 28.321Q65.147 27.911 64.908 27.659Q64.670 27.407 64.272 27.407Q63.877 27.407 63.571 27.669Q63.264 27.930 63.166 28.321M66.740 29.247Q66.740 28.801 67.154 28.544Q67.569 28.286 68.110 28.186Q68.651 28.087 69.158 28.079Q69.158 27.864 69.024 27.712Q68.889 27.559 68.682 27.483Q68.475 27.407 68.264 27.407Q67.920 27.407 67.760 27.430L67.760 27.489Q67.760 27.657 67.641 27.772Q67.522 27.887 67.358 27.887Q67.182 27.887 67.067 27.764Q66.951 27.641 66.951 27.473Q66.951 27.067 67.332 26.958Q67.713 26.848 68.272 26.848Q68.541 26.848 68.809 26.926Q69.076 27.005 69.301 27.155Q69.526 27.305 69.662 27.526Q69.799 27.747 69.799 28.024L69.799 29.743Q69.799 29.801 70.326 29.801Q70.522 29.821 70.572 30.032L70.572 30.122Q70.522 30.337 70.326 30.360L70.182 30.360Q69.838 30.360 69.610 30.313Q69.381 30.266 69.237 30.079Q68.776 30.399 68.069 30.399Q67.733 30.399 67.428 30.258Q67.123 30.118 66.932 29.856Q66.740 29.594 66.740 29.247M67.381 29.255Q67.381 29.528 67.623 29.684Q67.865 29.840 68.151 29.840Q68.369 29.840 68.602 29.782Q68.834 29.723 68.996 29.585Q69.158 29.446 69.158 29.223L69.158 28.633Q68.877 28.633 68.461 28.690Q68.045 28.747 67.713 28.885Q67.381 29.024 67.381 29.255M72.444 30.399Q71.979 30.399 71.613 30.149Q71.248 29.899 71.043 29.495Q70.838 29.090 70.838 28.633Q70.838 28.290 70.963 27.969Q71.088 27.649 71.321 27.399Q71.553 27.149 71.858 27.010Q72.162 26.872 72.518 26.872Q73.029 26.872 73.436 27.192L73.436 26.032L73.014 26.032Q72.803 26.008 72.764 25.794L72.764 25.704Q72.803 25.497 73.014 25.473L73.826 25.473Q74.026 25.497 74.076 25.704L74.076 29.801L74.502 29.801Q74.709 29.825 74.748 30.032L74.748 30.122Q74.709 30.337 74.502 30.360L73.686 30.360Q73.487 30.337 73.436 30.122L73.436 29.993Q73.240 30.184 72.979 30.292Q72.717 30.399 72.444 30.399M72.483 29.840Q72.842 29.840 73.094 29.571Q73.346 29.301 73.436 28.926L73.436 28.094Q73.377 27.907 73.250 27.756Q73.123 27.606 72.946 27.518Q72.768 27.430 72.572 27.430Q72.264 27.430 72.014 27.600Q71.764 27.770 71.619 28.055Q71.475 28.340 71.475 28.641Q71.475 29.090 71.760 29.465Q72.045 29.840 72.483 29.840M75.412 30.161L75.412 29.247Q75.440 29.040 75.651 29.016L75.819 29.016Q75.983 29.040 76.041 29.200Q76.244 29.840 76.971 29.840Q77.178 29.840 77.406 29.805Q77.635 29.770 77.803 29.655Q77.971 29.540 77.971 29.337Q77.971 29.126 77.748 29.012Q77.526 28.899 77.252 28.856L76.553 28.743Q75.412 28.532 75.412 27.809Q75.412 27.520 75.557 27.331Q75.701 27.141 75.942 27.034Q76.182 26.926 76.438 26.887Q76.694 26.848 76.971 26.848Q77.221 26.848 77.414 26.878Q77.608 26.907 77.772 26.985Q77.850 26.868 77.979 26.848L78.057 26.848Q78.154 26.860 78.217 26.922Q78.279 26.985 78.291 27.079L78.291 27.786Q78.279 27.880 78.217 27.946Q78.154 28.012 78.057 28.024L77.889 28.024Q77.795 28.012 77.729 27.946Q77.662 27.880 77.651 27.786Q77.651 27.407 76.955 27.407Q76.608 27.407 76.289 27.489Q75.971 27.571 75.971 27.817Q75.971 28.083 76.643 28.192L77.346 28.313Q77.830 28.395 78.180 28.643Q78.529 28.891 78.529 29.337Q78.529 29.727 78.293 29.969Q78.057 30.212 77.707 30.305Q77.358 30.399 76.971 30.399Q76.393 30.399 75.994 30.145Q75.924 30.270 75.875 30.327Q75.826 30.383 75.721 30.399L75.651 30.399Q75.436 30.376 75.412 30.161M80.791 31.473Q80.678 31.473 80.588 31.383Q80.498 31.294 80.498 31.184Q80.498 31.008 80.690 30.934Q80.908 30.883 81.080 30.725Q81.252 30.567 81.319 30.344Q81.295 30.344 81.264 30.360L81.201 30.360Q80.971 30.360 80.805 30.198Q80.639 30.036 80.639 29.801Q80.639 29.567 80.803 29.407Q80.967 29.247 81.201 29.247Q81.412 29.247 81.576 29.368Q81.740 29.489 81.830 29.682Q81.920 29.876 81.920 30.087Q81.920 30.563 81.621 30.952Q81.322 31.340 80.858 31.465Q80.834 31.473 80.791 31.473\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(117.344 -37.945)\">\u003Cpath d=\"M87.663 30.122L87.663 30.032Q87.706 29.825 87.913 29.801L88.335 29.801L88.335 27.473L87.913 27.473Q87.706 27.450 87.663 27.231L87.663 27.145Q87.710 26.934 87.913 26.911L88.730 26.911Q88.925 26.934 88.976 27.145L88.976 27.231L88.968 27.255Q89.195 27.075 89.468 26.973Q89.741 26.872 90.034 26.872Q90.382 26.872 90.620 27.012Q90.859 27.153 90.974 27.411Q91.089 27.669 91.089 28.024L91.089 29.801L91.515 29.801Q91.722 29.825 91.761 30.032L91.761 30.122Q91.722 30.337 91.515 30.360L90.120 30.360Q89.925 30.337 89.874 30.122L89.874 30.032Q89.925 29.821 90.120 29.801L90.448 29.801L90.448 28.055Q90.448 27.747 90.359 27.589Q90.269 27.430 89.976 27.430Q89.706 27.430 89.478 27.561Q89.249 27.692 89.113 27.921Q88.976 28.149 88.976 28.415L88.976 29.801L89.402 29.801Q89.609 29.825 89.648 30.032L89.648 30.122Q89.609 30.337 89.402 30.360L87.913 30.360Q87.706 30.337 87.663 30.122\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(117.344 -37.945)\">\u003Cpath d=\"M96.351 30.122L96.351 30.032Q96.390 29.825 96.597 29.801L97.003 29.801L97.933 28.583L97.062 27.473L96.644 27.473Q96.448 27.454 96.398 27.231L96.398 27.145Q96.448 26.934 96.644 26.911L97.804 26.911Q98.003 26.934 98.054 27.145L98.054 27.231Q98.003 27.450 97.804 27.473L97.695 27.473L98.191 28.130L98.667 27.473L98.550 27.473Q98.351 27.450 98.300 27.231L98.300 27.145Q98.351 26.934 98.550 26.911L99.710 26.911Q99.906 26.930 99.956 27.145L99.956 27.231Q99.906 27.450 99.710 27.473L99.300 27.473L98.452 28.583L99.413 29.801L99.820 29.801Q100.019 29.825 100.070 30.032L100.070 30.122Q100.031 30.337 99.820 30.360L98.667 30.360Q98.460 30.337 98.421 30.122L98.421 30.032Q98.460 29.825 98.667 29.801L98.796 29.801L98.191 28.946L97.605 29.801L97.749 29.801Q97.956 29.825 97.995 30.032L97.995 30.122Q97.956 30.337 97.749 30.360L96.597 30.360Q96.402 30.340 96.351 30.122\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(117.344 -37.945)\">\u003Cpath d=\"M106.714 30.438Q106.073 30.438 105.687 30.061Q105.300 29.684 105.142 29.112Q104.984 28.540 104.984 27.911Q104.984 27.446 105.144 26.991Q105.304 26.536 105.593 26.176Q105.882 25.817 106.290 25.604Q106.698 25.391 107.187 25.391Q107.460 25.391 107.710 25.489Q107.960 25.587 108.113 25.782Q108.265 25.977 108.265 26.270Q108.265 26.446 108.152 26.567Q108.038 26.688 107.866 26.688Q107.691 26.688 107.573 26.575Q107.456 26.462 107.456 26.290Q107.456 26.169 107.531 26.047Q107.421 25.954 107.187 25.954Q106.769 25.954 106.433 26.194Q106.097 26.434 105.892 26.821Q105.687 27.208 105.640 27.606Q105.878 27.407 106.187 27.299Q106.495 27.192 106.816 27.192Q107.156 27.192 107.454 27.317Q107.753 27.442 107.972 27.661Q108.191 27.880 108.316 28.178Q108.441 28.477 108.441 28.817Q108.441 29.165 108.302 29.465Q108.163 29.766 107.923 29.983Q107.683 30.200 107.368 30.319Q107.054 30.438 106.714 30.438M105.730 28.895Q105.792 29.157 105.921 29.381Q106.050 29.606 106.249 29.743Q106.448 29.880 106.714 29.880Q107.167 29.880 107.484 29.573Q107.800 29.266 107.800 28.817Q107.800 28.536 107.665 28.290Q107.531 28.044 107.294 27.901Q107.058 27.758 106.761 27.758Q106.370 27.758 106.038 27.985Q105.706 28.212 105.706 28.583Q105.706 28.622 105.722 28.700Q105.738 28.778 105.738 28.817Q105.738 28.844 105.736 28.860Q105.734 28.876 105.730 28.895M111.398 29.016L109.327 29.016Q109.132 28.993 109.077 28.774L109.077 28.536Q109.077 28.462 109.120 28.391L110.968 25.520Q111.054 25.391 111.206 25.391L111.663 25.391Q111.773 25.391 111.851 25.469Q111.929 25.547 111.929 25.657L111.929 28.458L112.593 28.458Q112.788 28.481 112.839 28.688L112.839 28.774Q112.788 28.993 112.593 29.016L111.929 29.016L111.929 29.801L112.503 29.801Q112.710 29.825 112.749 30.032L112.749 30.122Q112.710 30.337 112.503 30.360L110.823 30.360Q110.616 30.337 110.573 30.122L110.573 30.032Q110.616 29.825 110.823 29.801L111.398 29.801L111.398 29.016M111.398 25.864L109.726 28.458L111.398 28.458\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-soft-accent)\" stroke=\"var(--tk-accent)\">\u003Cpath d=\"M.972 70.194h96.74V47.43H.972Z\"\u002F>\u003Cg fill=\"currentColor\" stroke=\"none\" font-family=\"cmtt8\" font-size=\"8\">\u003Cg transform=\"translate(-36.126 30.897)\">\u003Cpath d=\"M49.901 28.633Q49.901 28.153 50.145 27.739Q50.389 27.325 50.805 27.087Q51.221 26.848 51.701 26.848Q52.256 26.848 52.635 26.958Q53.014 27.067 53.014 27.473Q53.014 27.641 52.901 27.764Q52.787 27.887 52.615 27.887Q52.444 27.887 52.324 27.772Q52.205 27.657 52.205 27.489L52.205 27.430Q52.045 27.407 51.709 27.407Q51.381 27.407 51.113 27.577Q50.846 27.747 50.694 28.030Q50.541 28.313 50.541 28.633Q50.541 28.954 50.713 29.235Q50.885 29.516 51.170 29.678Q51.455 29.840 51.783 29.840Q52.096 29.840 52.223 29.737Q52.350 29.633 52.467 29.444Q52.584 29.255 52.701 29.239L52.869 29.239Q52.975 29.251 53.039 29.319Q53.104 29.387 53.104 29.489Q53.104 29.536 53.084 29.575Q52.975 29.868 52.772 30.047Q52.569 30.227 52.293 30.313Q52.018 30.399 51.701 30.399Q51.217 30.399 50.801 30.161Q50.385 29.922 50.143 29.520Q49.901 29.118 49.901 28.633M55.709 30.399Q55.237 30.399 54.852 30.155Q54.467 29.911 54.244 29.501Q54.022 29.090 54.022 28.633Q54.022 28.290 54.147 27.967Q54.272 27.645 54.502 27.391Q54.733 27.137 55.039 26.993Q55.346 26.848 55.709 26.848Q56.072 26.848 56.385 26.995Q56.697 27.141 56.920 27.387Q57.143 27.633 57.270 27.954Q57.397 28.274 57.397 28.633Q57.397 29.090 57.172 29.503Q56.947 29.915 56.563 30.157Q56.178 30.399 55.709 30.399M55.709 29.840Q56.174 29.840 56.465 29.446Q56.756 29.051 56.756 28.567Q56.756 28.274 56.621 28.006Q56.487 27.739 56.246 27.573Q56.006 27.407 55.709 27.407Q55.404 27.407 55.166 27.573Q54.928 27.739 54.793 28.006Q54.658 28.274 54.658 28.567Q54.658 29.047 54.951 29.444Q55.244 29.840 55.709 29.840M57.904 30.122L57.904 30.032Q57.947 29.825 58.154 29.801L58.576 29.801L58.576 27.473L58.154 27.473Q57.947 27.450 57.904 27.231L57.904 27.145Q57.951 26.934 58.154 26.911L58.971 26.911Q59.166 26.934 59.217 27.145L59.217 27.231L59.209 27.255Q59.436 27.075 59.709 26.973Q59.983 26.872 60.276 26.872Q60.623 26.872 60.862 27.012Q61.100 27.153 61.215 27.411Q61.330 27.669 61.330 28.024L61.330 29.801L61.756 29.801Q61.963 29.825 62.002 30.032L62.002 30.122Q61.963 30.337 61.756 30.360L60.362 30.360Q60.166 30.337 60.115 30.122L60.115 30.032Q60.166 29.821 60.362 29.801L60.690 29.801L60.690 28.055Q60.690 27.747 60.600 27.589Q60.510 27.430 60.217 27.430Q59.947 27.430 59.719 27.561Q59.490 27.692 59.354 27.921Q59.217 28.149 59.217 28.415L59.217 29.801L59.643 29.801Q59.850 29.825 59.889 30.032L59.889 30.122Q59.850 30.337 59.643 30.360L58.154 30.360Q57.947 30.337 57.904 30.122M62.639 28.633Q62.639 28.153 62.883 27.739Q63.127 27.325 63.543 27.087Q63.959 26.848 64.440 26.848Q64.994 26.848 65.373 26.958Q65.752 27.067 65.752 27.473Q65.752 27.641 65.639 27.764Q65.526 27.887 65.354 27.887Q65.182 27.887 65.063 27.772Q64.944 27.657 64.944 27.489L64.944 27.430Q64.783 27.407 64.447 27.407Q64.119 27.407 63.852 27.577Q63.584 27.747 63.432 28.030Q63.279 28.313 63.279 28.633Q63.279 28.954 63.451 29.235Q63.623 29.516 63.908 29.678Q64.194 29.840 64.522 29.840Q64.834 29.840 64.961 29.737Q65.088 29.633 65.205 29.444Q65.322 29.255 65.440 29.239L65.608 29.239Q65.713 29.251 65.778 29.319Q65.842 29.387 65.842 29.489Q65.842 29.536 65.822 29.575Q65.713 29.868 65.510 30.047Q65.307 30.227 65.031 30.313Q64.756 30.399 64.440 30.399Q63.955 30.399 63.539 30.161Q63.123 29.922 62.881 29.520Q62.639 29.118 62.639 28.633M66.733 29.247Q66.733 28.801 67.147 28.544Q67.561 28.286 68.102 28.186Q68.643 28.087 69.151 28.079Q69.151 27.864 69.016 27.712Q68.881 27.559 68.674 27.483Q68.467 27.407 68.256 27.407Q67.912 27.407 67.752 27.430L67.752 27.489Q67.752 27.657 67.633 27.772Q67.514 27.887 67.350 27.887Q67.174 27.887 67.059 27.764Q66.944 27.641 66.944 27.473Q66.944 27.067 67.324 26.958Q67.705 26.848 68.264 26.848Q68.533 26.848 68.801 26.926Q69.069 27.005 69.293 27.155Q69.518 27.305 69.654 27.526Q69.791 27.747 69.791 28.024L69.791 29.743Q69.791 29.801 70.319 29.801Q70.514 29.821 70.565 30.032L70.565 30.122Q70.514 30.337 70.319 30.360L70.174 30.360Q69.830 30.360 69.602 30.313Q69.373 30.266 69.229 30.079Q68.768 30.399 68.061 30.399Q67.725 30.399 67.420 30.258Q67.115 30.118 66.924 29.856Q66.733 29.594 66.733 29.247M67.373 29.255Q67.373 29.528 67.615 29.684Q67.858 29.840 68.143 29.840Q68.362 29.840 68.594 29.782Q68.826 29.723 68.988 29.585Q69.151 29.446 69.151 29.223L69.151 28.633Q68.869 28.633 68.453 28.690Q68.037 28.747 67.705 28.885Q67.373 29.024 67.373 29.255M71.772 29.255L71.772 27.473L71.022 27.473Q70.822 27.450 70.772 27.231L70.772 27.145Q70.822 26.934 71.022 26.911L71.772 26.911L71.772 26.161Q71.822 25.954 72.022 25.926L72.166 25.926Q72.362 25.954 72.412 26.161L72.412 26.911L73.772 26.911Q73.963 26.930 74.022 27.145L74.022 27.231Q73.967 27.450 73.772 27.473L72.412 27.473L72.412 29.223Q72.412 29.840 72.987 29.840Q73.237 29.840 73.401 29.655Q73.565 29.469 73.565 29.223Q73.565 29.130 73.637 29.059Q73.709 28.989 73.811 28.977L73.955 28.977Q74.154 29.001 74.205 29.208L74.205 29.255Q74.205 29.579 74.022 29.842Q73.838 30.106 73.545 30.253Q73.252 30.399 72.932 30.399Q72.420 30.399 72.096 30.089Q71.772 29.778 71.772 29.255\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-36.126 30.897)\">\u003Cpath d=\"M80.659 29.801Q80.659 29.579 80.825 29.413Q80.991 29.247 81.222 29.247Q81.370 29.247 81.497 29.325Q81.624 29.403 81.698 29.528Q81.773 29.653 81.773 29.801Q81.773 30.028 81.607 30.194Q81.441 30.360 81.222 30.360Q80.995 30.360 80.827 30.192Q80.659 30.024 80.659 29.801M80.659 27.465Q80.659 27.243 80.825 27.077Q80.991 26.911 81.222 26.911Q81.370 26.911 81.497 26.989Q81.624 27.067 81.698 27.192Q81.773 27.317 81.773 27.465Q81.773 27.692 81.607 27.858Q81.441 28.024 81.222 28.024Q80.995 28.024 80.827 27.856Q80.659 27.688 80.659 27.465\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-36.126 30.897)\">\u003Cpath d=\"M91.913 30.122L91.913 30.032Q91.956 29.825 92.163 29.801L92.585 29.801L92.585 27.473L92.163 27.473Q91.956 27.450 91.913 27.231L91.913 27.145Q91.960 26.934 92.163 26.911L92.980 26.911Q93.175 26.934 93.226 27.145L93.226 27.231L93.218 27.255Q93.445 27.075 93.718 26.973Q93.991 26.872 94.284 26.872Q94.632 26.872 94.870 27.012Q95.109 27.153 95.224 27.411Q95.339 27.669 95.339 28.024L95.339 29.801L95.765 29.801Q95.972 29.825 96.011 30.032L96.011 30.122Q95.972 30.337 95.765 30.360L94.370 30.360Q94.175 30.337 94.124 30.122L94.124 30.032Q94.175 29.821 94.370 29.801L94.698 29.801L94.698 28.055Q94.698 27.747 94.609 27.589Q94.519 27.430 94.226 27.430Q93.956 27.430 93.728 27.561Q93.499 27.692 93.363 27.921Q93.226 28.149 93.226 28.415L93.226 29.801L93.652 29.801Q93.859 29.825 93.898 30.032L93.898 30.122Q93.859 30.337 93.652 30.360L92.163 30.360Q91.956 30.337 91.913 30.122\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-36.126 30.897)\">\u003Cpath d=\"M100.601 30.122L100.601 30.032Q100.640 29.825 100.847 29.801L101.253 29.801L102.183 28.583L101.312 27.473L100.894 27.473Q100.698 27.454 100.648 27.231L100.648 27.145Q100.698 26.934 100.894 26.911L102.054 26.911Q102.253 26.934 102.304 27.145L102.304 27.231Q102.253 27.450 102.054 27.473L101.945 27.473L102.441 28.130L102.917 27.473L102.800 27.473Q102.601 27.450 102.550 27.231L102.550 27.145Q102.601 26.934 102.800 26.911L103.960 26.911Q104.156 26.930 104.206 27.145L104.206 27.231Q104.156 27.450 103.960 27.473L103.550 27.473L102.702 28.583L103.663 29.801L104.070 29.801Q104.269 29.825 104.320 30.032L104.320 30.122Q104.281 30.337 104.070 30.360L102.917 30.360Q102.710 30.337 102.671 30.122L102.671 30.032Q102.710 29.825 102.917 29.801L103.046 29.801L102.441 28.946L101.855 29.801L101.999 29.801Q102.206 29.825 102.245 30.032L102.245 30.122Q102.206 30.337 101.999 30.360L100.847 30.360Q100.652 30.340 100.601 30.122\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-36.126 30.897)\">\u003Cpath d=\"M109.980 29.481Q110.097 29.684 110.331 29.782Q110.566 29.880 110.827 29.880Q111.124 29.880 111.398 29.751Q111.671 29.622 111.845 29.385Q112.019 29.149 112.019 28.848Q112.019 28.594 111.900 28.352Q111.781 28.110 111.566 27.964Q111.351 27.817 111.081 27.817Q110.476 27.817 110.140 28.137Q110.062 28.223 110.007 28.270Q109.952 28.317 109.866 28.329L109.765 28.329Q109.566 28.305 109.515 28.087L109.515 25.704Q109.566 25.497 109.765 25.473L112.124 25.473Q112.320 25.493 112.370 25.704L112.370 25.794Q112.320 26.008 112.124 26.032L110.156 26.032L110.156 27.458Q110.562 27.255 111.081 27.255Q111.511 27.255 111.876 27.471Q112.241 27.688 112.450 28.057Q112.659 28.426 112.659 28.848Q112.659 29.321 112.396 29.680Q112.132 30.040 111.708 30.239Q111.284 30.438 110.827 30.438Q110.573 30.438 110.296 30.364Q110.019 30.290 109.784 30.133Q109.550 29.977 109.409 29.743Q109.269 29.508 109.269 29.223Q109.269 29.055 109.384 28.932Q109.499 28.809 109.675 28.809Q109.757 28.809 109.827 28.839Q109.898 28.868 109.954 28.922Q110.011 28.977 110.042 29.053Q110.073 29.130 110.073 29.208Q110.073 29.368 109.980 29.481M113.945 30.122L113.945 30.032Q113.995 29.825 114.195 29.801L115.011 29.801L115.011 26.618Q114.632 26.926 114.179 26.926Q113.948 26.926 113.898 26.696L113.898 26.606Q113.948 26.391 114.144 26.368Q114.472 26.368 114.726 26.130Q114.980 25.891 115.120 25.544Q115.191 25.415 115.347 25.391L115.402 25.391Q115.597 25.411 115.648 25.626L115.648 29.801L116.464 29.801Q116.663 29.825 116.714 30.032L116.714 30.122Q116.663 30.337 116.464 30.360L114.195 30.360Q113.995 30.337 113.945 30.122M117.761 30.122L117.761 30.047Q117.792 29.880 117.894 29.833L119.183 28.766Q119.515 28.489 119.697 28.337Q119.878 28.184 120.073 27.964Q120.269 27.743 120.390 27.493Q120.511 27.243 120.511 26.977Q120.511 26.653 120.335 26.419Q120.159 26.184 119.880 26.069Q119.601 25.954 119.281 25.954Q119.023 25.954 118.794 26.077Q118.566 26.200 118.464 26.415Q118.566 26.547 118.566 26.696Q118.566 26.856 118.447 26.979Q118.327 27.102 118.167 27.102Q117.991 27.102 117.876 26.977Q117.761 26.852 117.761 26.680Q117.761 26.387 117.896 26.145Q118.031 25.903 118.269 25.731Q118.507 25.559 118.775 25.475Q119.042 25.391 119.335 25.391Q119.816 25.391 120.232 25.581Q120.648 25.770 120.900 26.131Q121.152 26.493 121.152 26.977Q121.152 27.321 121.019 27.624Q120.886 27.926 120.661 28.186Q120.437 28.446 120.128 28.708Q119.820 28.969 119.609 29.145L118.800 29.801L120.511 29.801L120.511 29.657Q120.562 29.446 120.761 29.422L120.902 29.422Q121.101 29.442 121.152 29.657L121.152 30.122Q121.101 30.337 120.902 30.360L118.007 30.360Q117.812 30.340 117.761 30.122\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-soft-accent)\" stroke=\"var(--tk-accent)\">\u003Cpath d=\"M.972 112.873h96.74V90.11H.972Z\"\u002F>\u003Cg fill=\"currentColor\" stroke=\"none\">\u003Cg transform=\"translate(-37.656 73.576)\">\u003Cpath d=\"M50.311 30.176L49.740 26.032L49.686 26.032Q49.490 26.008 49.440 25.794L49.440 25.704Q49.490 25.497 49.686 25.473L50.541 25.473Q50.752 25.497 50.791 25.704L50.791 25.794Q50.752 26.008 50.541 26.032L50.272 26.032Q50.311 26.317 50.447 27.350Q50.584 28.383 50.647 28.928Q50.709 29.473 50.717 29.719Q50.717 29.426 50.803 28.981Q50.889 28.536 51.076 27.688Q51.108 27.567 51.197 27.495Q51.287 27.422 51.404 27.422L51.518 27.422Q51.643 27.422 51.733 27.497Q51.822 27.571 51.846 27.688Q52.010 28.391 52.108 28.899Q52.205 29.407 52.205 29.719Q52.213 29.528 52.262 29.085Q52.311 28.641 52.360 28.251Q52.408 27.860 52.516 27.067Q52.623 26.274 52.654 26.032L52.381 26.032Q52.174 26.008 52.135 25.794L52.135 25.704Q52.174 25.497 52.381 25.473L53.237 25.473Q53.436 25.497 53.487 25.704L53.487 25.794Q53.436 26.008 53.237 26.032L53.182 26.032L52.615 30.176Q52.604 30.286 52.516 30.362Q52.428 30.438 52.311 30.438L52.205 30.438Q52.092 30.438 51.996 30.364Q51.901 30.290 51.877 30.176Q51.779 29.755 51.684 29.305Q51.588 28.856 51.529 28.538Q51.471 28.219 51.463 28.063Q51.463 28.340 51.045 30.176Q51.022 30.290 50.928 30.364Q50.834 30.438 50.717 30.438L50.615 30.438Q50.498 30.438 50.410 30.362Q50.322 30.286 50.311 30.176\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cpath d=\"M16.446 103.536h2.55v.4h-2.55z\"\u002F>\u003Cg transform=\"translate(-37.656 73.576)\">\u003Cpath d=\"M57.371 30.040Q57.277 29.887 57.226 29.667Q57.175 29.446 57.148 29.145Q57.121 28.844 57.115 28.569Q57.109 28.294 57.109 27.911Q57.109 27.399 57.121 27.038Q57.132 26.676 57.191 26.329Q57.250 25.981 57.371 25.794Q57.550 25.532 57.910 25.462Q58.269 25.391 58.773 25.391Q59.277 25.391 59.634 25.462Q59.992 25.532 60.172 25.794Q60.297 25.977 60.355 26.327Q60.414 26.676 60.425 27.038Q60.437 27.399 60.437 27.911Q60.437 28.430 60.425 28.792Q60.414 29.153 60.355 29.505Q60.297 29.856 60.172 30.040Q59.988 30.297 59.632 30.368Q59.277 30.438 58.773 30.438Q58.269 30.438 57.912 30.368Q57.554 30.297 57.371 30.040M57.910 29.551Q58.011 29.758 58.236 29.819Q58.461 29.880 58.773 29.880Q59.086 29.880 59.308 29.819Q59.531 29.758 59.636 29.551Q59.742 29.337 59.769 28.899Q59.797 28.462 59.797 27.817Q59.797 26.544 59.636 26.247Q59.531 26.059 59.312 26.006Q59.093 25.954 58.773 25.954Q58.453 25.954 58.236 26.005Q58.019 26.055 57.910 26.247Q57.750 26.544 57.750 27.817Q57.750 28.126 57.752 28.294Q57.754 28.462 57.771 28.747Q57.789 29.032 57.822 29.233Q57.855 29.434 57.910 29.551\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-37.656 73.576)\">\u003Cpath d=\"M67.273 30.399Q66.800 30.399 66.416 30.155Q66.031 29.911 65.808 29.501Q65.586 29.090 65.586 28.633Q65.586 28.290 65.711 27.967Q65.836 27.645 66.066 27.391Q66.297 27.137 66.603 26.993Q66.910 26.848 67.273 26.848Q67.636 26.848 67.949 26.995Q68.261 27.141 68.484 27.387Q68.707 27.633 68.834 27.954Q68.961 28.274 68.961 28.633Q68.961 29.090 68.736 29.503Q68.511 29.915 68.127 30.157Q67.742 30.399 67.273 30.399M67.273 29.840Q67.738 29.840 68.029 29.446Q68.320 29.051 68.320 28.567Q68.320 28.274 68.185 28.006Q68.050 27.739 67.810 27.573Q67.570 27.407 67.273 27.407Q66.968 27.407 66.730 27.573Q66.492 27.739 66.357 28.006Q66.222 28.274 66.222 28.567Q66.222 29.047 66.515 29.444Q66.808 29.840 67.273 29.840M70.140 29.505L70.140 27.473L69.718 27.473Q69.511 27.450 69.468 27.231L69.468 27.145Q69.515 26.934 69.718 26.911L70.535 26.911Q70.730 26.934 70.781 27.145L70.781 29.473Q70.781 29.708 70.951 29.774Q71.121 29.840 71.406 29.840Q71.613 29.840 71.808 29.764Q72.004 29.688 72.129 29.538Q72.254 29.387 72.254 29.176L72.254 27.473L71.832 27.473Q71.621 27.450 71.582 27.231L71.582 27.145Q71.621 26.934 71.832 26.911L72.644 26.911Q72.843 26.934 72.894 27.145L72.894 29.801L73.320 29.801Q73.527 29.825 73.566 30.032L73.566 30.122Q73.527 30.337 73.320 30.360L72.504 30.360Q72.304 30.337 72.254 30.137Q71.851 30.399 71.343 30.399Q71.109 30.399 70.894 30.358Q70.679 30.317 70.513 30.215Q70.347 30.114 70.244 29.936Q70.140 29.758 70.140 29.505M74.843 29.255L74.843 27.473L74.093 27.473Q73.894 27.450 73.843 27.231L73.843 27.145Q73.894 26.934 74.093 26.911L74.843 26.911L74.843 26.161Q74.894 25.954 75.093 25.926L75.238 25.926Q75.433 25.954 75.484 26.161L75.484 26.911L76.843 26.911Q77.035 26.930 77.093 27.145L77.093 27.231Q77.039 27.450 76.843 27.473L75.484 27.473L75.484 29.223Q75.484 29.840 76.058 29.840Q76.308 29.840 76.472 29.655Q76.636 29.469 76.636 29.223Q76.636 29.130 76.709 29.059Q76.781 28.989 76.882 28.977L77.027 28.977Q77.226 29.001 77.277 29.208L77.277 29.255Q77.277 29.579 77.093 29.842Q76.910 30.106 76.617 30.253Q76.324 30.399 76.004 30.399Q75.492 30.399 75.168 30.089Q74.843 29.778 74.843 29.255\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-37.656 73.576)\">\u003Cpath d=\"M83.719 29.801Q83.719 29.579 83.885 29.413Q84.051 29.247 84.282 29.247Q84.430 29.247 84.557 29.325Q84.684 29.403 84.758 29.528Q84.833 29.653 84.833 29.801Q84.833 30.028 84.667 30.194Q84.501 30.360 84.282 30.360Q84.055 30.360 83.887 30.192Q83.719 30.024 83.719 29.801M83.719 27.465Q83.719 27.243 83.885 27.077Q84.051 26.911 84.282 26.911Q84.430 26.911 84.557 26.989Q84.684 27.067 84.758 27.192Q84.833 27.317 84.833 27.465Q84.833 27.692 84.667 27.858Q84.501 28.024 84.282 28.024Q84.055 28.024 83.887 27.856Q83.719 27.688 83.719 27.465\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-37.656 73.576)\">\u003Cpath d=\"M94.973 30.122L94.973 30.032Q95.016 29.825 95.223 29.801L95.645 29.801L95.645 27.473L95.223 27.473Q95.016 27.450 94.973 27.231L94.973 27.145Q95.020 26.934 95.223 26.911L96.040 26.911Q96.235 26.934 96.286 27.145L96.286 27.231L96.278 27.255Q96.505 27.075 96.778 26.973Q97.051 26.872 97.344 26.872Q97.692 26.872 97.930 27.012Q98.169 27.153 98.284 27.411Q98.399 27.669 98.399 28.024L98.399 29.801L98.825 29.801Q99.032 29.825 99.071 30.032L99.071 30.122Q99.032 30.337 98.825 30.360L97.430 30.360Q97.235 30.337 97.184 30.122L97.184 30.032Q97.235 29.821 97.430 29.801L97.758 29.801L97.758 28.055Q97.758 27.747 97.669 27.589Q97.579 27.430 97.286 27.430Q97.016 27.430 96.788 27.561Q96.559 27.692 96.423 27.921Q96.286 28.149 96.286 28.415L96.286 29.801L96.712 29.801Q96.919 29.825 96.958 30.032L96.958 30.122Q96.919 30.337 96.712 30.360L95.223 30.360Q95.016 30.337 94.973 30.122\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-37.656 73.576)\">\u003Cpath d=\"M103.661 30.122L103.661 30.032Q103.700 29.825 103.907 29.801L104.313 29.801L105.243 28.583L104.372 27.473L103.954 27.473Q103.758 27.454 103.708 27.231L103.708 27.145Q103.758 26.934 103.954 26.911L105.114 26.911Q105.313 26.934 105.364 27.145L105.364 27.231Q105.313 27.450 105.114 27.473L105.005 27.473L105.501 28.130L105.977 27.473L105.860 27.473Q105.661 27.450 105.610 27.231L105.610 27.145Q105.661 26.934 105.860 26.911L107.020 26.911Q107.216 26.930 107.266 27.145L107.266 27.231Q107.216 27.450 107.020 27.473L106.610 27.473L105.762 28.583L106.723 29.801L107.130 29.801Q107.329 29.825 107.380 30.032L107.380 30.122Q107.341 30.337 107.130 30.360L105.977 30.360Q105.770 30.337 105.731 30.122L105.731 30.032Q105.770 29.825 105.977 29.801L106.106 29.801L105.501 28.946L104.915 29.801L105.059 29.801Q105.266 29.825 105.305 30.032L105.305 30.122Q105.266 30.337 105.059 30.360L103.907 30.360Q103.712 30.340 103.661 30.122\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-37.656 73.576)\">\u003Cpath d=\"M113.040 29.481Q113.157 29.684 113.391 29.782Q113.626 29.880 113.887 29.880Q114.184 29.880 114.458 29.751Q114.731 29.622 114.905 29.385Q115.079 29.149 115.079 28.848Q115.079 28.594 114.960 28.352Q114.841 28.110 114.626 27.964Q114.411 27.817 114.141 27.817Q113.536 27.817 113.200 28.137Q113.122 28.223 113.067 28.270Q113.012 28.317 112.926 28.329L112.825 28.329Q112.626 28.305 112.575 28.087L112.575 25.704Q112.626 25.497 112.825 25.473L115.184 25.473Q115.380 25.493 115.430 25.704L115.430 25.794Q115.380 26.008 115.184 26.032L113.216 26.032L113.216 27.458Q113.622 27.255 114.141 27.255Q114.571 27.255 114.936 27.471Q115.301 27.688 115.510 28.057Q115.719 28.426 115.719 28.848Q115.719 29.321 115.456 29.680Q115.192 30.040 114.768 30.239Q114.344 30.438 113.887 30.438Q113.633 30.438 113.356 30.364Q113.079 30.290 112.844 30.133Q112.610 29.977 112.469 29.743Q112.329 29.508 112.329 29.223Q112.329 29.055 112.444 28.932Q112.559 28.809 112.735 28.809Q112.817 28.809 112.887 28.839Q112.958 28.868 113.014 28.922Q113.071 28.977 113.102 29.053Q113.133 29.130 113.133 29.208Q113.133 29.368 113.040 29.481M117.005 30.122L117.005 30.032Q117.055 29.825 117.255 29.801L118.071 29.801L118.071 26.618Q117.692 26.926 117.239 26.926Q117.008 26.926 116.958 26.696L116.958 26.606Q117.008 26.391 117.204 26.368Q117.532 26.368 117.786 26.130Q118.040 25.891 118.180 25.544Q118.251 25.415 118.407 25.391L118.462 25.391Q118.657 25.411 118.708 25.626L118.708 29.801L119.524 29.801Q119.723 29.825 119.774 30.032L119.774 30.122Q119.723 30.337 119.524 30.360L117.255 30.360Q117.055 30.337 117.005 30.122M120.821 30.122L120.821 30.047Q120.852 29.880 120.954 29.833L122.243 28.766Q122.575 28.489 122.757 28.337Q122.938 28.184 123.133 27.964Q123.329 27.743 123.450 27.493Q123.571 27.243 123.571 26.977Q123.571 26.653 123.395 26.419Q123.219 26.184 122.940 26.069Q122.661 25.954 122.341 25.954Q122.083 25.954 121.854 26.077Q121.626 26.200 121.524 26.415Q121.626 26.547 121.626 26.696Q121.626 26.856 121.507 26.979Q121.387 27.102 121.227 27.102Q121.051 27.102 120.936 26.977Q120.821 26.852 120.821 26.680Q120.821 26.387 120.956 26.145Q121.091 25.903 121.329 25.731Q121.567 25.559 121.835 25.475Q122.102 25.391 122.395 25.391Q122.876 25.391 123.292 25.581Q123.708 25.770 123.960 26.131Q124.212 26.493 124.212 26.977Q124.212 27.321 124.079 27.624Q123.946 27.926 123.721 28.186Q123.497 28.446 123.188 28.708Q122.880 28.969 122.669 29.145L121.860 29.801L123.571 29.801L123.571 29.657Q123.622 29.446 123.821 29.422L123.962 29.422Q124.161 29.442 124.212 29.657L124.212 30.122Q124.161 30.337 123.962 30.360L121.067 30.360Q120.872 30.340 120.821 30.122\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M44.771-49.108 34.648-23.473\"\u002F>\u003Cpath stroke=\"none\" d=\"m33.693-21.055 3.463-3.105-2.508.687-1.361-2.215\"\u002F>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"m53.498-49.108 9.192 25.606\"\u002F>\u003Cpath stroke=\"none\" d=\"m63.569-21.055.552-4.618-1.43 2.171-2.485-.765\"\u002F>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"m-42.93-4.733 92.272 33.67M-13.054-2.85l62.396 31.787M16.822 1.493l32.52 27.444M38.411 2.107l10.931 26.83M59.28 2.107l-9.938 26.83M80.44 1.852 49.343 28.937M110.316-2.729 49.342 28.937M140.192-4.665l-90.85 33.602M49.342 28.937v16.294\"\u002F>\u003Cpath stroke=\"none\" d=\"m49.342 47.231 1.6-3.2-1.6 1.2-1.6-1.2\"\u002F>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M49.342 70.394V87.31\"\u002F>\u003Cpath stroke=\"none\" d=\"m49.342 89.91 2.08-4.16-2.08 1.56-2.08-1.56\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fsvg>\u003Cfigcaption class=\"tikz-cap\">Multi-head width bookkeeping for \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6944em;\">\u003C\u002Fspan>\u003Cspan class=\"mord mathnormal\">d\u003C\u002Fspan>\u003Cspan class=\"mspace\" style=\"margin-right:0.2778em;\">\u003C\u002Fspan>\u003Cspan class=\"mrel\">=\u003C\u002Fspan>\u003Cspan class=\"mspace\" style=\"margin-right:0.2778em;\">\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6444em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">512\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>, \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6944em;\">\u003C\u002Fspan>\u003Cspan class=\"mord mathnormal\">h\u003C\u002Fspan>\u003Cspan class=\"mspace\" style=\"margin-right:0.2778em;\">\u003C\u002Fspan>\u003Cspan class=\"mrel\">=\u003C\u002Fspan>\u003Cspan class=\"mspace\" style=\"margin-right:0.2778em;\">\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6444em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">8\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>. Each head projects the \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6667em;vertical-align:-0.0833em;\">\u003C\u002Fspan>\u003Cspan class=\"mord mathnormal\">n\u003C\u002Fspan>\u003Cspan class=\"mspace\" style=\"margin-right:0.2222em;\">\u003C\u002Fspan>\u003Cspan class=\"mbin\">×\u003C\u002Fspan>\u003Cspan class=\"mspace\" style=\"margin-right:0.2222em;\">\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6444em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">512\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan> input down to \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6667em;vertical-align:-0.0833em;\">\u003C\u002Fspan>\u003Cspan class=\"mord mathnormal\">n\u003C\u002Fspan>\u003Cspan class=\"mspace\" style=\"margin-right:0.2222em;\">\u003C\u002Fspan>\u003Cspan class=\"mbin\">×\u003C\u002Fspan>\u003Cspan class=\"mspace\" style=\"margin-right:0.2222em;\">\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6444em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">64\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>, attends there, and the \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6444em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">8\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan> blocks concatenate back to \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6667em;vertical-align:-0.0833em;\">\u003C\u002Fspan>\u003Cspan class=\"mord mathnormal\">n\u003C\u002Fspan>\u003Cspan class=\"mspace\" style=\"margin-right:0.2222em;\">\u003C\u002Fspan>\u003Cspan class=\"mbin\">×\u003C\u002Fspan>\u003Cspan class=\"mspace\" style=\"margin-right:0.2222em;\">\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6444em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">512\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan> before \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.8333em;vertical-align:-0.15em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">\u003Cspan class=\"mord mathnormal\" style=\"margin-right:0.1389em;\">W\u003C\u002Fspan>\u003Cspan class=\"msupsub\">\u003Cspan class=\"vlist-t vlist-t2\">\u003Cspan class=\"vlist-r\">\u003Cspan class=\"vlist\" style=\"height:0.3283em;\">\u003Cspan style=\"top:-2.55em;margin-left:-0.1389em;margin-right:0.05em;\">\u003Cspan class=\"pstrut\" style=\"height:2.7em;\">\u003C\u002Fspan>\u003Cspan class=\"sizing reset-size6 size3 mtight\">\u003Cspan class=\"mord mathnormal mtight\" style=\"margin-right:0.0278em;\">O\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"vlist-s\">​\u003C\u002Fspan>\u003C\u002Fspan>\u003Cspan class=\"vlist-r\">\u003Cspan class=\"vlist\" style=\"height:0.15em;\">\u003Cspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>.\u003C\u002Ffigcaption>","\u003Csvg style=\"width:100%;max-width:377.675px;height:auto\" xmlns=\"http:\u002F\u002Fwww.w3.org\u002F2000\u002Fsvg\" viewBox=\"-75 -75 283.256 100.840\">\u003Cg stroke=\"currentColor\" style=\"stroke-miterlimit:10;stroke-width:.4\">\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M-58.964-17.664h219.331\"\u002F>\u003Cpath stroke=\"none\" d=\"m162.967-17.664-4.16-2.08 1.56 2.08-1.56 2.08\"\u002F>\u003Cg stroke=\"none\" font-family=\"cmr7\" font-size=\"7\">\u003Cg transform=\"translate(219.974 1.674)\">\u003Cpath d=\"M-51.316-16.307L-52.946-16.307L-52.946-16.587Q-52.717-16.587-52.568-16.622Q-52.420-16.656-52.420-16.796L-52.420-20.142Q-52.420-20.313-52.556-20.354Q-52.693-20.395-52.946-20.395L-52.946-20.675L-51.866-20.750L-51.866-20.344Q-51.644-20.545-51.357-20.648Q-51.069-20.750-50.762-20.750Q-50.335-20.750-49.971-20.537Q-49.607-20.323-49.393-19.959Q-49.179-19.595-49.179-19.175Q-49.179-18.730-49.419-18.366Q-49.658-18.002-50.051-17.799Q-50.444-17.596-50.888-17.596Q-51.155-17.596-51.403-17.696Q-51.650-17.797-51.838-17.978L-51.838-16.796Q-51.838-16.659-51.690-16.623Q-51.541-16.587-51.316-16.587L-51.316-16.307M-51.838-19.995L-51.838-18.385Q-51.705-18.132-51.462-17.975Q-51.220-17.818-50.943-17.818Q-50.615-17.818-50.362-18.019Q-50.109-18.221-49.976-18.539Q-49.842-18.857-49.842-19.175Q-49.842-19.404-49.907-19.633Q-49.972-19.862-50.100-20.060Q-50.229-20.258-50.423-20.378Q-50.618-20.497-50.851-20.497Q-51.145-20.497-51.413-20.368Q-51.681-20.238-51.838-19.995\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(219.974 1.674)\">\u003Cpath d=\"M-48.369-19.147Q-48.369-19.489-48.234-19.788Q-48.099-20.087-47.859-20.311Q-47.620-20.535-47.302-20.660Q-46.984-20.785-46.653-20.785Q-46.208-20.785-45.809-20.569Q-45.409-20.354-45.174-19.976Q-44.940-19.599-44.940-19.147Q-44.940-18.806-45.082-18.522Q-45.224-18.238-45.468-18.031Q-45.713-17.825-46.022-17.710Q-46.331-17.596-46.653-17.596Q-47.083-17.596-47.485-17.797Q-47.887-17.999-48.128-18.351Q-48.369-18.703-48.369-19.147M-46.653-17.845Q-46.051-17.845-45.827-18.223Q-45.603-18.601-45.603-19.233Q-45.603-19.845-45.838-20.204Q-46.072-20.562-46.653-20.562Q-47.705-20.562-47.705-19.233Q-47.705-18.601-47.480-18.223Q-47.254-17.845-46.653-17.845M-44.346-17.671L-44.346-18.734Q-44.346-18.758-44.318-18.785Q-44.291-18.812-44.267-18.812L-44.158-18.812Q-44.093-18.812-44.079-18.754Q-43.983-18.320-43.737-18.069Q-43.491-17.818-43.078-17.818Q-42.736-17.818-42.483-17.951Q-42.230-18.084-42.230-18.392Q-42.230-18.549-42.324-18.664Q-42.418-18.778-42.556-18.847Q-42.695-18.915-42.862-18.953L-43.443-19.052Q-43.799-19.120-44.072-19.341Q-44.346-19.561-44.346-19.903Q-44.346-20.152-44.235-20.327Q-44.123-20.501-43.937-20.600Q-43.751-20.699-43.536-20.742Q-43.320-20.785-43.078-20.785Q-42.664-20.785-42.384-20.603L-42.168-20.778Q-42.158-20.781-42.151-20.783Q-42.144-20.785-42.134-20.785L-42.083-20.785Q-42.056-20.785-42.032-20.761Q-42.008-20.737-42.008-20.709L-42.008-19.862Q-42.008-19.841-42.032-19.814Q-42.056-19.787-42.083-19.787L-42.196-19.787Q-42.223-19.787-42.249-19.812Q-42.274-19.838-42.274-19.862Q-42.274-20.098-42.380-20.262Q-42.486-20.426-42.669-20.508Q-42.852-20.590-43.084-20.590Q-43.413-20.590-43.669-20.487Q-43.925-20.385-43.925-20.108Q-43.925-19.913-43.742-19.804Q-43.559-19.694-43.330-19.653L-42.756-19.547Q-42.510-19.499-42.297-19.371Q-42.083-19.243-41.946-19.040Q-41.809-18.836-41.809-18.587Q-41.809-18.074-42.175-17.835Q-42.541-17.596-43.078-17.596Q-43.573-17.596-43.905-17.890L-44.171-17.616Q-44.192-17.596-44.219-17.596L-44.267-17.596Q-44.291-17.596-44.318-17.623Q-44.346-17.650-44.346-17.671M-39.564-17.664L-41.116-17.664L-41.116-17.944Q-40.890-17.944-40.741-17.978Q-40.593-18.013-40.593-18.153L-40.593-20.002Q-40.593-20.190-40.641-20.274Q-40.688-20.357-40.786-20.376Q-40.883-20.395-41.095-20.395L-41.095-20.675L-40.039-20.750L-40.039-18.153Q-40.039-18.013-39.907-17.978Q-39.776-17.944-39.564-17.944L-39.564-17.664M-40.835-21.971Q-40.835-22.142-40.712-22.261Q-40.589-22.381-40.418-22.381Q-40.251-22.381-40.128-22.261Q-40.005-22.142-40.005-21.971Q-40.005-21.796-40.128-21.673Q-40.251-21.550-40.418-21.550Q-40.589-21.550-40.712-21.673Q-40.835-21.796-40.835-21.971M-38.392-18.505L-38.392-20.402L-39.031-20.402L-39.031-20.624Q-38.713-20.624-38.496-20.834Q-38.279-21.044-38.178-21.354Q-38.077-21.663-38.077-21.971L-37.810-21.971L-37.810-20.682L-36.734-20.682L-36.734-20.402L-37.810-20.402L-37.810-18.518Q-37.810-18.242-37.706-18.043Q-37.602-17.845-37.342-17.845Q-37.185-17.845-37.079-17.949Q-36.973-18.054-36.923-18.207Q-36.874-18.361-36.874-18.518L-36.874-18.932L-36.607-18.932L-36.607-18.505Q-36.607-18.279-36.706-18.069Q-36.806-17.859-36.990-17.727Q-37.175-17.596-37.404-17.596Q-37.841-17.596-38.116-17.833Q-38.392-18.071-38.392-18.505M-34.181-17.664L-35.732-17.664L-35.732-17.944Q-35.507-17.944-35.358-17.978Q-35.209-18.013-35.209-18.153L-35.209-20.002Q-35.209-20.190-35.257-20.274Q-35.305-20.357-35.402-20.376Q-35.500-20.395-35.712-20.395L-35.712-20.675L-34.656-20.750L-34.656-18.153Q-34.656-18.013-34.524-17.978Q-34.392-17.944-34.181-17.944L-34.181-17.664M-35.452-21.971Q-35.452-22.142-35.329-22.261Q-35.206-22.381-35.035-22.381Q-34.868-22.381-34.745-22.261Q-34.621-22.142-34.621-21.971Q-34.621-21.796-34.745-21.673Q-34.868-21.550-35.035-21.550Q-35.206-21.550-35.329-21.673Q-35.452-21.796-35.452-21.971M-33.576-19.147Q-33.576-19.489-33.441-19.788Q-33.306-20.087-33.066-20.311Q-32.827-20.535-32.509-20.660Q-32.191-20.785-31.860-20.785Q-31.415-20.785-31.016-20.569Q-30.616-20.354-30.382-19.976Q-30.147-19.599-30.147-19.147Q-30.147-18.806-30.289-18.522Q-30.431-18.238-30.675-18.031Q-30.920-17.825-31.229-17.710Q-31.538-17.596-31.860-17.596Q-32.290-17.596-32.692-17.797Q-33.094-17.999-33.335-18.351Q-33.576-18.703-33.576-19.147M-31.860-17.845Q-31.258-17.845-31.034-18.223Q-30.810-18.601-30.810-19.233Q-30.810-19.845-31.045-20.204Q-31.279-20.562-31.860-20.562Q-32.913-20.562-32.913-19.233Q-32.913-18.601-32.687-18.223Q-32.461-17.845-31.860-17.845M-27.871-17.664L-29.505-17.664L-29.505-17.944Q-29.276-17.944-29.127-17.978Q-28.978-18.013-28.978-18.153L-28.978-20.002Q-28.978-20.272-29.086-20.333Q-29.194-20.395-29.505-20.395L-29.505-20.675L-28.445-20.750L-28.445-20.101Q-28.274-20.409-27.970-20.580Q-27.666-20.750-27.321-20.750Q-26.815-20.750-26.531-20.527Q-26.247-20.303-26.247-19.807L-26.247-18.153Q-26.247-18.016-26.099-17.980Q-25.950-17.944-25.725-17.944L-25.725-17.664L-27.355-17.664L-27.355-17.944Q-27.126-17.944-26.977-17.978Q-26.829-18.013-26.829-18.153L-26.829-19.793Q-26.829-20.128-26.948-20.328Q-27.068-20.528-27.382-20.528Q-27.652-20.528-27.886-20.392Q-28.121-20.255-28.259-20.021Q-28.397-19.787-28.397-19.513L-28.397-18.153Q-28.397-18.016-28.247-17.980Q-28.097-17.944-27.871-17.944\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(219.974 1.674)\">\u003Cpath d=\"M-20.780-16.307L-22.410-16.307L-22.410-16.587Q-22.181-16.587-22.032-16.622Q-21.884-16.656-21.884-16.796L-21.884-20.142Q-21.884-20.313-22.020-20.354Q-22.157-20.395-22.410-20.395L-22.410-20.675L-21.330-20.750L-21.330-20.344Q-21.108-20.545-20.821-20.648Q-20.533-20.750-20.226-20.750Q-19.799-20.750-19.435-20.537Q-19.071-20.323-18.857-19.959Q-18.643-19.595-18.643-19.175Q-18.643-18.730-18.883-18.366Q-19.122-18.002-19.515-17.799Q-19.908-17.596-20.352-17.596Q-20.619-17.596-20.867-17.696Q-21.114-17.797-21.302-17.978L-21.302-16.796Q-21.302-16.659-21.154-16.623Q-21.005-16.587-20.780-16.587L-20.780-16.307M-21.302-19.995L-21.302-18.385Q-21.169-18.132-20.926-17.975Q-20.684-17.818-20.407-17.818Q-20.079-17.818-19.826-18.019Q-19.573-18.221-19.440-18.539Q-19.306-18.857-19.306-19.175Q-19.306-19.404-19.371-19.633Q-19.436-19.862-19.564-20.060Q-19.693-20.258-19.887-20.378Q-20.082-20.497-20.315-20.497Q-20.609-20.497-20.877-20.368Q-21.145-20.238-21.302-19.995\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M-53.274 22.17v-79.913\"\u002F>\u003Cpath stroke=\"none\" d=\"m-53.274-60.343-2.08 4.16 2.08-1.56 2.08 1.56\"\u002F>\u003Cg stroke=\"none\" font-family=\"cmr7\" font-size=\"7\">\u003Cg transform=\"translate(-8.997 -46.412)\">\u003Cpath d=\"M-51.370-17.691L-52.498-20.190Q-52.570-20.337-52.700-20.369Q-52.830-20.402-53.059-20.402L-53.059-20.682L-51.545-20.682L-51.545-20.402Q-51.897-20.402-51.897-20.255Q-51.897-20.210-51.886-20.190L-51.022-18.272L-50.242-20.002Q-50.208-20.070-50.208-20.149Q-50.208-20.262-50.292-20.332Q-50.376-20.402-50.495-20.402L-50.495-20.682L-49.299-20.682L-49.299-20.402Q-49.518-20.402-49.689-20.299Q-49.859-20.197-49.948-20.002L-50.984-17.691Q-51.032-17.596-51.138-17.596L-51.216-17.596Q-51.322-17.596-51.370-17.691\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-8.997 -46.412)\">\u003Cpath d=\"M-49.130-18.392Q-49.130-18.724-48.907-18.951Q-48.683-19.178-48.339-19.306Q-47.996-19.435-47.623-19.487Q-47.251-19.540-46.946-19.540L-46.946-19.793Q-46.946-19.998-47.054-20.178Q-47.162-20.357-47.343-20.460Q-47.524-20.562-47.732-20.562Q-48.139-20.562-48.375-20.470Q-48.286-20.433-48.240-20.349Q-48.194-20.265-48.194-20.163Q-48.194-20.067-48.240-19.988Q-48.286-19.910-48.367-19.865Q-48.447-19.821-48.536-19.821Q-48.686-19.821-48.787-19.918Q-48.888-20.016-48.888-20.163Q-48.888-20.785-47.732-20.785Q-47.521-20.785-47.271-20.721Q-47.022-20.658-46.820-20.539Q-46.618-20.419-46.492-20.234Q-46.365-20.050-46.365-19.807L-46.365-18.231Q-46.365-18.115-46.304-18.019Q-46.242-17.924-46.129-17.924Q-46.020-17.924-45.955-18.018Q-45.890-18.112-45.890-18.231L-45.890-18.679L-45.624-18.679L-45.624-18.231Q-45.624-17.961-45.851-17.796Q-46.078-17.630-46.358-17.630Q-46.567-17.630-46.704-17.784Q-46.840-17.937-46.864-18.153Q-47.011-17.886-47.293-17.741Q-47.575-17.596-47.900-17.596Q-48.177-17.596-48.461-17.671Q-48.744-17.746-48.937-17.925Q-49.130-18.105-49.130-18.392M-48.515-18.392Q-48.515-18.218-48.414-18.088Q-48.314-17.958-48.158-17.888Q-48.003-17.818-47.838-17.818Q-47.620-17.818-47.411-17.915Q-47.203-18.013-47.075-18.194Q-46.946-18.375-46.946-18.601L-46.946-19.329Q-47.271-19.329-47.637-19.238Q-48.003-19.147-48.259-18.935Q-48.515-18.724-48.515-18.392M-43.539-17.664L-45.142-17.664L-45.142-17.944Q-44.916-17.944-44.767-17.978Q-44.619-18.013-44.619-18.153L-44.619-21.772Q-44.619-22.042-44.726-22.104Q-44.834-22.165-45.142-22.165L-45.142-22.446L-44.065-22.521L-44.065-18.153Q-44.065-18.016-43.915-17.980Q-43.764-17.944-43.539-17.944L-43.539-17.664M-42.370-18.498L-42.370-20.002Q-42.370-20.272-42.477-20.333Q-42.585-20.395-42.896-20.395L-42.896-20.675L-41.789-20.750L-41.789-18.518L-41.789-18.498Q-41.789-18.218-41.737-18.074Q-41.686-17.931-41.544-17.874Q-41.402-17.818-41.115-17.818Q-40.862-17.818-40.657-17.958Q-40.452-18.098-40.336-18.324Q-40.220-18.549-40.220-18.799L-40.220-20.002Q-40.220-20.272-40.327-20.333Q-40.435-20.395-40.746-20.395L-40.746-20.675L-39.639-20.750L-39.639-18.337Q-39.639-18.146-39.586-18.064Q-39.533-17.982-39.432-17.963Q-39.331-17.944-39.116-17.944L-39.116-17.664L-40.192-17.596L-40.192-18.160Q-40.302-17.978-40.447-17.855Q-40.592-17.732-40.779-17.664Q-40.965-17.596-41.167-17.596Q-42.370-17.596-42.370-18.498M-38.569-19.199Q-38.569-19.520-38.444-19.809Q-38.319-20.098-38.094-20.321Q-37.868-20.545-37.573-20.665Q-37.277-20.785-36.959-20.785Q-36.631-20.785-36.369-20.685Q-36.108-20.586-35.932-20.404Q-35.756-20.221-35.662-19.963Q-35.568-19.705-35.568-19.373Q-35.568-19.281-35.650-19.260L-37.906-19.260L-37.906-19.199Q-37.906-18.611-37.622-18.228Q-37.338-17.845-36.771-17.845Q-36.450-17.845-36.181-18.038Q-35.913-18.231-35.824-18.546Q-35.817-18.587-35.742-18.601L-35.650-18.601Q-35.568-18.577-35.568-18.505Q-35.568-18.498-35.575-18.471Q-35.688-18.074-36.058-17.835Q-36.429-17.596-36.853-17.596Q-37.291-17.596-37.691-17.804Q-38.090-18.013-38.330-18.380Q-38.569-18.747-38.569-19.199M-37.899-19.469L-36.084-19.469Q-36.084-19.746-36.181-19.998Q-36.279-20.251-36.477-20.407Q-36.675-20.562-36.959-20.562Q-37.236-20.562-37.450-20.404Q-37.663-20.245-37.781-19.990Q-37.899-19.735-37.899-19.469\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"M-53.274-17.664h210.55\"\u002F>\u003Cpath fill=\"none\" stroke=\"var(--tk-accent)\" d=\"m-53.274-17.664 1.745-4.97 1.746-4.844 1.745-4.595 1.745-4.229 1.746-3.756 1.745-3.188 1.745-2.539 1.746-1.824 1.745-1.065 1.745-.277 1.745.516 1.746 1.298 1.745 2.045 1.745 2.742 1.746 3.368 1.745 3.91 1.745 4.351 1.746 4.684 1.745 4.895 1.745 4.982 1.745 4.946 1.746 4.782 1.745 4.496 1.745 4.097 1.746 3.595 1.745 3 1.745 2.329 1.746 1.6 1.745.829 1.745.037 1.746-.754 1.745-1.528L4.32 9.008l1.745-2.94 1.746-3.54 1.745-4.054 1.745-4.464 1.746-4.76 1.745-4.935 1.745-4.985 1.746-4.909 1.745-4.708 1.745-4.389 1.745-3.955 1.746-3.424 1.745-2.804 1.745-2.115 1.746-1.37 1.745-.592 1.745.201 1.746.991 1.745 1.755 1.745 2.473 1.745 3.129 1.746 3.706 1.745 4.189 1.745 4.564 1.746 4.827 1.745 4.963 1.745 4.975 1.746 4.862 1.745 4.623L56.68.591l1.745 3.805 1.746 3.247 1.745 2.603 1.745 1.894 1.746 1.138 1.745.353 1.745-.441 1.746-1.224 1.745-1.977 1.745-2.677 1.745-3.314L77.624.136l1.745-4.314 1.745-4.656 1.746-4.88 1.745-4.98 1.745-4.955 1.746-4.803 1.745-4.529 1.745-4.14 1.745-3.645 1.746-3.06 1.745-2.397 1.745-1.67 1.746-.904 1.745-.113 1.745.68 1.746 1.457 1.745 2.194 1.745 2.877 1.745 3.487 1.746 4.01 1.745 4.43 1.745 4.737 1.746 4.923 1.745 4.985 1.745 4.922 1.746 4.733 1.745 4.423 1.745 4.001 1.745 3.478 1.746 2.867 1.745 2.183 1.745 1.443 1.746.667 1.745-.126 1.745-.918 1.746-1.683 1.745-2.407 1.745-3.07 1.745-3.655 1.746-4.147 1.745-4.535 1.745-4.806 1.746-4.956 1.745-4.979\" style=\"stroke-width:1.2\"\u002F>\u003Cpath fill=\"none\" d=\"m-53.274-17.664 1.745-2.11 1.746-2.1 1.745-2.082 1.745-2.052 1.746-2.015 1.745-1.967 1.745-1.912 1.746-1.847 1.745-1.773 1.745-1.693 1.745-1.603 1.746-1.507 1.745-1.404 1.745-1.295 1.746-1.179 1.745-1.057 1.745-.933 1.746-.802 1.745-.67 1.745-.532 1.745-.394 1.746-.253 1.745-.111 1.745.032 1.746.173 1.745.315 1.745.455 1.746.593 1.745.729 1.745.86 1.746.988 1.745 1.113 1.745 1.23 1.745 1.343 1.746 1.45 1.745 1.55 1.745 1.643 1.746 1.73 1.745 1.807 1.745 1.876 1.746 1.938 1.745 1.99 1.745 2.033 1.745 2.065 1.746 2.091 1.745 2.106 1.745 2.11 1.746 2.107 1.745 2.094 1.745 2.07 1.746 2.037 1.745 1.996 1.745 1.944 1.745 1.883 1.746 1.815 1.745 1.74 1.745 1.654 1.746 1.561 1.745 1.463 1.745 1.356L53.19 8.23l1.745 1.125 1.745 1.004 1.745.875 1.746.745 1.745.61 1.745.472 1.746.332 1.745.19 1.745.048 1.746-.095 1.745-.236 1.745-.376 1.745-.516 1.746-.653 1.745-.787 1.745-.918 1.746-1.043 1.745-1.165 1.745-1.28 1.746-1.392 1.745-1.495 1.745-1.592L93.331.4l1.746-1.764 1.745-1.838 1.745-1.904 1.746-1.962 1.745-2.01 1.745-2.048 1.746-2.079 1.745-2.098 1.745-2.11 1.745-2.11 1.746-2.102 1.745-2.084 1.745-2.056 1.746-2.02 1.745-1.974 1.745-1.918 1.746-1.856 1.745-1.782 1.745-1.702 1.745-1.615 1.746-1.518 1.745-1.417 1.745-1.308 1.746-1.192 1.745-1.073 1.745-.948 1.746-.819 1.745-.686 1.745-.548 1.745-.41 1.746-.27 1.745-.128 1.745.015 1.746.156 1.745.297\" style=\"stroke-width:1.2\"\u002F>\u003Cpath fill=\"none\" stroke=\"var(--tk-warn)\" d=\"m-53.274-17.664 1.745-.862 1.746-.864 1.745-.861 1.745-.86 1.746-.858 1.745-.853 1.745-.85 1.746-.846 1.745-.84 1.745-.835 1.745-.827 1.746-.822 1.745-.812 1.745-.805 1.746-.795 1.745-.787 1.745-.775 1.746-.765 1.745-.754 1.745-.741 1.745-.73 1.746-.716 1.745-.702 1.745-.689 1.746-.673 1.745-.66 1.745-.642 1.746-.627 1.745-.61 1.745-.592 1.746-.577 1.745-.556 1.745-.54 1.745-.52 1.746-.5 1.745-.482 1.745-.46 1.746-.443 1.745-.42 1.745-.398 1.746-.379 1.745-.356 1.745-.336 1.745-.31 1.746-.291 1.745-.267 1.745-.246 1.746-.222 1.745-.198 1.745-.176 1.746-.151 1.745-.129 1.745-.105 1.745-.08 1.746-.059 1.745-.032 1.745-.01 1.746.014 1.745.038 1.745.061 1.746.086 1.745.109 1.745.133 1.745.157 1.746.18 1.745.202 1.745.225 1.746.25 1.745.271 1.745.296 1.746.316 1.745.338 1.745.361 1.745.381 1.746.404 1.745.425 1.745.444 1.746.466 1.745.484 1.745.505 1.746.523 1.745.542 1.745.56 1.745.58 1.746.596 1.745.613 1.745.63 1.746.645 1.745.662 1.745.676 1.746.692 1.745.705 1.745.718 1.745.732 1.746.744 1.745.756 1.745.767 1.746.777 1.745.789 1.745.797 1.746.806 1.745.815 1.745.822 1.745.83 1.746.834 1.745.841 1.745.847 1.746.85 1.745.855 1.745.858 1.746.86 1.745.863 1.745.862 1.745.863 1.746.864 1.745.863 1.745.861 1.746.86 1.745.856\" style=\"stroke-width:1.2\"\u002F>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\">\u003Cg transform=\"translate(179.94 -33.135)\">\u003Cpath d=\"M-51.162-17.664L-52.895-17.664L-52.895-17.944Q-52.669-17.944-52.520-17.978Q-52.372-18.013-52.372-18.153L-52.372-20.402L-52.960-20.402L-52.960-20.682L-52.372-20.682L-52.372-21.499Q-52.372-21.817-52.194-22.065Q-52.016-22.312-51.726-22.453Q-51.435-22.593-51.124-22.593Q-50.868-22.593-50.664-22.451Q-50.461-22.309-50.461-22.066Q-50.461-21.930-50.560-21.831Q-50.659-21.731-50.796-21.731Q-50.933-21.731-51.032-21.831Q-51.131-21.930-51.131-22.066Q-51.131-22.247-50.991-22.340Q-51.069-22.367-51.169-22.367Q-51.377-22.367-51.531-22.234Q-51.685-22.101-51.765-21.897Q-51.845-21.694-51.845-21.485L-51.845-20.682L-50.957-20.682L-50.957-20.402L-51.818-20.402L-51.818-18.153Q-51.818-17.944-51.162-17.944L-51.162-17.664M-50.423-18.392Q-50.423-18.724-50.200-18.951Q-49.976-19.178-49.632-19.306Q-49.289-19.435-48.916-19.487Q-48.544-19.540-48.239-19.540L-48.239-19.793Q-48.239-19.998-48.347-20.178Q-48.455-20.357-48.636-20.460Q-48.817-20.562-49.025-20.562Q-49.432-20.562-49.668-20.470Q-49.579-20.433-49.533-20.349Q-49.487-20.265-49.487-20.163Q-49.487-20.067-49.533-19.988Q-49.579-19.910-49.659-19.865Q-49.740-19.821-49.829-19.821Q-49.979-19.821-50.080-19.918Q-50.181-20.016-50.181-20.163Q-50.181-20.785-49.025-20.785Q-48.814-20.785-48.564-20.721Q-48.315-20.658-48.113-20.539Q-47.911-20.419-47.785-20.234Q-47.658-20.050-47.658-19.807L-47.658-18.231Q-47.658-18.115-47.597-18.019Q-47.535-17.924-47.422-17.924Q-47.313-17.924-47.248-18.018Q-47.183-18.112-47.183-18.231L-47.183-18.679L-46.917-18.679L-46.917-18.231Q-46.917-17.961-47.144-17.796Q-47.371-17.630-47.651-17.630Q-47.860-17.630-47.997-17.784Q-48.133-17.937-48.157-18.153Q-48.304-17.886-48.586-17.741Q-48.868-17.596-49.193-17.596Q-49.470-17.596-49.753-17.671Q-50.037-17.746-50.230-17.925Q-50.423-18.105-50.423-18.392M-49.808-18.392Q-49.808-18.218-49.707-18.088Q-49.607-17.958-49.451-17.888Q-49.295-17.818-49.131-17.818Q-48.913-17.818-48.704-17.915Q-48.496-18.013-48.368-18.194Q-48.239-18.375-48.239-18.601L-48.239-19.329Q-48.564-19.329-48.930-19.238Q-49.295-19.147-49.552-18.935Q-49.808-18.724-49.808-18.392M-46.500-17.671L-46.500-18.734Q-46.500-18.758-46.472-18.785Q-46.445-18.812-46.421-18.812L-46.312-18.812Q-46.247-18.812-46.233-18.754Q-46.137-18.320-45.891-18.069Q-45.645-17.818-45.232-17.818Q-44.890-17.818-44.637-17.951Q-44.384-18.084-44.384-18.392Q-44.384-18.549-44.478-18.664Q-44.572-18.778-44.710-18.847Q-44.849-18.915-45.016-18.953L-45.597-19.052Q-45.953-19.120-46.226-19.341Q-46.500-19.561-46.500-19.903Q-46.500-20.152-46.389-20.327Q-46.277-20.501-46.091-20.600Q-45.905-20.699-45.690-20.742Q-45.474-20.785-45.232-20.785Q-44.818-20.785-44.538-20.603L-44.322-20.778Q-44.312-20.781-44.305-20.783Q-44.298-20.785-44.288-20.785L-44.237-20.785Q-44.210-20.785-44.186-20.761Q-44.162-20.737-44.162-20.709L-44.162-19.862Q-44.162-19.841-44.186-19.814Q-44.210-19.787-44.237-19.787L-44.350-19.787Q-44.377-19.787-44.403-19.812Q-44.428-19.838-44.428-19.862Q-44.428-20.098-44.534-20.262Q-44.640-20.426-44.823-20.508Q-45.006-20.590-45.238-20.590Q-45.566-20.590-45.823-20.487Q-46.079-20.385-46.079-20.108Q-46.079-19.913-45.896-19.804Q-45.713-19.694-45.484-19.653L-44.910-19.547Q-44.664-19.499-44.451-19.371Q-44.237-19.243-44.100-19.040Q-43.963-18.836-43.963-18.587Q-43.963-18.074-44.329-17.835Q-44.695-17.596-45.232-17.596Q-45.727-17.596-46.059-17.890L-46.325-17.616Q-46.346-17.596-46.373-17.596L-46.421-17.596Q-46.445-17.596-46.472-17.623Q-46.500-17.650-46.500-17.671M-42.808-18.505L-42.808-20.402L-43.447-20.402L-43.447-20.624Q-43.129-20.624-42.912-20.834Q-42.695-21.044-42.595-21.354Q-42.494-21.663-42.494-21.971L-42.227-21.971L-42.227-20.682L-41.150-20.682L-41.150-20.402L-42.227-20.402L-42.227-18.518Q-42.227-18.242-42.123-18.043Q-42.019-17.845-41.759-17.845Q-41.602-17.845-41.496-17.949Q-41.390-18.054-41.340-18.207Q-41.291-18.361-41.291-18.518L-41.291-18.932L-41.024-18.932L-41.024-18.505Q-41.024-18.279-41.123-18.069Q-41.222-17.859-41.407-17.727Q-41.591-17.596-41.820-17.596Q-42.258-17.596-42.533-17.833Q-42.808-18.071-42.808-18.505\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-warn)\" stroke=\"var(--tk-warn)\">\u003Cg fill=\"var(--tk-warn)\" stroke=\"none\" font-family=\"cmr7\" font-size=\"7\">\u003Cg transform=\"translate(179.94 18.08)\">\u003Cpath d=\"M-52.960-17.671L-52.960-18.734Q-52.960-18.758-52.932-18.785Q-52.905-18.812-52.881-18.812L-52.772-18.812Q-52.707-18.812-52.693-18.754Q-52.597-18.320-52.351-18.069Q-52.105-17.818-51.691-17.818Q-51.350-17.818-51.097-17.951Q-50.844-18.084-50.844-18.392Q-50.844-18.549-50.938-18.664Q-51.032-18.778-51.170-18.847Q-51.309-18.915-51.476-18.953L-52.057-19.052Q-52.413-19.120-52.686-19.341Q-52.960-19.561-52.960-19.903Q-52.960-20.152-52.848-20.327Q-52.737-20.501-52.551-20.600Q-52.365-20.699-52.149-20.742Q-51.934-20.785-51.691-20.785Q-51.278-20.785-50.998-20.603L-50.782-20.778Q-50.772-20.781-50.765-20.783Q-50.758-20.785-50.748-20.785L-50.697-20.785Q-50.670-20.785-50.646-20.761Q-50.622-20.737-50.622-20.709L-50.622-19.862Q-50.622-19.841-50.646-19.814Q-50.670-19.787-50.697-19.787L-50.810-19.787Q-50.837-19.787-50.863-19.812Q-50.888-19.838-50.888-19.862Q-50.888-20.098-50.994-20.262Q-51.100-20.426-51.283-20.508Q-51.466-20.590-51.698-20.590Q-52.026-20.590-52.283-20.487Q-52.539-20.385-52.539-20.108Q-52.539-19.913-52.356-19.804Q-52.173-19.694-51.944-19.653L-51.370-19.547Q-51.124-19.499-50.910-19.371Q-50.697-19.243-50.560-19.040Q-50.423-18.836-50.423-18.587Q-50.423-18.074-50.789-17.835Q-51.155-17.596-51.691-17.596Q-52.187-17.596-52.519-17.890L-52.785-17.616Q-52.806-17.596-52.833-17.596L-52.881-17.596Q-52.905-17.596-52.932-17.623Q-52.960-17.650-52.960-17.671M-48.127-17.664L-49.730-17.664L-49.730-17.944Q-49.504-17.944-49.355-17.978Q-49.207-18.013-49.207-18.153L-49.207-21.772Q-49.207-22.042-49.314-22.104Q-49.422-22.165-49.730-22.165L-49.730-22.446L-48.653-22.521L-48.653-18.153Q-48.653-18.016-48.503-17.980Q-48.352-17.944-48.127-17.944L-48.127-17.664M-47.573-19.147Q-47.573-19.489-47.438-19.788Q-47.303-20.087-47.064-20.311Q-46.824-20.535-46.506-20.660Q-46.189-20.785-45.857-20.785Q-45.413-20.785-45.013-20.569Q-44.613-20.354-44.379-19.976Q-44.145-19.599-44.145-19.147Q-44.145-18.806-44.286-18.522Q-44.428-18.238-44.673-18.031Q-44.917-17.825-45.226-17.710Q-45.536-17.596-45.857-17.596Q-46.288-17.596-46.689-17.797Q-47.091-17.999-47.332-18.351Q-47.573-18.703-47.573-19.147M-45.857-17.845Q-45.255-17.845-45.032-18.223Q-44.808-18.601-44.808-19.233Q-44.808-19.845-45.042-20.204Q-45.276-20.562-45.857-20.562Q-46.910-20.562-46.910-19.233Q-46.910-18.601-46.684-18.223Q-46.459-17.845-45.857-17.845\" fill=\"var(--tk-warn)\" stroke=\"var(--tk-warn)\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(179.94 18.08)\">\u003Cpath d=\"M-42.369-17.691L-43.350-20.190Q-43.411-20.333-43.529-20.368Q-43.647-20.402-43.863-20.402L-43.863-20.682L-42.383-20.682L-42.383-20.402Q-42.762-20.402-42.762-20.241Q-42.762-20.231-42.748-20.190L-42.034-18.358L-41.361-20.063Q-41.391-20.135-41.391-20.163Q-41.391-20.190-41.419-20.190Q-41.480-20.337-41.598-20.369Q-41.716-20.402-41.928-20.402L-41.928-20.682L-40.530-20.682L-40.530-20.402Q-40.906-20.402-40.906-20.241Q-40.906-20.210-40.899-20.190L-40.144-18.252L-39.457-20.002Q-39.436-20.053-39.436-20.108Q-39.436-20.248-39.549-20.325Q-39.662-20.402-39.802-20.402L-39.802-20.682L-38.582-20.682L-38.582-20.402Q-38.787-20.402-38.942-20.296Q-39.098-20.190-39.170-20.002L-40.075-17.691Q-40.110-17.596-40.222-17.596L-40.291-17.596Q-40.400-17.596-40.438-17.691L-41.220-19.694L-42.007-17.691Q-42.041-17.596-42.154-17.596L-42.222-17.596Q-42.331-17.596-42.369-17.691\" fill=\"var(--tk-warn)\" stroke=\"var(--tk-warn)\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fsvg>\u003Cfigcaption class=\"tikz-cap\">Sinusoidal positional encodings. Each curve is one embedding dimension over position \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.625em;vertical-align:-0.1944em;\">\u003C\u002Fspan>\u003Cspan class=\"mord mathnormal\">p\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>: low dimensions (blue) oscillate fast, high ones (red) slowly.\u003C\u002Ffigcaption>","\u003Csvg style=\"width:100%;max-width:241.901px;height:auto\" xmlns=\"http:\u002F\u002Fwww.w3.org\u002F2000\u002Fsvg\" viewBox=\"-75 -75 181.426 209.998\">\u003Cg stroke=\"currentColor\" style=\"stroke-miterlimit:10;stroke-width:.4\">\u003Cg transform=\"translate(-9.917 1.905)\">\u003Cpath d=\"M-3.355 124.934L-5.133 124.934L-5.133 124.637Q-4.859 124.637-4.691 124.590Q-4.523 124.543-4.523 124.375L-4.523 122.239Q-4.523 122.024-4.580 121.928Q-4.637 121.832-4.750 121.811Q-4.863 121.789-5.109 121.789L-5.109 121.493L-3.910 121.407L-3.910 124.375Q-3.910 124.543-3.764 124.590Q-3.617 124.637-3.355 124.637L-3.355 124.934M-4.797 120.012Q-4.797 119.821-4.662 119.690Q-4.527 119.559-4.332 119.559Q-4.211 119.559-4.107 119.621Q-4.004 119.684-3.941 119.788Q-3.879 119.891-3.879 120.012Q-3.879 120.207-4.010 120.342Q-4.141 120.477-4.332 120.477Q-4.531 120.477-4.664 120.344Q-4.797 120.211-4.797 120.012M-0.926 124.934L-2.781 124.934L-2.781 124.637Q-2.508 124.637-2.340 124.590Q-2.172 124.543-2.172 124.375L-2.172 122.239Q-2.172 122.024-2.234 121.928Q-2.297 121.832-2.416 121.811Q-2.535 121.789-2.781 121.789L-2.781 121.493L-1.590 121.407L-1.590 122.141Q-1.476 121.926-1.283 121.758Q-1.090 121.590-0.851 121.498Q-0.613 121.407-0.359 121.407Q0.809 121.407 0.809 122.485L0.809 124.375Q0.809 124.543 0.979 124.590Q1.149 124.637 1.418 124.637L1.418 124.934L-0.437 124.934L-0.437 124.637Q-0.164 124.637 0.004 124.590Q0.172 124.543 0.172 124.375L0.172 122.500Q0.172 122.118 0.051 121.889Q-0.070 121.661-0.422 121.661Q-0.734 121.661-0.988 121.823Q-1.242 121.985-1.389 122.254Q-1.535 122.524-1.535 122.821L-1.535 124.375Q-1.535 124.543-1.365 124.590Q-1.195 124.637-0.926 124.637L-0.926 124.934M3.746 126.485L1.891 126.485L1.891 126.192Q2.160 126.192 2.328 126.147Q2.496 126.102 2.496 125.926L2.496 122.102Q2.496 121.895 2.340 121.842Q2.184 121.789 1.891 121.789L1.891 121.493L3.113 121.407L3.113 121.871Q3.344 121.649 3.658 121.528Q3.973 121.407 4.313 121.407Q4.785 121.407 5.190 121.653Q5.594 121.899 5.826 122.315Q6.059 122.731 6.059 123.207Q6.059 123.582 5.910 123.911Q5.762 124.239 5.492 124.491Q5.223 124.743 4.879 124.877Q4.535 125.012 4.176 125.012Q3.887 125.012 3.615 124.891Q3.344 124.770 3.137 124.559L3.137 125.926Q3.137 126.102 3.305 126.147Q3.473 126.192 3.746 126.192L3.746 126.485M3.137 122.270L3.137 124.110Q3.289 124.399 3.551 124.579Q3.813 124.758 4.121 124.758Q4.406 124.758 4.629 124.620Q4.852 124.481 5.004 124.250Q5.156 124.020 5.234 123.748Q5.313 123.477 5.313 123.207Q5.313 122.875 5.188 122.518Q5.063 122.161 4.815 121.924Q4.567 121.688 4.219 121.688Q3.895 121.688 3.600 121.844Q3.305 122 3.137 122.270M7.266 123.981L7.266 122.239Q7.266 122.024 7.203 121.928Q7.141 121.832 7.022 121.811Q6.902 121.789 6.656 121.789L6.656 121.493L7.902 121.407L7.902 123.957L7.902 123.981Q7.902 124.293 7.957 124.455Q8.012 124.618 8.162 124.688Q8.313 124.758 8.633 124.758Q9.063 124.758 9.336 124.420Q9.610 124.082 9.610 123.637L9.610 122.239Q9.610 122.024 9.547 121.928Q9.485 121.832 9.365 121.811Q9.246 121.789 9 121.789L9 121.493L10.246 121.407L10.246 124.192Q10.246 124.403 10.309 124.498Q10.371 124.594 10.490 124.616Q10.610 124.637 10.856 124.637L10.856 124.934L9.633 125.012L9.633 124.391Q9.465 124.680 9.184 124.846Q8.902 125.012 8.582 125.012Q7.266 125.012 7.266 123.981M11.926 123.973L11.926 121.782L11.223 121.782L11.223 121.528Q11.578 121.528 11.820 121.295Q12.063 121.063 12.174 120.715Q12.285 120.368 12.285 120.012L12.567 120.012L12.567 121.485L13.742 121.485L13.742 121.782L12.567 121.782L12.567 123.957Q12.567 124.278 12.686 124.506Q12.805 124.735 13.086 124.735Q13.266 124.735 13.383 124.612Q13.500 124.489 13.553 124.309Q13.606 124.129 13.606 123.957L13.606 123.485L13.887 123.485L13.887 123.973Q13.887 124.227 13.781 124.467Q13.676 124.707 13.479 124.860Q13.281 125.012 13.024 125.012Q12.707 125.012 12.455 124.889Q12.203 124.766 12.065 124.532Q11.926 124.297 11.926 123.973\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M-65.203 100.75H54.298V75.141H-65.203Z\"\u002F>\u003Cg fill=\"var(--tk-accent)\" stroke=\"none\" font-family=\"cmtt8\" font-size=\"8\">\u003Cg transform=\"translate(-53.126 -34.544)\">\u003Cpath d=\"M-5.508 124.696L-5.508 124.606Q-5.457 124.395-5.262 124.375L-5.062 124.375L-5.062 122.047L-5.262 122.047Q-5.469 122.024-5.508 121.805L-5.508 121.719Q-5.457 121.504-5.262 121.485L-4.781 121.485Q-4.590 121.508-4.531 121.719Q-4.211 121.446-3.797 121.446Q-3.605 121.446-3.437 121.555Q-3.269 121.664-3.195 121.836Q-3.019 121.645-2.797 121.545Q-2.574 121.446-2.332 121.446Q-1.914 121.446-1.760 121.788Q-1.605 122.129-1.605 122.598L-1.605 124.375L-1.406 124.375Q-1.195 124.399-1.156 124.606L-1.156 124.696Q-1.207 124.911-1.406 124.934L-2.223 124.934Q-2.418 124.911-2.469 124.696L-2.469 124.606Q-2.418 124.399-2.223 124.375L-2.133 124.375L-2.133 122.629Q-2.133 122.004-2.383 122.004Q-2.715 122.004-2.892 122.315Q-3.070 122.625-3.070 122.989L-3.070 124.375L-2.867 124.375Q-2.660 124.399-2.621 124.606L-2.621 124.696Q-2.672 124.911-2.867 124.934L-3.683 124.934Q-3.883 124.911-3.933 124.696L-3.933 124.606Q-3.883 124.399-3.683 124.375L-3.598 124.375L-3.598 122.629Q-3.598 122.004-3.844 122.004Q-4.176 122.004-4.353 122.317Q-4.531 122.629-4.531 122.989L-4.531 124.375L-4.332 124.375Q-4.125 124.399-4.086 124.606L-4.086 124.696Q-4.137 124.914-4.332 124.934L-5.262 124.934Q-5.469 124.911-5.508 124.696M-0.465 124.079L-0.465 122.047L-0.887 122.047Q-1.094 122.024-1.137 121.805L-1.137 121.719Q-1.090 121.508-0.887 121.485L-0.070 121.485Q0.125 121.508 0.176 121.719L0.176 124.047Q0.176 124.282 0.346 124.348Q0.516 124.414 0.801 124.414Q1.008 124.414 1.203 124.338Q1.399 124.262 1.524 124.112Q1.649 123.961 1.649 123.750L1.649 122.047L1.227 122.047Q1.016 122.024 0.977 121.805L0.977 121.719Q1.016 121.508 1.227 121.485L2.039 121.485Q2.238 121.508 2.289 121.719L2.289 124.375L2.715 124.375Q2.922 124.399 2.961 124.606L2.961 124.696Q2.922 124.911 2.715 124.934L1.899 124.934Q1.699 124.911 1.649 124.711Q1.246 124.973 0.738 124.973Q0.504 124.973 0.289 124.932Q0.074 124.891-0.092 124.789Q-0.258 124.688-0.361 124.510Q-0.465 124.332-0.465 124.079M3.488 124.696L3.488 124.606Q3.539 124.399 3.734 124.375L4.840 124.375L4.840 120.606L3.734 120.606Q3.539 120.582 3.488 120.368L3.488 120.278Q3.539 120.071 3.734 120.047L5.231 120.047Q5.422 120.071 5.481 120.278L5.481 124.375L6.582 124.375Q6.781 124.399 6.832 124.606L6.832 124.696Q6.781 124.911 6.582 124.934L3.734 124.934Q3.539 124.911 3.488 124.696M8.485 123.829L8.485 122.047L7.734 122.047Q7.535 122.024 7.484 121.805L7.484 121.719Q7.535 121.508 7.734 121.485L8.485 121.485L8.485 120.735Q8.535 120.528 8.735 120.500L8.879 120.500Q9.074 120.528 9.125 120.735L9.125 121.485L10.485 121.485Q10.676 121.504 10.735 121.719L10.735 121.805Q10.680 122.024 10.485 122.047L9.125 122.047L9.125 123.797Q9.125 124.414 9.699 124.414Q9.949 124.414 10.113 124.229Q10.277 124.043 10.277 123.797Q10.277 123.704 10.350 123.633Q10.422 123.563 10.524 123.551L10.668 123.551Q10.867 123.575 10.918 123.782L10.918 123.829Q10.918 124.153 10.735 124.416Q10.551 124.680 10.258 124.827Q9.965 124.973 9.645 124.973Q9.133 124.973 8.809 124.663Q8.485 124.352 8.485 123.829M12.149 124.696L12.149 124.606Q12.199 124.399 12.395 124.375L13.434 124.375L13.434 122.047L12.461 122.047Q12.262 122.024 12.211 121.805L12.211 121.719Q12.262 121.508 12.461 121.485L13.828 121.485Q14.024 121.504 14.074 121.719L14.074 124.375L14.988 124.375Q15.184 124.399 15.235 124.606L15.235 124.696Q15.184 124.911 14.988 124.934L12.395 124.934Q12.199 124.911 12.149 124.696M13.180 120.508L13.180 120.454Q13.180 120.282 13.317 120.161Q13.453 120.039 13.629 120.039Q13.801 120.039 13.938 120.161Q14.074 120.282 14.074 120.454L14.074 120.508Q14.074 120.684 13.938 120.805Q13.801 120.926 13.629 120.926Q13.453 120.926 13.317 120.805Q13.180 120.684 13.180 120.508M19.274 122.797L16.531 122.797Q16.406 122.786 16.320 122.700Q16.235 122.614 16.235 122.485Q16.235 122.356 16.320 122.274Q16.406 122.192 16.531 122.180L19.274 122.180Q19.399 122.192 19.481 122.274Q19.563 122.356 19.563 122.485Q19.563 122.614 19.481 122.700Q19.399 122.786 19.274 122.797M20.094 124.696L20.094 124.606Q20.137 124.399 20.344 124.375L20.766 124.375L20.766 120.606L20.344 120.606Q20.137 120.582 20.094 120.368L20.094 120.278Q20.137 120.071 20.344 120.047L21.160 120.047Q21.356 120.071 21.406 120.278L21.406 121.829Q21.867 121.446 22.465 121.446Q22.813 121.446 23.051 121.586Q23.289 121.727 23.404 121.985Q23.520 122.243 23.520 122.598L23.520 124.375L23.945 124.375Q24.152 124.399 24.192 124.606L24.192 124.696Q24.152 124.911 23.945 124.934L22.551 124.934Q22.356 124.911 22.305 124.696L22.305 124.606Q22.356 124.395 22.551 124.375L22.879 124.375L22.879 122.629Q22.879 122.321 22.789 122.163Q22.699 122.004 22.406 122.004Q22.137 122.004 21.908 122.135Q21.680 122.266 21.543 122.495Q21.406 122.723 21.406 122.989L21.406 124.375L21.832 124.375Q22.039 124.399 22.078 124.606L22.078 124.696Q22.039 124.911 21.832 124.934L20.344 124.934Q20.137 124.911 20.094 124.696M27.781 123.446L25.340 123.446Q25.395 123.723 25.592 123.946Q25.789 124.168 26.067 124.291Q26.344 124.414 26.629 124.414Q27.102 124.414 27.324 124.125Q27.332 124.114 27.389 124.008Q27.445 123.903 27.494 123.860Q27.543 123.817 27.637 123.805L27.781 123.805Q27.973 123.825 28.031 124.039L28.031 124.094Q27.965 124.395 27.735 124.592Q27.504 124.789 27.192 124.881Q26.879 124.973 26.574 124.973Q26.090 124.973 25.651 124.745Q25.211 124.516 24.943 124.116Q24.676 123.715 24.676 123.223L24.676 123.164Q24.676 122.696 24.922 122.293Q25.168 121.891 25.576 121.657Q25.985 121.422 26.453 121.422Q26.957 121.422 27.311 121.645Q27.664 121.868 27.848 122.256Q28.031 122.645 28.031 123.149L28.031 123.207Q27.973 123.422 27.781 123.446M25.348 122.895L27.375 122.895Q27.328 122.485 27.090 122.233Q26.852 121.981 26.453 121.981Q26.059 121.981 25.752 122.243Q25.445 122.504 25.348 122.895M28.922 123.821Q28.922 123.375 29.336 123.118Q29.750 122.860 30.291 122.760Q30.832 122.661 31.340 122.653Q31.340 122.438 31.205 122.286Q31.070 122.133 30.863 122.057Q30.656 121.981 30.445 121.981Q30.102 121.981 29.942 122.004L29.942 122.063Q29.942 122.231 29.822 122.346Q29.703 122.461 29.539 122.461Q29.363 122.461 29.248 122.338Q29.133 122.215 29.133 122.047Q29.133 121.641 29.514 121.532Q29.895 121.422 30.453 121.422Q30.723 121.422 30.990 121.500Q31.258 121.579 31.483 121.729Q31.707 121.879 31.844 122.100Q31.981 122.321 31.981 122.598L31.981 124.317Q31.981 124.375 32.508 124.375Q32.703 124.395 32.754 124.606L32.754 124.696Q32.703 124.911 32.508 124.934L32.363 124.934Q32.020 124.934 31.791 124.887Q31.563 124.840 31.418 124.653Q30.957 124.973 30.250 124.973Q29.914 124.973 29.610 124.832Q29.305 124.692 29.113 124.430Q28.922 124.168 28.922 123.821M29.563 123.829Q29.563 124.102 29.805 124.258Q30.047 124.414 30.332 124.414Q30.551 124.414 30.783 124.356Q31.016 124.297 31.178 124.159Q31.340 124.020 31.340 123.797L31.340 123.207Q31.059 123.207 30.643 123.264Q30.227 123.321 29.895 123.459Q29.563 123.598 29.563 123.829M34.625 124.973Q34.160 124.973 33.795 124.723Q33.430 124.473 33.225 124.069Q33.020 123.664 33.020 123.207Q33.020 122.864 33.145 122.543Q33.270 122.223 33.502 121.973Q33.734 121.723 34.039 121.584Q34.344 121.446 34.699 121.446Q35.211 121.446 35.617 121.766L35.617 120.606L35.195 120.606Q34.984 120.582 34.945 120.368L34.945 120.278Q34.984 120.071 35.195 120.047L36.008 120.047Q36.207 120.071 36.258 120.278L36.258 124.375L36.684 124.375Q36.891 124.399 36.930 124.606L36.930 124.696Q36.891 124.911 36.684 124.934L35.867 124.934Q35.668 124.911 35.617 124.696L35.617 124.567Q35.422 124.758 35.160 124.866Q34.899 124.973 34.625 124.973M34.664 124.414Q35.024 124.414 35.276 124.145Q35.527 123.875 35.617 123.500L35.617 122.668Q35.559 122.481 35.432 122.330Q35.305 122.180 35.127 122.092Q34.949 122.004 34.754 122.004Q34.445 122.004 34.195 122.174Q33.945 122.344 33.801 122.629Q33.656 122.914 33.656 123.215Q33.656 123.664 33.942 124.039Q34.227 124.414 34.664 124.414\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-53.126 -34.544)\">\u003Cpath d=\"M41.884 124.735L41.884 123.821Q41.911 123.614 42.122 123.590L42.290 123.590Q42.454 123.614 42.513 123.774Q42.716 124.414 43.443 124.414Q43.650 124.414 43.878 124.379Q44.107 124.344 44.275 124.229Q44.443 124.114 44.443 123.911Q44.443 123.700 44.220 123.586Q43.997 123.473 43.724 123.430L43.025 123.317Q41.884 123.106 41.884 122.383Q41.884 122.094 42.028 121.905Q42.173 121.715 42.413 121.608Q42.653 121.500 42.909 121.461Q43.165 121.422 43.443 121.422Q43.693 121.422 43.886 121.452Q44.079 121.481 44.243 121.559Q44.321 121.442 44.450 121.422L44.528 121.422Q44.626 121.434 44.689 121.496Q44.751 121.559 44.763 121.653L44.763 122.360Q44.751 122.454 44.689 122.520Q44.626 122.586 44.528 122.598L44.361 122.598Q44.267 122.586 44.200 122.520Q44.134 122.454 44.122 122.360Q44.122 121.981 43.427 121.981Q43.079 121.981 42.761 122.063Q42.443 122.145 42.443 122.391Q42.443 122.657 43.114 122.766L43.818 122.887Q44.302 122.969 44.652 123.217Q45.001 123.465 45.001 123.911Q45.001 124.301 44.765 124.543Q44.528 124.786 44.179 124.879Q43.829 124.973 43.443 124.973Q42.864 124.973 42.466 124.719Q42.396 124.844 42.347 124.901Q42.298 124.957 42.193 124.973L42.122 124.973Q41.907 124.950 41.884 124.735M49.056 123.446L46.614 123.446Q46.669 123.723 46.866 123.946Q47.064 124.168 47.341 124.291Q47.618 124.414 47.903 124.414Q48.376 124.414 48.599 124.125Q48.607 124.114 48.663 124.008Q48.720 123.903 48.769 123.860Q48.818 123.817 48.911 123.805L49.056 123.805Q49.247 123.825 49.306 124.039L49.306 124.094Q49.239 124.395 49.009 124.592Q48.778 124.789 48.466 124.881Q48.153 124.973 47.849 124.973Q47.364 124.973 46.925 124.745Q46.486 124.516 46.218 124.116Q45.950 123.715 45.950 123.223L45.950 123.164Q45.950 122.696 46.196 122.293Q46.443 121.891 46.851 121.657Q47.259 121.422 47.728 121.422Q48.232 121.422 48.585 121.645Q48.939 121.868 49.122 122.256Q49.306 122.645 49.306 123.149L49.306 123.207Q49.247 123.422 49.056 123.446M46.622 122.895L48.650 122.895Q48.603 122.485 48.364 122.233Q48.126 121.981 47.728 121.981Q47.333 121.981 47.027 122.243Q46.720 122.504 46.622 122.895M50.239 124.696L50.239 124.606Q50.290 124.399 50.486 124.375L51.591 124.375L51.591 120.606L50.486 120.606Q50.290 120.582 50.239 120.368L50.239 120.278Q50.290 120.071 50.486 120.047L51.982 120.047Q52.173 120.071 52.232 120.278L52.232 124.375L53.333 124.375Q53.532 124.399 53.583 124.606L53.583 124.696Q53.532 124.911 53.333 124.934L50.486 124.934Q50.290 124.911 50.239 124.696M54.357 124.696L54.357 124.606Q54.407 124.399 54.603 124.375L55.486 124.375L55.486 122.047L54.630 122.047Q54.431 122.024 54.380 121.805L54.380 121.719Q54.431 121.508 54.630 121.485L55.486 121.485L55.486 121.032Q55.486 120.567 55.892 120.286Q56.298 120.004 56.778 120.004Q57.091 120.004 57.335 120.125Q57.579 120.246 57.579 120.528Q57.579 120.692 57.470 120.809Q57.361 120.926 57.196 120.926Q57.048 120.926 56.927 120.821Q56.806 120.715 56.806 120.567L56.724 120.567Q56.571 120.567 56.435 120.627Q56.298 120.688 56.212 120.803Q56.126 120.918 56.126 121.063L56.126 121.485L57.157 121.485Q57.353 121.504 57.403 121.719L57.403 121.805Q57.353 122.024 57.157 122.047L56.126 122.047L56.126 124.375L57.005 124.375Q57.200 124.399 57.251 124.606L57.251 124.696Q57.200 124.911 57.005 124.934L54.603 124.934Q54.407 124.911 54.357 124.696M61.778 122.797L59.036 122.797Q58.911 122.786 58.825 122.700Q58.739 122.614 58.739 122.485Q58.739 122.356 58.825 122.274Q58.911 122.192 59.036 122.180L61.778 122.180Q61.903 122.192 61.986 122.274Q62.068 122.356 62.068 122.485Q62.068 122.614 61.986 122.700Q61.903 122.786 61.778 122.797M62.935 123.821Q62.935 123.375 63.349 123.118Q63.763 122.860 64.304 122.760Q64.845 122.661 65.353 122.653Q65.353 122.438 65.218 122.286Q65.083 122.133 64.876 122.057Q64.669 121.981 64.458 121.981Q64.114 121.981 63.954 122.004L63.954 122.063Q63.954 122.231 63.835 122.346Q63.716 122.461 63.552 122.461Q63.376 122.461 63.261 122.338Q63.146 122.215 63.146 122.047Q63.146 121.641 63.527 121.532Q63.907 121.422 64.466 121.422Q64.736 121.422 65.003 121.500Q65.271 121.579 65.495 121.729Q65.720 121.879 65.857 122.100Q65.993 122.321 65.993 122.598L65.993 124.317Q65.993 124.375 66.521 124.375Q66.716 124.395 66.767 124.606L66.767 124.696Q66.716 124.911 66.521 124.934L66.376 124.934Q66.032 124.934 65.804 124.887Q65.575 124.840 65.431 124.653Q64.970 124.973 64.263 124.973Q63.927 124.973 63.622 124.832Q63.318 124.692 63.126 124.430Q62.935 124.168 62.935 123.821M63.575 123.829Q63.575 124.102 63.818 124.258Q64.060 124.414 64.345 124.414Q64.564 124.414 64.796 124.356Q65.028 124.297 65.191 124.159Q65.353 124.020 65.353 123.797L65.353 123.207Q65.071 123.207 64.655 123.264Q64.239 123.321 63.907 123.459Q63.575 123.598 63.575 123.829M67.974 123.829L67.974 122.047L67.224 122.047Q67.025 122.024 66.974 121.805L66.974 121.719Q67.025 121.508 67.224 121.485L67.974 121.485L67.974 120.735Q68.025 120.528 68.224 120.500L68.368 120.500Q68.564 120.528 68.614 120.735L68.614 121.485L69.974 121.485Q70.165 121.504 70.224 121.719L70.224 121.805Q70.169 122.024 69.974 122.047L68.614 122.047L68.614 123.797Q68.614 124.414 69.189 124.414Q69.439 124.414 69.603 124.229Q69.767 124.043 69.767 123.797Q69.767 123.704 69.839 123.633Q69.911 123.563 70.013 123.551L70.157 123.551Q70.357 123.575 70.407 123.782L70.407 123.829Q70.407 124.153 70.224 124.416Q70.040 124.680 69.747 124.827Q69.454 124.973 69.134 124.973Q68.622 124.973 68.298 124.663Q67.974 124.352 67.974 123.829M72.220 123.829L72.220 122.047L71.470 122.047Q71.271 122.024 71.220 121.805L71.220 121.719Q71.271 121.508 71.470 121.485L72.220 121.485L72.220 120.735Q72.271 120.528 72.470 120.500L72.614 120.500Q72.810 120.528 72.861 120.735L72.861 121.485L74.220 121.485Q74.411 121.504 74.470 121.719L74.470 121.805Q74.415 122.024 74.220 122.047L72.861 122.047L72.861 123.797Q72.861 124.414 73.435 124.414Q73.685 124.414 73.849 124.229Q74.013 124.043 74.013 123.797Q74.013 123.704 74.085 123.633Q74.157 123.563 74.259 123.551L74.403 123.551Q74.603 123.575 74.653 123.782L74.653 123.829Q74.653 124.153 74.470 124.416Q74.286 124.680 73.993 124.827Q73.700 124.973 73.380 124.973Q72.868 124.973 72.544 124.663Q72.220 124.352 72.220 123.829M78.778 123.446L76.337 123.446Q76.392 123.723 76.589 123.946Q76.786 124.168 77.064 124.291Q77.341 124.414 77.626 124.414Q78.099 124.414 78.321 124.125Q78.329 124.114 78.386 124.008Q78.443 123.903 78.491 123.860Q78.540 123.817 78.634 123.805L78.778 123.805Q78.970 123.825 79.028 124.039L79.028 124.094Q78.962 124.395 78.732 124.592Q78.501 124.789 78.189 124.881Q77.876 124.973 77.571 124.973Q77.087 124.973 76.648 124.745Q76.208 124.516 75.941 124.116Q75.673 123.715 75.673 123.223L75.673 123.164Q75.673 122.696 75.919 122.293Q76.165 121.891 76.573 121.657Q76.982 121.422 77.450 121.422Q77.954 121.422 78.308 121.645Q78.661 121.868 78.845 122.256Q79.028 122.645 79.028 123.149L79.028 123.207Q78.970 123.422 78.778 123.446M76.345 122.895L78.372 122.895Q78.325 122.485 78.087 122.233Q77.849 121.981 77.450 121.981Q77.056 121.981 76.749 122.243Q76.443 122.504 76.345 122.895M79.583 124.696L79.583 124.606Q79.626 124.399 79.833 124.375L80.255 124.375L80.255 122.047L79.833 122.047Q79.626 122.024 79.583 121.805L79.583 121.719Q79.630 121.508 79.833 121.485L80.650 121.485Q80.845 121.508 80.896 121.719L80.896 121.805L80.888 121.829Q81.114 121.649 81.388 121.547Q81.661 121.446 81.954 121.446Q82.302 121.446 82.540 121.586Q82.778 121.727 82.894 121.985Q83.009 122.243 83.009 122.598L83.009 124.375L83.435 124.375Q83.642 124.399 83.681 124.606L83.681 124.696Q83.642 124.911 83.435 124.934L82.040 124.934Q81.845 124.911 81.794 124.696L81.794 124.606Q81.845 124.395 82.040 124.375L82.368 124.375L82.368 122.629Q82.368 122.321 82.278 122.163Q82.189 122.004 81.896 122.004Q81.626 122.004 81.398 122.135Q81.169 122.266 81.032 122.495Q80.896 122.723 80.896 122.989L80.896 124.375L81.321 124.375Q81.528 124.399 81.568 124.606L81.568 124.696Q81.528 124.911 81.321 124.934L79.833 124.934Q79.626 124.911 79.583 124.696M84.958 123.829L84.958 122.047L84.208 122.047Q84.009 122.024 83.958 121.805L83.958 121.719Q84.009 121.508 84.208 121.485L84.958 121.485L84.958 120.735Q85.009 120.528 85.208 120.500L85.353 120.500Q85.548 120.528 85.599 120.735L85.599 121.485L86.958 121.485Q87.150 121.504 87.208 121.719L87.208 121.805Q87.153 122.024 86.958 122.047L85.599 122.047L85.599 123.797Q85.599 124.414 86.173 124.414Q86.423 124.414 86.587 124.229Q86.751 124.043 86.751 123.797Q86.751 123.704 86.823 123.633Q86.896 123.563 86.997 123.551L87.142 123.551Q87.341 123.575 87.392 123.782L87.392 123.829Q87.392 124.153 87.208 124.416Q87.025 124.680 86.732 124.827Q86.439 124.973 86.118 124.973Q85.607 124.973 85.282 124.663Q84.958 124.352 84.958 123.829M88.622 124.696L88.622 124.606Q88.673 124.399 88.868 124.375L89.907 124.375L89.907 122.047L88.935 122.047Q88.736 122.024 88.685 121.805L88.685 121.719Q88.736 121.508 88.935 121.485L90.302 121.485Q90.497 121.504 90.548 121.719L90.548 124.375L91.462 124.375Q91.657 124.399 91.708 124.606L91.708 124.696Q91.657 124.911 91.462 124.934L88.868 124.934Q88.673 124.911 88.622 124.696M89.653 120.508L89.653 120.454Q89.653 120.282 89.790 120.161Q89.927 120.039 90.103 120.039Q90.275 120.039 90.411 120.161Q90.548 120.282 90.548 120.454L90.548 120.508Q90.548 120.684 90.411 120.805Q90.275 120.926 90.103 120.926Q89.927 120.926 89.790 120.805Q89.653 120.684 89.653 120.508M94.372 124.973Q93.900 124.973 93.515 124.729Q93.130 124.485 92.907 124.075Q92.685 123.664 92.685 123.207Q92.685 122.864 92.810 122.541Q92.935 122.219 93.165 121.965Q93.396 121.711 93.702 121.567Q94.009 121.422 94.372 121.422Q94.736 121.422 95.048 121.569Q95.361 121.715 95.583 121.961Q95.806 122.207 95.933 122.528Q96.060 122.848 96.060 123.207Q96.060 123.664 95.835 124.077Q95.611 124.489 95.226 124.731Q94.841 124.973 94.372 124.973M94.372 124.414Q94.837 124.414 95.128 124.020Q95.419 123.625 95.419 123.141Q95.419 122.848 95.284 122.580Q95.150 122.313 94.909 122.147Q94.669 121.981 94.372 121.981Q94.068 121.981 93.829 122.147Q93.591 122.313 93.456 122.580Q93.321 122.848 93.321 123.141Q93.321 123.621 93.614 124.018Q93.907 124.414 94.372 124.414M96.568 124.696L96.568 124.606Q96.611 124.399 96.818 124.375L97.239 124.375L97.239 122.047L96.818 122.047Q96.611 122.024 96.568 121.805L96.568 121.719Q96.614 121.508 96.818 121.485L97.634 121.485Q97.829 121.508 97.880 121.719L97.880 121.805L97.872 121.829Q98.099 121.649 98.372 121.547Q98.646 121.446 98.939 121.446Q99.286 121.446 99.525 121.586Q99.763 121.727 99.878 121.985Q99.993 122.243 99.993 122.598L99.993 124.375L100.419 124.375Q100.626 124.399 100.665 124.606L100.665 124.696Q100.626 124.911 100.419 124.934L99.025 124.934Q98.829 124.911 98.778 124.696L98.778 124.606Q98.829 124.395 99.025 124.375L99.353 124.375L99.353 122.629Q99.353 122.321 99.263 122.163Q99.173 122.004 98.880 122.004Q98.611 122.004 98.382 122.135Q98.153 122.266 98.017 122.495Q97.880 122.723 97.880 122.989L97.880 124.375L98.306 124.375Q98.513 124.399 98.552 124.606L98.552 124.696Q98.513 124.911 98.306 124.934L96.818 124.934Q96.611 124.911 96.568 124.696\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-soft-neutral)\">\u003Cpath d=\"M-65.203 59.493H54.298V36.731H-65.203Z\"\u002F>\u003Cg fill=\"currentColor\" stroke=\"none\" font-family=\"cmtt8\" font-size=\"8\">\u003Cg transform=\"translate(-21.25 -74.378)\">\u003Cpath d=\"M-5.047 123.821Q-5.047 123.375-4.633 123.118Q-4.219 122.860-3.678 122.760Q-3.137 122.661-2.629 122.653Q-2.629 122.438-2.764 122.286Q-2.898 122.133-3.105 122.057Q-3.312 121.981-3.523 121.981Q-3.867 121.981-4.027 122.004L-4.027 122.063Q-4.027 122.231-4.146 122.346Q-4.266 122.461-4.430 122.461Q-4.605 122.461-4.721 122.338Q-4.836 122.215-4.836 122.047Q-4.836 121.641-4.455 121.532Q-4.074 121.422-3.516 121.422Q-3.246 121.422-2.978 121.500Q-2.711 121.579-2.486 121.729Q-2.262 121.879-2.125 122.100Q-1.988 122.321-1.988 122.598L-1.988 124.317Q-1.988 124.375-1.461 124.375Q-1.266 124.395-1.215 124.606L-1.215 124.696Q-1.266 124.911-1.461 124.934L-1.605 124.934Q-1.949 124.934-2.178 124.887Q-2.406 124.840-2.551 124.653Q-3.012 124.973-3.719 124.973Q-4.055 124.973-4.359 124.832Q-4.664 124.692-4.855 124.430Q-5.047 124.168-5.047 123.821M-4.406 123.829Q-4.406 124.102-4.164 124.258Q-3.922 124.414-3.637 124.414Q-3.418 124.414-3.185 124.356Q-2.953 124.297-2.791 124.159Q-2.629 124.020-2.629 123.797L-2.629 123.207Q-2.910 123.207-3.326 123.264Q-3.742 123.321-4.074 123.459Q-4.406 123.598-4.406 123.829M0.656 124.973Q0.192 124.973-0.174 124.723Q-0.539 124.473-0.744 124.069Q-0.949 123.664-0.949 123.207Q-0.949 122.864-0.824 122.543Q-0.699 122.223-0.467 121.973Q-0.234 121.723 0.070 121.584Q0.375 121.446 0.731 121.446Q1.242 121.446 1.649 121.766L1.649 120.606L1.227 120.606Q1.016 120.582 0.977 120.368L0.977 120.278Q1.016 120.071 1.227 120.047L2.039 120.047Q2.238 120.071 2.289 120.278L2.289 124.375L2.715 124.375Q2.922 124.399 2.961 124.606L2.961 124.696Q2.922 124.911 2.715 124.934L1.899 124.934Q1.699 124.911 1.649 124.696L1.649 124.567Q1.453 124.758 1.192 124.866Q0.930 124.973 0.656 124.973M0.695 124.414Q1.055 124.414 1.307 124.145Q1.559 123.875 1.649 123.500L1.649 122.668Q1.590 122.481 1.463 122.330Q1.336 122.180 1.158 122.092Q0.981 122.004 0.785 122.004Q0.477 122.004 0.227 122.174Q-0.023 122.344-0.168 122.629Q-0.312 122.914-0.312 123.215Q-0.312 123.664-0.027 124.039Q0.258 124.414 0.695 124.414M4.902 124.973Q4.438 124.973 4.072 124.723Q3.707 124.473 3.502 124.069Q3.297 123.664 3.297 123.207Q3.297 122.864 3.422 122.543Q3.547 122.223 3.779 121.973Q4.012 121.723 4.317 121.584Q4.621 121.446 4.977 121.446Q5.488 121.446 5.895 121.766L5.895 120.606L5.473 120.606Q5.262 120.582 5.223 120.368L5.223 120.278Q5.262 120.071 5.473 120.047L6.285 120.047Q6.484 120.071 6.535 120.278L6.535 124.375L6.961 124.375Q7.168 124.399 7.207 124.606L7.207 124.696Q7.168 124.911 6.961 124.934L6.145 124.934Q5.945 124.911 5.895 124.696L5.895 124.567Q5.699 124.758 5.438 124.866Q5.176 124.973 4.902 124.973M4.942 124.414Q5.301 124.414 5.553 124.145Q5.805 123.875 5.895 123.500L5.895 122.668Q5.836 122.481 5.709 122.330Q5.582 122.180 5.404 122.092Q5.227 122.004 5.031 122.004Q4.723 122.004 4.473 122.174Q4.223 122.344 4.078 122.629Q3.934 122.914 3.934 123.215Q3.934 123.664 4.219 124.039Q4.504 124.414 4.942 124.414\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-21.25 -74.378)\">\u003Cpath d=\"M12.763 125.012Q12.466 125.012 12.241 124.840Q12.017 124.668 11.894 124.393Q11.771 124.118 11.771 123.836L11.771 123.782Q11.771 123.524 11.862 123.288Q11.954 123.051 12.114 122.895L12.571 122.438Q12.349 121.801 12.349 121.149L12.349 121.094Q12.349 120.829 12.456 120.565Q12.564 120.301 12.769 120.133Q12.974 119.965 13.251 119.965Q13.528 119.965 13.708 120.137Q13.888 120.309 13.966 120.571Q14.044 120.832 14.044 121.094L14.044 121.149Q14.044 121.418 13.917 121.674Q13.790 121.930 13.624 122.129Q13.458 122.329 13.196 122.590Q13.376 123 13.739 123.567Q13.954 123.164 14.267 122.325L14.380 122.032Q14.239 121.996 14.196 121.805L14.196 121.719Q14.247 121.508 14.443 121.485L15.267 121.485Q15.466 121.504 15.517 121.719L15.517 121.805Q15.466 122.024 15.267 122.047L14.939 122.047L14.747 122.551Q14.364 123.575 14.083 124.047Q14.255 124.235 14.400 124.344Q14.544 124.454 14.724 124.454Q14.845 124.454 14.900 124.317Q14.954 124.180 14.954 124.032Q15.017 123.797 15.220 123.797Q15.423 123.797 15.486 124.032L15.486 124.086Q15.486 124.450 15.277 124.731Q15.068 125.012 14.724 125.012Q14.247 125.012 13.771 124.493Q13.314 125.012 12.763 125.012M12.763 124.454Q13.114 124.454 13.427 124.063Q13.013 123.422 12.806 122.996Q12.693 123.102 12.622 123.182Q12.552 123.262 12.503 123.344Q12.454 123.426 12.433 123.538Q12.411 123.649 12.411 123.782Q12.411 124 12.495 124.227Q12.579 124.454 12.763 124.454M12.989 122.004Q13.216 121.758 13.366 121.543Q13.517 121.329 13.517 121.118Q13.517 120.930 13.456 120.729Q13.396 120.528 13.251 120.528Q13.068 120.528 12.972 120.725Q12.876 120.922 12.876 121.133Q12.876 121.571 12.989 122.004\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-21.25 -74.378)\">\u003Cpath d=\"M20.118 124.696L20.118 124.606Q20.161 124.399 20.368 124.375L20.790 124.375L20.790 122.047L20.368 122.047Q20.161 122.024 20.118 121.805L20.118 121.719Q20.165 121.508 20.368 121.485L21.185 121.485Q21.380 121.508 21.431 121.719L21.431 121.805L21.423 121.829Q21.650 121.649 21.923 121.547Q22.196 121.446 22.489 121.446Q22.837 121.446 23.075 121.586Q23.314 121.727 23.429 121.985Q23.544 122.243 23.544 122.598L23.544 124.375L23.970 124.375Q24.177 124.399 24.216 124.606L24.216 124.696Q24.177 124.911 23.970 124.934L22.575 124.934Q22.380 124.911 22.329 124.696L22.329 124.606Q22.380 124.395 22.575 124.375L22.903 124.375L22.903 122.629Q22.903 122.321 22.814 122.163Q22.724 122.004 22.431 122.004Q22.161 122.004 21.933 122.135Q21.704 122.266 21.568 122.495Q21.431 122.723 21.431 122.989L21.431 124.375L21.857 124.375Q22.064 124.399 22.103 124.606L22.103 124.696Q22.064 124.911 21.857 124.934L20.368 124.934Q20.161 124.911 20.118 124.696M26.415 124.973Q25.943 124.973 25.558 124.729Q25.173 124.485 24.950 124.075Q24.728 123.664 24.728 123.207Q24.728 122.864 24.853 122.541Q24.978 122.219 25.208 121.965Q25.439 121.711 25.745 121.567Q26.052 121.422 26.415 121.422Q26.778 121.422 27.091 121.569Q27.403 121.715 27.626 121.961Q27.849 122.207 27.976 122.528Q28.103 122.848 28.103 123.207Q28.103 123.664 27.878 124.077Q27.653 124.489 27.269 124.731Q26.884 124.973 26.415 124.973M26.415 124.414Q26.880 124.414 27.171 124.020Q27.462 123.625 27.462 123.141Q27.462 122.848 27.327 122.580Q27.193 122.313 26.952 122.147Q26.712 121.981 26.415 121.981Q26.110 121.981 25.872 122.147Q25.634 122.313 25.499 122.580Q25.364 122.848 25.364 123.141Q25.364 123.621 25.657 124.018Q25.950 124.414 26.415 124.414M28.763 124.696L28.763 124.606Q28.821 124.399 29.013 124.375L29.724 124.375L29.724 122.047L29.013 122.047Q28.818 122.024 28.763 121.805L28.763 121.719Q28.821 121.508 29.013 121.485L30.114 121.485Q30.314 121.504 30.364 121.719L30.364 122.047Q30.626 121.762 30.982 121.604Q31.337 121.446 31.724 121.446Q32.017 121.446 32.251 121.580Q32.486 121.715 32.486 121.981Q32.486 122.149 32.376 122.266Q32.267 122.383 32.099 122.383Q31.946 122.383 31.831 122.272Q31.716 122.161 31.716 122.004Q31.341 122.004 31.027 122.205Q30.712 122.407 30.538 122.741Q30.364 123.075 30.364 123.454L30.364 124.375L31.310 124.375Q31.517 124.399 31.556 124.606L31.556 124.696Q31.517 124.911 31.310 124.934L29.013 124.934Q28.821 124.911 28.763 124.696M32.732 124.696L32.732 124.606Q32.782 124.395 32.978 124.375L33.177 124.375L33.177 122.047L32.978 122.047Q32.771 122.024 32.732 121.805L32.732 121.719Q32.782 121.504 32.978 121.485L33.458 121.485Q33.650 121.508 33.708 121.719Q34.028 121.446 34.443 121.446Q34.634 121.446 34.802 121.555Q34.970 121.664 35.044 121.836Q35.220 121.645 35.443 121.545Q35.665 121.446 35.907 121.446Q36.325 121.446 36.480 121.788Q36.634 122.129 36.634 122.598L36.634 124.375L36.833 124.375Q37.044 124.399 37.083 124.606L37.083 124.696Q37.032 124.911 36.833 124.934L36.017 124.934Q35.821 124.911 35.771 124.696L35.771 124.606Q35.821 124.399 36.017 124.375L36.107 124.375L36.107 122.629Q36.107 122.004 35.857 122.004Q35.525 122.004 35.347 122.315Q35.169 122.625 35.169 122.989L35.169 124.375L35.372 124.375Q35.579 124.399 35.618 124.606L35.618 124.696Q35.568 124.911 35.372 124.934L34.556 124.934Q34.357 124.911 34.306 124.696L34.306 124.606Q34.357 124.399 34.556 124.375L34.642 124.375L34.642 122.629Q34.642 122.004 34.396 122.004Q34.064 122.004 33.886 122.317Q33.708 122.629 33.708 122.989L33.708 124.375L33.907 124.375Q34.114 124.399 34.153 124.606L34.153 124.696Q34.103 124.914 33.907 124.934L32.978 124.934Q32.771 124.911 32.732 124.696\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M-65.203 21.082H54.298V-4.526H-65.203Z\"\u002F>\u003Cg transform=\"translate(-25.5 -114.212)\">\u003Cpath d=\"M-5.133 124.696L-5.133 124.606Q-5.082 124.399-4.887 124.375L-4.004 124.375L-4.004 122.047L-4.859 122.047Q-5.058 122.024-5.109 121.805L-5.109 121.719Q-5.058 121.508-4.859 121.485L-4.004 121.485L-4.004 121.032Q-4.004 120.567-3.598 120.286Q-3.191 120.004-2.711 120.004Q-2.398 120.004-2.154 120.125Q-1.910 120.246-1.910 120.528Q-1.910 120.692-2.019 120.809Q-2.129 120.926-2.293 120.926Q-2.441 120.926-2.562 120.821Q-2.683 120.715-2.683 120.567L-2.766 120.567Q-2.918 120.567-3.055 120.627Q-3.191 120.688-3.277 120.803Q-3.363 120.918-3.363 121.063L-3.363 121.485L-2.332 121.485Q-2.137 121.504-2.086 121.719L-2.086 121.805Q-2.137 122.024-2.332 122.047L-3.363 122.047L-3.363 124.375L-2.484 124.375Q-2.289 124.399-2.238 124.606L-2.238 124.696Q-2.289 124.911-2.484 124.934L-4.887 124.934Q-5.082 124.911-5.133 124.696M2.305 123.446L-0.137 123.446Q-0.082 123.723 0.115 123.946Q0.313 124.168 0.590 124.291Q0.867 124.414 1.152 124.414Q1.625 124.414 1.848 124.125Q1.856 124.114 1.912 124.008Q1.969 123.903 2.018 123.860Q2.067 123.817 2.160 123.805L2.305 123.805Q2.496 123.825 2.555 124.039L2.555 124.094Q2.488 124.395 2.258 124.592Q2.027 124.789 1.715 124.881Q1.402 124.973 1.098 124.973Q0.613 124.973 0.174 124.745Q-0.266 124.516-0.533 124.116Q-0.801 123.715-0.801 123.223L-0.801 123.164Q-0.801 122.696-0.555 122.293Q-0.308 121.891 0.100 121.657Q0.508 121.422 0.977 121.422Q1.481 121.422 1.834 121.645Q2.188 121.868 2.371 122.256Q2.555 122.645 2.555 123.149L2.555 123.207Q2.496 123.422 2.305 123.446M-0.129 122.895L1.899 122.895Q1.852 122.485 1.613 122.233Q1.375 121.981 0.977 121.981Q0.582 121.981 0.276 122.243Q-0.031 122.504-0.129 122.895M6.551 123.446L4.109 123.446Q4.164 123.723 4.361 123.946Q4.559 124.168 4.836 124.291Q5.113 124.414 5.399 124.414Q5.871 124.414 6.094 124.125Q6.102 124.114 6.158 124.008Q6.215 123.903 6.264 123.860Q6.313 123.817 6.406 123.805L6.551 123.805Q6.742 123.825 6.801 124.039L6.801 124.094Q6.734 124.395 6.504 124.592Q6.274 124.789 5.961 124.881Q5.649 124.973 5.344 124.973Q4.859 124.973 4.420 124.745Q3.981 124.516 3.713 124.116Q3.445 123.715 3.445 123.223L3.445 123.164Q3.445 122.696 3.692 122.293Q3.938 121.891 4.346 121.657Q4.754 121.422 5.223 121.422Q5.727 121.422 6.080 121.645Q6.434 121.868 6.617 122.256Q6.801 122.645 6.801 123.149L6.801 123.207Q6.742 123.422 6.551 123.446M4.117 122.895L6.145 122.895Q6.098 122.485 5.859 122.233Q5.621 121.981 5.223 121.981Q4.828 121.981 4.522 122.243Q4.215 122.504 4.117 122.895M9.149 124.973Q8.684 124.973 8.318 124.723Q7.953 124.473 7.748 124.069Q7.543 123.664 7.543 123.207Q7.543 122.864 7.668 122.543Q7.793 122.223 8.026 121.973Q8.258 121.723 8.563 121.584Q8.867 121.446 9.223 121.446Q9.735 121.446 10.141 121.766L10.141 120.606L9.719 120.606Q9.508 120.582 9.469 120.368L9.469 120.278Q9.508 120.071 9.719 120.047L10.531 120.047Q10.731 120.071 10.781 120.278L10.781 124.375L11.207 124.375Q11.414 124.399 11.453 124.606L11.453 124.696Q11.414 124.911 11.207 124.934L10.391 124.934Q10.192 124.911 10.141 124.696L10.141 124.567Q9.945 124.758 9.684 124.866Q9.422 124.973 9.149 124.973M9.188 124.414Q9.547 124.414 9.799 124.145Q10.051 123.875 10.141 123.500L10.141 122.668Q10.082 122.481 9.955 122.330Q9.828 122.180 9.651 122.092Q9.473 122.004 9.277 122.004Q8.969 122.004 8.719 122.174Q8.469 122.344 8.324 122.629Q8.180 122.914 8.180 123.215Q8.180 123.664 8.465 124.039Q8.750 124.414 9.188 124.414M15.027 122.797L12.285 122.797Q12.160 122.786 12.074 122.700Q11.988 122.614 11.988 122.485Q11.988 122.356 12.074 122.274Q12.160 122.192 12.285 122.180L15.027 122.180Q15.152 122.192 15.235 122.274Q15.317 122.356 15.317 122.485Q15.317 122.614 15.235 122.700Q15.152 122.786 15.027 122.797M16.098 124.696L16.098 124.606Q16.149 124.399 16.344 124.375L17.227 124.375L17.227 122.047L16.371 122.047Q16.172 122.024 16.121 121.805L16.121 121.719Q16.172 121.508 16.371 121.485L17.227 121.485L17.227 121.032Q17.227 120.567 17.633 120.286Q18.039 120.004 18.520 120.004Q18.832 120.004 19.076 120.125Q19.320 120.246 19.320 120.528Q19.320 120.692 19.211 120.809Q19.102 120.926 18.938 120.926Q18.789 120.926 18.668 120.821Q18.547 120.715 18.547 120.567L18.465 120.567Q18.313 120.567 18.176 120.627Q18.039 120.688 17.953 120.803Q17.867 120.918 17.867 121.063L17.867 121.485L18.899 121.485Q19.094 121.504 19.145 121.719L19.145 121.805Q19.094 122.024 18.899 122.047L17.867 122.047L17.867 124.375L18.746 124.375Q18.942 124.399 18.992 124.606L18.992 124.696Q18.942 124.911 18.746 124.934L16.344 124.934Q16.149 124.911 16.098 124.696M22.145 124.973Q21.672 124.973 21.287 124.729Q20.902 124.485 20.680 124.075Q20.457 123.664 20.457 123.207Q20.457 122.864 20.582 122.541Q20.707 122.219 20.938 121.965Q21.168 121.711 21.475 121.567Q21.781 121.422 22.145 121.422Q22.508 121.422 22.820 121.569Q23.133 121.715 23.356 121.961Q23.578 122.207 23.705 122.528Q23.832 122.848 23.832 123.207Q23.832 123.664 23.608 124.077Q23.383 124.489 22.998 124.731Q22.613 124.973 22.145 124.973M22.145 124.414Q22.610 124.414 22.901 124.020Q23.192 123.625 23.192 123.141Q23.192 122.848 23.057 122.580Q22.922 122.313 22.682 122.147Q22.442 121.981 22.145 121.981Q21.840 121.981 21.602 122.147Q21.363 122.313 21.229 122.580Q21.094 122.848 21.094 123.141Q21.094 123.621 21.387 124.018Q21.680 124.414 22.145 124.414M24.492 124.696L24.492 124.606Q24.551 124.399 24.742 124.375L25.453 124.375L25.453 122.047L24.742 122.047Q24.547 122.024 24.492 121.805L24.492 121.719Q24.551 121.508 24.742 121.485L25.844 121.485Q26.043 121.504 26.094 121.719L26.094 122.047Q26.356 121.762 26.711 121.604Q27.067 121.446 27.453 121.446Q27.746 121.446 27.981 121.580Q28.215 121.715 28.215 121.981Q28.215 122.149 28.106 122.266Q27.996 122.383 27.828 122.383Q27.676 122.383 27.561 122.272Q27.445 122.161 27.445 122.004Q27.070 122.004 26.756 122.205Q26.442 122.407 26.268 122.741Q26.094 123.075 26.094 123.454L26.094 124.375L27.039 124.375Q27.246 124.399 27.285 124.606L27.285 124.696Q27.246 124.911 27.039 124.934L24.742 124.934Q24.551 124.911 24.492 124.696M29.524 124.711L29.059 122.047L28.899 122.047Q28.692 122.024 28.652 121.805L28.652 121.719Q28.692 121.508 28.899 121.485L30.067 121.485Q30.277 121.508 30.317 121.719L30.317 121.805Q30.277 122.024 30.067 122.047L29.594 122.047Q29.707 122.692 29.785 123.137Q29.863 123.582 29.914 123.901Q29.965 124.219 29.965 124.293Q29.973 124.121 30.258 123.125Q30.297 123.012 30.395 122.938Q30.492 122.864 30.613 122.864L30.692 122.864Q30.813 122.864 30.910 122.938Q31.008 123.012 31.043 123.125Q31.152 123.508 31.235 123.832Q31.317 124.157 31.317 124.293Q31.328 124.004 31.676 122.047L31.203 122.047Q30.996 122.024 30.957 121.805L30.957 121.719Q30.996 121.508 31.203 121.485L32.379 121.485Q32.570 121.508 32.629 121.719L32.629 121.805Q32.574 122.024 32.379 122.047L32.211 122.047L31.746 124.711Q31.723 124.829 31.635 124.901Q31.547 124.973 31.426 124.973L31.285 124.973Q31.164 124.973 31.068 124.901Q30.973 124.829 30.930 124.711Q30.883 124.532 30.811 124.276Q30.738 124.020 30.695 123.811Q30.652 123.602 30.652 123.500Q30.645 123.782 30.363 124.711Q30.336 124.821 30.238 124.897Q30.141 124.973 30.020 124.973L29.844 124.973Q29.723 124.973 29.635 124.901Q29.547 124.829 29.524 124.711M33.168 123.821Q33.168 123.375 33.582 123.118Q33.996 122.860 34.537 122.760Q35.078 122.661 35.586 122.653Q35.586 122.438 35.451 122.286Q35.317 122.133 35.109 122.057Q34.902 121.981 34.692 121.981Q34.348 121.981 34.188 122.004L34.188 122.063Q34.188 122.231 34.068 122.346Q33.949 122.461 33.785 122.461Q33.609 122.461 33.494 122.338Q33.379 122.215 33.379 122.047Q33.379 121.641 33.760 121.532Q34.141 121.422 34.699 121.422Q34.969 121.422 35.236 121.500Q35.504 121.579 35.729 121.729Q35.953 121.879 36.090 122.100Q36.227 122.321 36.227 122.598L36.227 124.317Q36.227 124.375 36.754 124.375Q36.949 124.395 37 124.606L37 124.696Q36.949 124.911 36.754 124.934L36.609 124.934Q36.266 124.934 36.037 124.887Q35.809 124.840 35.664 124.653Q35.203 124.973 34.496 124.973Q34.160 124.973 33.856 124.832Q33.551 124.692 33.359 124.430Q33.168 124.168 33.168 123.821M33.809 123.829Q33.809 124.102 34.051 124.258Q34.293 124.414 34.578 124.414Q34.797 124.414 35.029 124.356Q35.262 124.297 35.424 124.159Q35.586 124.020 35.586 123.797L35.586 123.207Q35.305 123.207 34.889 123.264Q34.473 123.321 34.141 123.459Q33.809 123.598 33.809 123.829M37.231 124.696L37.231 124.606Q37.289 124.399 37.481 124.375L38.192 124.375L38.192 122.047L37.481 122.047Q37.285 122.024 37.231 121.805L37.231 121.719Q37.289 121.508 37.481 121.485L38.582 121.485Q38.781 121.504 38.832 121.719L38.832 122.047Q39.094 121.762 39.449 121.604Q39.805 121.446 40.192 121.446Q40.484 121.446 40.719 121.580Q40.953 121.715 40.953 121.981Q40.953 122.149 40.844 122.266Q40.734 122.383 40.567 122.383Q40.414 122.383 40.299 122.272Q40.184 122.161 40.184 122.004Q39.809 122.004 39.494 122.205Q39.180 122.407 39.006 122.741Q38.832 123.075 38.832 123.454L38.832 124.375L39.777 124.375Q39.984 124.399 40.024 124.606L40.024 124.696Q39.984 124.911 39.777 124.934L37.481 124.934Q37.289 124.911 37.231 124.696M43.117 124.973Q42.652 124.973 42.287 124.723Q41.922 124.473 41.717 124.069Q41.512 123.664 41.512 123.207Q41.512 122.864 41.637 122.543Q41.762 122.223 41.994 121.973Q42.227 121.723 42.531 121.584Q42.836 121.446 43.192 121.446Q43.703 121.446 44.109 121.766L44.109 120.606L43.688 120.606Q43.477 120.582 43.438 120.368L43.438 120.278Q43.477 120.071 43.688 120.047L44.500 120.047Q44.699 120.071 44.750 120.278L44.750 124.375L45.176 124.375Q45.383 124.399 45.422 124.606L45.422 124.696Q45.383 124.911 45.176 124.934L44.359 124.934Q44.160 124.911 44.109 124.696L44.109 124.567Q43.914 124.758 43.652 124.866Q43.391 124.973 43.117 124.973M43.156 124.414Q43.516 124.414 43.768 124.145Q44.020 123.875 44.109 123.500L44.109 122.668Q44.051 122.481 43.924 122.330Q43.797 122.180 43.619 122.092Q43.442 122.004 43.246 122.004Q42.938 122.004 42.688 122.174Q42.438 122.344 42.293 122.629Q42.149 122.914 42.149 123.215Q42.149 123.664 42.434 124.039Q42.719 124.414 43.156 124.414\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-soft-neutral)\">\u003Cpath d=\"M-65.203-20.175H54.298v-22.762H-65.203Z\"\u002F>\u003Cg fill=\"currentColor\" stroke=\"none\" font-family=\"cmtt8\" font-size=\"8\">\u003Cg transform=\"translate(-21.25 -154.046)\">\u003Cpath d=\"M-5.047 123.821Q-5.047 123.375-4.633 123.118Q-4.219 122.860-3.678 122.760Q-3.137 122.661-2.629 122.653Q-2.629 122.438-2.764 122.286Q-2.898 122.133-3.105 122.057Q-3.312 121.981-3.523 121.981Q-3.867 121.981-4.027 122.004L-4.027 122.063Q-4.027 122.231-4.146 122.346Q-4.266 122.461-4.430 122.461Q-4.605 122.461-4.721 122.338Q-4.836 122.215-4.836 122.047Q-4.836 121.641-4.455 121.532Q-4.074 121.422-3.516 121.422Q-3.246 121.422-2.978 121.500Q-2.711 121.579-2.486 121.729Q-2.262 121.879-2.125 122.100Q-1.988 122.321-1.988 122.598L-1.988 124.317Q-1.988 124.375-1.461 124.375Q-1.266 124.395-1.215 124.606L-1.215 124.696Q-1.266 124.911-1.461 124.934L-1.605 124.934Q-1.949 124.934-2.178 124.887Q-2.406 124.840-2.551 124.653Q-3.012 124.973-3.719 124.973Q-4.055 124.973-4.359 124.832Q-4.664 124.692-4.855 124.430Q-5.047 124.168-5.047 123.821M-4.406 123.829Q-4.406 124.102-4.164 124.258Q-3.922 124.414-3.637 124.414Q-3.418 124.414-3.185 124.356Q-2.953 124.297-2.791 124.159Q-2.629 124.020-2.629 123.797L-2.629 123.207Q-2.910 123.207-3.326 123.264Q-3.742 123.321-4.074 123.459Q-4.406 123.598-4.406 123.829M0.656 124.973Q0.192 124.973-0.174 124.723Q-0.539 124.473-0.744 124.069Q-0.949 123.664-0.949 123.207Q-0.949 122.864-0.824 122.543Q-0.699 122.223-0.467 121.973Q-0.234 121.723 0.070 121.584Q0.375 121.446 0.731 121.446Q1.242 121.446 1.649 121.766L1.649 120.606L1.227 120.606Q1.016 120.582 0.977 120.368L0.977 120.278Q1.016 120.071 1.227 120.047L2.039 120.047Q2.238 120.071 2.289 120.278L2.289 124.375L2.715 124.375Q2.922 124.399 2.961 124.606L2.961 124.696Q2.922 124.911 2.715 124.934L1.899 124.934Q1.699 124.911 1.649 124.696L1.649 124.567Q1.453 124.758 1.192 124.866Q0.930 124.973 0.656 124.973M0.695 124.414Q1.055 124.414 1.307 124.145Q1.559 123.875 1.649 123.500L1.649 122.668Q1.590 122.481 1.463 122.330Q1.336 122.180 1.158 122.092Q0.981 122.004 0.785 122.004Q0.477 122.004 0.227 122.174Q-0.023 122.344-0.168 122.629Q-0.312 122.914-0.312 123.215Q-0.312 123.664-0.027 124.039Q0.258 124.414 0.695 124.414M4.902 124.973Q4.438 124.973 4.072 124.723Q3.707 124.473 3.502 124.069Q3.297 123.664 3.297 123.207Q3.297 122.864 3.422 122.543Q3.547 122.223 3.779 121.973Q4.012 121.723 4.317 121.584Q4.621 121.446 4.977 121.446Q5.488 121.446 5.895 121.766L5.895 120.606L5.473 120.606Q5.262 120.582 5.223 120.368L5.223 120.278Q5.262 120.071 5.473 120.047L6.285 120.047Q6.484 120.071 6.535 120.278L6.535 124.375L6.961 124.375Q7.168 124.399 7.207 124.606L7.207 124.696Q7.168 124.911 6.961 124.934L6.145 124.934Q5.945 124.911 5.895 124.696L5.895 124.567Q5.699 124.758 5.438 124.866Q5.176 124.973 4.902 124.973M4.942 124.414Q5.301 124.414 5.553 124.145Q5.805 123.875 5.895 123.500L5.895 122.668Q5.836 122.481 5.709 122.330Q5.582 122.180 5.404 122.092Q5.227 122.004 5.031 122.004Q4.723 122.004 4.473 122.174Q4.223 122.344 4.078 122.629Q3.934 122.914 3.934 123.215Q3.934 123.664 4.219 124.039Q4.504 124.414 4.942 124.414\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-21.25 -154.046)\">\u003Cpath d=\"M12.763 125.012Q12.466 125.012 12.241 124.840Q12.017 124.668 11.894 124.393Q11.771 124.118 11.771 123.836L11.771 123.782Q11.771 123.524 11.862 123.288Q11.954 123.051 12.114 122.895L12.571 122.438Q12.349 121.801 12.349 121.149L12.349 121.094Q12.349 120.829 12.456 120.565Q12.564 120.301 12.769 120.133Q12.974 119.965 13.251 119.965Q13.528 119.965 13.708 120.137Q13.888 120.309 13.966 120.571Q14.044 120.832 14.044 121.094L14.044 121.149Q14.044 121.418 13.917 121.674Q13.790 121.930 13.624 122.129Q13.458 122.329 13.196 122.590Q13.376 123 13.739 123.567Q13.954 123.164 14.267 122.325L14.380 122.032Q14.239 121.996 14.196 121.805L14.196 121.719Q14.247 121.508 14.443 121.485L15.267 121.485Q15.466 121.504 15.517 121.719L15.517 121.805Q15.466 122.024 15.267 122.047L14.939 122.047L14.747 122.551Q14.364 123.575 14.083 124.047Q14.255 124.235 14.400 124.344Q14.544 124.454 14.724 124.454Q14.845 124.454 14.900 124.317Q14.954 124.180 14.954 124.032Q15.017 123.797 15.220 123.797Q15.423 123.797 15.486 124.032L15.486 124.086Q15.486 124.450 15.277 124.731Q15.068 125.012 14.724 125.012Q14.247 125.012 13.771 124.493Q13.314 125.012 12.763 125.012M12.763 124.454Q13.114 124.454 13.427 124.063Q13.013 123.422 12.806 122.996Q12.693 123.102 12.622 123.182Q12.552 123.262 12.503 123.344Q12.454 123.426 12.433 123.538Q12.411 123.649 12.411 123.782Q12.411 124 12.495 124.227Q12.579 124.454 12.763 124.454M12.989 122.004Q13.216 121.758 13.366 121.543Q13.517 121.329 13.517 121.118Q13.517 120.930 13.456 120.729Q13.396 120.528 13.251 120.528Q13.068 120.528 12.972 120.725Q12.876 120.922 12.876 121.133Q12.876 121.571 12.989 122.004\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-21.25 -154.046)\">\u003Cpath d=\"M20.118 124.696L20.118 124.606Q20.161 124.399 20.368 124.375L20.790 124.375L20.790 122.047L20.368 122.047Q20.161 122.024 20.118 121.805L20.118 121.719Q20.165 121.508 20.368 121.485L21.185 121.485Q21.380 121.508 21.431 121.719L21.431 121.805L21.423 121.829Q21.650 121.649 21.923 121.547Q22.196 121.446 22.489 121.446Q22.837 121.446 23.075 121.586Q23.314 121.727 23.429 121.985Q23.544 122.243 23.544 122.598L23.544 124.375L23.970 124.375Q24.177 124.399 24.216 124.606L24.216 124.696Q24.177 124.911 23.970 124.934L22.575 124.934Q22.380 124.911 22.329 124.696L22.329 124.606Q22.380 124.395 22.575 124.375L22.903 124.375L22.903 122.629Q22.903 122.321 22.814 122.163Q22.724 122.004 22.431 122.004Q22.161 122.004 21.933 122.135Q21.704 122.266 21.568 122.495Q21.431 122.723 21.431 122.989L21.431 124.375L21.857 124.375Q22.064 124.399 22.103 124.606L22.103 124.696Q22.064 124.911 21.857 124.934L20.368 124.934Q20.161 124.911 20.118 124.696M26.415 124.973Q25.943 124.973 25.558 124.729Q25.173 124.485 24.950 124.075Q24.728 123.664 24.728 123.207Q24.728 122.864 24.853 122.541Q24.978 122.219 25.208 121.965Q25.439 121.711 25.745 121.567Q26.052 121.422 26.415 121.422Q26.778 121.422 27.091 121.569Q27.403 121.715 27.626 121.961Q27.849 122.207 27.976 122.528Q28.103 122.848 28.103 123.207Q28.103 123.664 27.878 124.077Q27.653 124.489 27.269 124.731Q26.884 124.973 26.415 124.973M26.415 124.414Q26.880 124.414 27.171 124.020Q27.462 123.625 27.462 123.141Q27.462 122.848 27.327 122.580Q27.193 122.313 26.952 122.147Q26.712 121.981 26.415 121.981Q26.110 121.981 25.872 122.147Q25.634 122.313 25.499 122.580Q25.364 122.848 25.364 123.141Q25.364 123.621 25.657 124.018Q25.950 124.414 26.415 124.414M28.763 124.696L28.763 124.606Q28.821 124.399 29.013 124.375L29.724 124.375L29.724 122.047L29.013 122.047Q28.818 122.024 28.763 121.805L28.763 121.719Q28.821 121.508 29.013 121.485L30.114 121.485Q30.314 121.504 30.364 121.719L30.364 122.047Q30.626 121.762 30.982 121.604Q31.337 121.446 31.724 121.446Q32.017 121.446 32.251 121.580Q32.486 121.715 32.486 121.981Q32.486 122.149 32.376 122.266Q32.267 122.383 32.099 122.383Q31.946 122.383 31.831 122.272Q31.716 122.161 31.716 122.004Q31.341 122.004 31.027 122.205Q30.712 122.407 30.538 122.741Q30.364 123.075 30.364 123.454L30.364 124.375L31.310 124.375Q31.517 124.399 31.556 124.606L31.556 124.696Q31.517 124.911 31.310 124.934L29.013 124.934Q28.821 124.911 28.763 124.696M32.732 124.696L32.732 124.606Q32.782 124.395 32.978 124.375L33.177 124.375L33.177 122.047L32.978 122.047Q32.771 122.024 32.732 121.805L32.732 121.719Q32.782 121.504 32.978 121.485L33.458 121.485Q33.650 121.508 33.708 121.719Q34.028 121.446 34.443 121.446Q34.634 121.446 34.802 121.555Q34.970 121.664 35.044 121.836Q35.220 121.645 35.443 121.545Q35.665 121.446 35.907 121.446Q36.325 121.446 36.480 121.788Q36.634 122.129 36.634 122.598L36.634 124.375L36.833 124.375Q37.044 124.399 37.083 124.606L37.083 124.696Q37.032 124.911 36.833 124.934L36.017 124.934Q35.821 124.911 35.771 124.696L35.771 124.606Q35.821 124.399 36.017 124.375L36.107 124.375L36.107 122.629Q36.107 122.004 35.857 122.004Q35.525 122.004 35.347 122.315Q35.169 122.625 35.169 122.989L35.169 124.375L35.372 124.375Q35.579 124.399 35.618 124.606L35.618 124.696Q35.568 124.911 35.372 124.934L34.556 124.934Q34.357 124.911 34.306 124.696L34.306 124.606Q34.357 124.399 34.556 124.375L34.642 124.375L34.642 122.629Q34.642 122.004 34.396 122.004Q34.064 122.004 33.886 122.317Q33.708 122.629 33.708 122.989L33.708 124.375L33.907 124.375Q34.114 124.399 34.153 124.606L34.153 124.696Q34.103 124.914 33.907 124.934L32.978 124.934Q32.771 124.911 32.732 124.696\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003Cg transform=\"translate(-12.514 -188.95)\">\u003Cpath d=\"M-5.215 123.239Q-5.215 122.735-4.959 122.303Q-4.703 121.871-4.267 121.620Q-3.832 121.368-3.332 121.368Q-2.945 121.368-2.603 121.512Q-2.262 121.657-2 121.918Q-1.738 122.180-1.596 122.516Q-1.453 122.852-1.453 123.239Q-1.453 123.731-1.717 124.141Q-1.980 124.551-2.410 124.782Q-2.840 125.012-3.332 125.012Q-3.824 125.012-4.258 124.780Q-4.691 124.547-4.953 124.139Q-5.215 123.731-5.215 123.239M-3.332 124.735Q-2.875 124.735-2.623 124.512Q-2.371 124.289-2.283 123.938Q-2.195 123.586-2.195 123.141Q-2.195 122.711-2.289 122.373Q-2.383 122.036-2.637 121.829Q-2.891 121.621-3.332 121.621Q-3.980 121.621-4.224 122.038Q-4.469 122.454-4.469 123.141Q-4.469 123.586-4.381 123.938Q-4.293 124.289-4.041 124.512Q-3.789 124.735-3.332 124.735M-0.285 123.981L-0.285 122.239Q-0.285 122.024-0.348 121.928Q-0.410 121.832-0.529 121.811Q-0.648 121.789-0.894 121.789L-0.894 121.493L0.352 121.407L0.352 123.957L0.352 123.981Q0.352 124.293 0.406 124.455Q0.461 124.618 0.611 124.688Q0.762 124.758 1.082 124.758Q1.512 124.758 1.785 124.420Q2.059 124.082 2.059 123.637L2.059 122.239Q2.059 122.024 1.996 121.928Q1.934 121.832 1.815 121.811Q1.695 121.789 1.449 121.789L1.449 121.493L2.695 121.407L2.695 124.192Q2.695 124.403 2.758 124.498Q2.820 124.594 2.940 124.616Q3.059 124.637 3.305 124.637L3.305 124.934L2.082 125.012L2.082 124.391Q1.914 124.680 1.633 124.846Q1.352 125.012 1.031 125.012Q-0.285 125.012-0.285 123.981M4.375 123.973L4.375 121.782L3.672 121.782L3.672 121.528Q4.027 121.528 4.270 121.295Q4.512 121.063 4.623 120.715Q4.734 120.368 4.734 120.012L5.016 120.012L5.016 121.485L6.192 121.485L6.192 121.782L5.016 121.782L5.016 123.957Q5.016 124.278 5.135 124.506Q5.254 124.735 5.535 124.735Q5.715 124.735 5.832 124.612Q5.949 124.489 6.002 124.309Q6.055 124.129 6.055 123.957L6.055 123.485L6.336 123.485L6.336 123.973Q6.336 124.227 6.231 124.467Q6.125 124.707 5.928 124.860Q5.731 125.012 5.473 125.012Q5.156 125.012 4.904 124.889Q4.652 124.766 4.514 124.532Q4.375 124.297 4.375 123.973M8.938 126.485L7.082 126.485L7.082 126.192Q7.352 126.192 7.520 126.147Q7.688 126.102 7.688 125.926L7.688 122.102Q7.688 121.895 7.531 121.842Q7.375 121.789 7.082 121.789L7.082 121.493L8.305 121.407L8.305 121.871Q8.535 121.649 8.850 121.528Q9.164 121.407 9.504 121.407Q9.977 121.407 10.381 121.653Q10.785 121.899 11.018 122.315Q11.250 122.731 11.250 123.207Q11.250 123.582 11.102 123.911Q10.953 124.239 10.684 124.491Q10.414 124.743 10.070 124.877Q9.727 125.012 9.367 125.012Q9.078 125.012 8.807 124.891Q8.535 124.770 8.328 124.559L8.328 125.926Q8.328 126.102 8.496 126.147Q8.664 126.192 8.938 126.192L8.938 126.485M8.328 122.270L8.328 124.110Q8.481 124.399 8.742 124.579Q9.004 124.758 9.313 124.758Q9.598 124.758 9.820 124.620Q10.043 124.481 10.195 124.250Q10.348 124.020 10.426 123.748Q10.504 123.477 10.504 123.207Q10.504 122.875 10.379 122.518Q10.254 122.161 10.006 121.924Q9.758 121.688 9.410 121.688Q9.086 121.688 8.791 121.844Q8.496 122 8.328 122.270M12.457 123.981L12.457 122.239Q12.457 122.024 12.395 121.928Q12.332 121.832 12.213 121.811Q12.094 121.789 11.848 121.789L11.848 121.493L13.094 121.407L13.094 123.957L13.094 123.981Q13.094 124.293 13.149 124.455Q13.203 124.618 13.354 124.688Q13.504 124.758 13.824 124.758Q14.254 124.758 14.527 124.420Q14.801 124.082 14.801 123.637L14.801 122.239Q14.801 122.024 14.738 121.928Q14.676 121.832 14.557 121.811Q14.438 121.789 14.192 121.789L14.192 121.493L15.438 121.407L15.438 124.192Q15.438 124.403 15.500 124.498Q15.563 124.594 15.682 124.616Q15.801 124.637 16.047 124.637L16.047 124.934L14.824 125.012L14.824 124.391Q14.656 124.680 14.375 124.846Q14.094 125.012 13.774 125.012Q12.457 125.012 12.457 123.981M17.117 123.973L17.117 121.782L16.414 121.782L16.414 121.528Q16.770 121.528 17.012 121.295Q17.254 121.063 17.365 120.715Q17.477 120.368 17.477 120.012L17.758 120.012L17.758 121.485L18.934 121.485L18.934 121.782L17.758 121.782L17.758 123.957Q17.758 124.278 17.877 124.506Q17.996 124.735 18.277 124.735Q18.457 124.735 18.574 124.612Q18.692 124.489 18.744 124.309Q18.797 124.129 18.797 123.957L18.797 123.485L19.078 123.485L19.078 123.973Q19.078 124.227 18.973 124.467Q18.867 124.707 18.670 124.860Q18.473 125.012 18.215 125.012Q17.899 125.012 17.647 124.889Q17.395 124.766 17.256 124.532Q17.117 124.297 17.117 123.973\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M-5.453 117.941V103.75\"\u002F>\u003Cpath stroke=\"none\" d=\"m-5.453 101.15-2.08 4.16 2.08-1.56 2.08 1.56\"\u002F>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M-5.453 74.742V62.293\"\u002F>\u003Cpath stroke=\"none\" d=\"m-5.453 59.693-2.08 4.16 2.08-1.56 2.08 1.56\"\u002F>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M-5.453 36.531v-12.45\"\u002F>\u003Cpath stroke=\"none\" d=\"m-5.453 21.482-2.08 4.16 2.08-1.56 2.08 1.56\"\u002F>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M-5.453-4.926v-12.449\"\u002F>\u003Cpath stroke=\"none\" d=\"m-5.453-19.975-2.08 4.16 2.08-1.56 2.08 1.56\"\u002F>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M-5.453-43.137v-13.19\"\u002F>\u003Cpath stroke=\"none\" d=\"m-5.453-58.928-2.08 4.16 2.08-1.56 2.08 1.56\"\u002F>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M54.298 119.244c36.989-19.917 36.989-51.215 2.29-69.9\"\u002F>\u003Cpath stroke=\"none\" d=\"m54.298 48.112 2.677 3.803-.388-2.57 2.36-1.092\"\u002F>\u003Cg transform=\"translate(91.225 -39.506)\">\u003Cpath d=\"M-5.139 124.927L-5.139 123.864Q-5.139 123.840-5.111 123.813Q-5.084 123.786-5.060 123.786L-4.951 123.786Q-4.886 123.786-4.872 123.844Q-4.776 124.278-4.530 124.529Q-4.284 124.780-3.870 124.780Q-3.529 124.780-3.276 124.647Q-3.023 124.514-3.023 124.206Q-3.023 124.049-3.117 123.934Q-3.211 123.820-3.349 123.751Q-3.488 123.683-3.655 123.645L-4.236 123.546Q-4.592 123.478-4.865 123.257Q-5.139 123.037-5.139 122.695Q-5.139 122.446-5.027 122.271Q-4.916 122.097-4.730 121.998Q-4.544 121.899-4.328 121.856Q-4.113 121.813-3.870 121.813Q-3.457 121.813-3.177 121.995L-2.961 121.820Q-2.951 121.817-2.944 121.815Q-2.937 121.813-2.927 121.813L-2.876 121.813Q-2.849 121.813-2.825 121.837Q-2.801 121.861-2.801 121.889L-2.801 122.736Q-2.801 122.757-2.825 122.784Q-2.849 122.811-2.876 122.811L-2.989 122.811Q-3.016 122.811-3.042 122.786Q-3.067 122.760-3.067 122.736Q-3.067 122.500-3.173 122.336Q-3.279 122.172-3.462 122.090Q-3.645 122.008-3.877 122.008Q-4.205 122.008-4.462 122.111Q-4.718 122.213-4.718 122.490Q-4.718 122.685-4.535 122.794Q-4.352 122.904-4.123 122.945L-3.549 123.051Q-3.303 123.099-3.089 123.227Q-2.876 123.355-2.739 123.558Q-2.602 123.762-2.602 124.011Q-2.602 124.524-2.968 124.763Q-3.334 125.002-3.870 125.002Q-4.366 125.002-4.698 124.708L-4.964 124.982Q-4.985 125.002-5.012 125.002L-5.060 125.002Q-5.084 125.002-5.111 124.975Q-5.139 124.948-5.139 124.927M-0.377 124.934L-1.960 124.934L-1.960 124.654Q-1.731 124.654-1.582 124.620Q-1.433 124.585-1.433 124.445L-1.433 120.826Q-1.433 120.556-1.541 120.494Q-1.649 120.433-1.960 120.433L-1.960 120.152L-0.880 120.077L-0.880 123.365L0.105 122.596Q0.310 122.459 0.310 122.309Q0.310 122.265 0.269 122.230Q0.228 122.196 0.183 122.196L0.183 121.916L1.547 121.916L1.547 122.196Q1.058 122.196 0.539 122.596L-0.018 123.030L0.959 124.254Q1.161 124.500 1.294 124.577Q1.427 124.654 1.714 124.654L1.714 124.934L0.282 124.934L0.282 124.654Q0.470 124.654 0.470 124.541Q0.470 124.445 0.317 124.254L-0.418 123.345L-0.900 123.724L-0.900 124.445Q-0.900 124.582-0.752 124.618Q-0.603 124.654-0.377 124.654L-0.377 124.934M3.844 124.934L2.292 124.934L2.292 124.654Q2.518 124.654 2.666 124.620Q2.815 124.585 2.815 124.445L2.815 122.596Q2.815 122.408 2.767 122.324Q2.719 122.241 2.622 122.222Q2.525 122.203 2.313 122.203L2.313 121.923L3.369 121.848L3.369 124.445Q3.369 124.585 3.500 124.620Q3.632 124.654 3.844 124.654L3.844 124.934M2.572 120.627Q2.572 120.456 2.695 120.337Q2.818 120.217 2.989 120.217Q3.157 120.217 3.280 120.337Q3.403 120.456 3.403 120.627Q3.403 120.802 3.280 120.925Q3.157 121.048 2.989 121.048Q2.818 121.048 2.695 120.925Q2.572 120.802 2.572 120.627M6.134 126.291L4.504 126.291L4.504 126.011Q4.733 126.011 4.881 125.976Q5.030 125.942 5.030 125.802L5.030 122.456Q5.030 122.285 4.893 122.244Q4.756 122.203 4.504 122.203L4.504 121.923L5.584 121.848L5.584 122.254Q5.806 122.053 6.093 121.950Q6.380 121.848 6.688 121.848Q7.115 121.848 7.479 122.061Q7.843 122.275 8.057 122.639Q8.270 123.003 8.270 123.423Q8.270 123.868 8.031 124.232Q7.792 124.596 7.399 124.799Q7.005 125.002 6.561 125.002Q6.295 125.002 6.047 124.902Q5.799 124.801 5.611 124.620L5.611 125.802Q5.611 125.939 5.760 125.975Q5.908 126.011 6.134 126.011L6.134 126.291M5.611 122.603L5.611 124.213Q5.744 124.466 5.987 124.623Q6.230 124.780 6.506 124.780Q6.835 124.780 7.088 124.579Q7.340 124.377 7.474 124.059Q7.607 123.741 7.607 123.423Q7.607 123.194 7.542 122.965Q7.477 122.736 7.349 122.538Q7.221 122.340 7.026 122.220Q6.831 122.101 6.599 122.101Q6.305 122.101 6.037 122.230Q5.768 122.360 5.611 122.603\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M54.298 39.576c36.989-19.917 36.989-51.215 2.29-69.9\"\u002F>\u003Cpath stroke=\"none\" d=\"m54.298-31.556 2.677 3.804-.388-2.571 2.36-1.092\"\u002F>\u003Cg transform=\"translate(91.225 -119.174)\">\u003Cpath d=\"M-5.139 124.927L-5.139 123.864Q-5.139 123.840-5.111 123.813Q-5.084 123.786-5.060 123.786L-4.951 123.786Q-4.886 123.786-4.872 123.844Q-4.776 124.278-4.530 124.529Q-4.284 124.780-3.870 124.780Q-3.529 124.780-3.276 124.647Q-3.023 124.514-3.023 124.206Q-3.023 124.049-3.117 123.934Q-3.211 123.820-3.349 123.751Q-3.488 123.683-3.655 123.645L-4.236 123.546Q-4.592 123.478-4.865 123.257Q-5.139 123.037-5.139 122.695Q-5.139 122.446-5.027 122.271Q-4.916 122.097-4.730 121.998Q-4.544 121.899-4.328 121.856Q-4.113 121.813-3.870 121.813Q-3.457 121.813-3.177 121.995L-2.961 121.820Q-2.951 121.817-2.944 121.815Q-2.937 121.813-2.927 121.813L-2.876 121.813Q-2.849 121.813-2.825 121.837Q-2.801 121.861-2.801 121.889L-2.801 122.736Q-2.801 122.757-2.825 122.784Q-2.849 122.811-2.876 122.811L-2.989 122.811Q-3.016 122.811-3.042 122.786Q-3.067 122.760-3.067 122.736Q-3.067 122.500-3.173 122.336Q-3.279 122.172-3.462 122.090Q-3.645 122.008-3.877 122.008Q-4.205 122.008-4.462 122.111Q-4.718 122.213-4.718 122.490Q-4.718 122.685-4.535 122.794Q-4.352 122.904-4.123 122.945L-3.549 123.051Q-3.303 123.099-3.089 123.227Q-2.876 123.355-2.739 123.558Q-2.602 123.762-2.602 124.011Q-2.602 124.524-2.968 124.763Q-3.334 125.002-3.870 125.002Q-4.366 125.002-4.698 124.708L-4.964 124.982Q-4.985 125.002-5.012 125.002L-5.060 125.002Q-5.084 125.002-5.111 124.975Q-5.139 124.948-5.139 124.927M-0.377 124.934L-1.960 124.934L-1.960 124.654Q-1.731 124.654-1.582 124.620Q-1.433 124.585-1.433 124.445L-1.433 120.826Q-1.433 120.556-1.541 120.494Q-1.649 120.433-1.960 120.433L-1.960 120.152L-0.880 120.077L-0.880 123.365L0.105 122.596Q0.310 122.459 0.310 122.309Q0.310 122.265 0.269 122.230Q0.228 122.196 0.183 122.196L0.183 121.916L1.547 121.916L1.547 122.196Q1.058 122.196 0.539 122.596L-0.018 123.030L0.959 124.254Q1.161 124.500 1.294 124.577Q1.427 124.654 1.714 124.654L1.714 124.934L0.282 124.934L0.282 124.654Q0.470 124.654 0.470 124.541Q0.470 124.445 0.317 124.254L-0.418 123.345L-0.900 123.724L-0.900 124.445Q-0.900 124.582-0.752 124.618Q-0.603 124.654-0.377 124.654L-0.377 124.934M3.844 124.934L2.292 124.934L2.292 124.654Q2.518 124.654 2.666 124.620Q2.815 124.585 2.815 124.445L2.815 122.596Q2.815 122.408 2.767 122.324Q2.719 122.241 2.622 122.222Q2.525 122.203 2.313 122.203L2.313 121.923L3.369 121.848L3.369 124.445Q3.369 124.585 3.500 124.620Q3.632 124.654 3.844 124.654L3.844 124.934M2.572 120.627Q2.572 120.456 2.695 120.337Q2.818 120.217 2.989 120.217Q3.157 120.217 3.280 120.337Q3.403 120.456 3.403 120.627Q3.403 120.802 3.280 120.925Q3.157 121.048 2.989 121.048Q2.818 121.048 2.695 120.925Q2.572 120.802 2.572 120.627M6.134 126.291L4.504 126.291L4.504 126.011Q4.733 126.011 4.881 125.976Q5.030 125.942 5.030 125.802L5.030 122.456Q5.030 122.285 4.893 122.244Q4.756 122.203 4.504 122.203L4.504 121.923L5.584 121.848L5.584 122.254Q5.806 122.053 6.093 121.950Q6.380 121.848 6.688 121.848Q7.115 121.848 7.479 122.061Q7.843 122.275 8.057 122.639Q8.270 123.003 8.270 123.423Q8.270 123.868 8.031 124.232Q7.792 124.596 7.399 124.799Q7.005 125.002 6.561 125.002Q6.295 125.002 6.047 124.902Q5.799 124.801 5.611 124.620L5.611 125.802Q5.611 125.939 5.760 125.975Q5.908 126.011 6.134 126.011L6.134 126.291M5.611 122.603L5.611 124.213Q5.744 124.466 5.987 124.623Q6.230 124.780 6.506 124.780Q6.835 124.780 7.088 124.579Q7.340 124.377 7.474 124.059Q7.607 123.741 7.607 123.423Q7.607 123.194 7.542 122.965Q7.477 122.736 7.349 122.538Q7.221 122.340 7.026 122.220Q6.831 122.101 6.599 122.101Q6.305 122.101 6.037 122.230Q5.768 122.360 5.611 122.603\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fsvg>\u003Cfigcaption class=\"tikz-cap\">A Transformer encoder block: multi-head self-attention then a position-wise FFN, each wrapped in an add-and-normalize. The two curved skips are the residual connections.\u003C\u002Ffigcaption>","\u003Csvg style=\"width:100%;max-width:204.379px;height:auto\" xmlns=\"http:\u002F\u002Fwww.w3.org\u002F2000\u002Fsvg\" viewBox=\"-75 -75 153.284 145.621\">\u003Cg stroke=\"currentColor\" style=\"stroke-miterlimit:10;stroke-width:.4\">\u003Cg stroke=\"none\">\u003Cg transform=\"translate(8.727 -131.227)\">\u003Cpath d=\"M-52.696 66.990Q-52.696 66.949-52.689 66.925L-51.698 62.930Q-51.660 62.834-51.660 62.742Q-51.660 62.650-52.094 62.650Q-52.180 62.622-52.180 62.537L-52.152 62.427Q-52.145 62.386-52.074 62.369L-51.093 62.294Q-51.048 62.294-51.019 62.320Q-50.990 62.345-50.990 62.397L-51.711 65.275Q-51.499 65.186-51.351 65.061Q-51.202 64.936-50.864 64.632Q-50.525 64.328-50.284 64.196Q-50.043 64.065-49.784 64.065Q-49.565 64.065-49.416 64.210Q-49.267 64.355-49.267 64.574Q-49.267 64.769-49.384 64.914Q-49.500 65.059-49.688 65.059Q-49.804 65.059-49.886 64.986Q-49.968 64.912-49.968 64.793Q-49.968 64.639-49.855 64.509Q-49.743 64.379-49.589 64.379Q-49.671 64.287-49.801 64.287Q-50.060 64.287-50.310 64.456Q-50.559 64.625-50.911 64.946Q-51.264 65.268-51.438 65.374Q-50.337 65.493-50.337 66.122Q-50.337 66.221-50.373 66.385Q-50.409 66.549-50.409 66.638Q-50.409 66.997-50.170 66.997Q-49.927 66.997-49.780 66.725Q-49.633 66.454-49.555 66.122Q-49.531 66.061-49.476 66.061L-49.367 66.061Q-49.332 66.061-49.310 66.086Q-49.288 66.112-49.288 66.143Q-49.288 66.156-49.295 66.170Q-49.394 66.570-49.616 66.895Q-49.838 67.219-50.183 67.219Q-50.512 67.219-50.732 67.021Q-50.952 66.823-50.952 66.502Q-50.952 66.443-50.932 66.324Q-50.911 66.204-50.911 66.143Q-50.911 65.876-51.202 65.729Q-51.493 65.582-51.786 65.582L-52.132 66.977Q-52.162 67.079-52.256 67.149Q-52.350 67.219-52.453 67.219Q-52.552 67.219-52.624 67.156Q-52.696 67.093-52.696 66.990\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(8.727 -131.227)\">\u003Cpath d=\"M-46.190 68.151L-48.231 68.151L-48.231 67.912Q-47.450 67.912-47.450 67.792L-47.450 65.246Q-47.628 65.321-47.832 65.351Q-48.036 65.380-48.265 65.380L-48.265 65.141Q-47.929 65.141-47.645 65.072Q-47.362 65.004-47.152 64.821L-47.047 64.821Q-47.020 64.821-46.996 64.845Q-46.971 64.870-46.971 64.897L-46.971 67.792Q-46.971 67.912-46.190 67.912\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.150\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"none\">\u003Cg transform=\"translate(34.334 -131.227)\">\u003Cpath d=\"M-52.696 66.990Q-52.696 66.949-52.689 66.925L-51.698 62.930Q-51.660 62.834-51.660 62.742Q-51.660 62.650-52.094 62.650Q-52.180 62.622-52.180 62.537L-52.152 62.427Q-52.145 62.386-52.074 62.369L-51.093 62.294Q-51.048 62.294-51.019 62.320Q-50.990 62.345-50.990 62.397L-51.711 65.275Q-51.499 65.186-51.351 65.061Q-51.202 64.936-50.864 64.632Q-50.525 64.328-50.284 64.196Q-50.043 64.065-49.784 64.065Q-49.565 64.065-49.416 64.210Q-49.267 64.355-49.267 64.574Q-49.267 64.769-49.384 64.914Q-49.500 65.059-49.688 65.059Q-49.804 65.059-49.886 64.986Q-49.968 64.912-49.968 64.793Q-49.968 64.639-49.855 64.509Q-49.743 64.379-49.589 64.379Q-49.671 64.287-49.801 64.287Q-50.060 64.287-50.310 64.456Q-50.559 64.625-50.911 64.946Q-51.264 65.268-51.438 65.374Q-50.337 65.493-50.337 66.122Q-50.337 66.221-50.373 66.385Q-50.409 66.549-50.409 66.638Q-50.409 66.997-50.170 66.997Q-49.927 66.997-49.780 66.725Q-49.633 66.454-49.555 66.122Q-49.531 66.061-49.476 66.061L-49.367 66.061Q-49.332 66.061-49.310 66.086Q-49.288 66.112-49.288 66.143Q-49.288 66.156-49.295 66.170Q-49.394 66.570-49.616 66.895Q-49.838 67.219-50.183 67.219Q-50.512 67.219-50.732 67.021Q-50.952 66.823-50.952 66.502Q-50.952 66.443-50.932 66.324Q-50.911 66.204-50.911 66.143Q-50.911 65.876-51.202 65.729Q-51.493 65.582-51.786 65.582L-52.132 66.977Q-52.162 67.079-52.256 67.149Q-52.350 67.219-52.453 67.219Q-52.552 67.219-52.624 67.156Q-52.696 67.093-52.696 66.990\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(34.334 -131.227)\">\u003Cpath d=\"M-46.190 68.151L-48.517 68.151L-48.517 67.970Q-48.514 67.958-48.495 67.931L-47.455 67.055Q-47.147 66.796-46.993 66.652Q-46.840 66.508-46.710 66.291Q-46.581 66.073-46.581 65.832Q-46.581 65.590-46.708 65.414Q-46.835 65.238-47.039 65.149Q-47.242 65.060-47.482 65.060Q-47.689 65.060-47.885 65.148Q-48.080 65.236-48.185 65.402Q-48.065 65.402-47.988 65.494Q-47.911 65.585-47.911 65.700Q-47.911 65.827-47.998 65.916Q-48.085 66.005-48.212 66.005Q-48.341 66.005-48.429 65.915Q-48.517 65.824-48.517 65.700Q-48.517 65.419-48.344 65.220Q-48.170 65.021-47.899 64.921Q-47.628 64.821-47.355 64.821Q-47.030 64.821-46.725 64.928Q-46.420 65.036-46.223 65.263Q-46.027 65.490-46.027 65.827Q-46.027 66.064-46.140 66.256Q-46.254 66.449-46.414 66.587Q-46.574 66.725-46.856 66.913Q-47.138 67.101-47.216 67.160L-47.882 67.651L-47.426 67.651Q-46.993 67.651-46.699 67.644Q-46.405 67.638-46.390 67.626Q-46.312 67.528-46.256 67.167L-46.027 67.167\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.150\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"none\">\u003Cg transform=\"translate(59.941 -131.227)\">\u003Cpath d=\"M-52.696 66.990Q-52.696 66.949-52.689 66.925L-51.698 62.930Q-51.660 62.834-51.660 62.742Q-51.660 62.650-52.094 62.650Q-52.180 62.622-52.180 62.537L-52.152 62.427Q-52.145 62.386-52.074 62.369L-51.093 62.294Q-51.048 62.294-51.019 62.320Q-50.990 62.345-50.990 62.397L-51.711 65.275Q-51.499 65.186-51.351 65.061Q-51.202 64.936-50.864 64.632Q-50.525 64.328-50.284 64.196Q-50.043 64.065-49.784 64.065Q-49.565 64.065-49.416 64.210Q-49.267 64.355-49.267 64.574Q-49.267 64.769-49.384 64.914Q-49.500 65.059-49.688 65.059Q-49.804 65.059-49.886 64.986Q-49.968 64.912-49.968 64.793Q-49.968 64.639-49.855 64.509Q-49.743 64.379-49.589 64.379Q-49.671 64.287-49.801 64.287Q-50.060 64.287-50.310 64.456Q-50.559 64.625-50.911 64.946Q-51.264 65.268-51.438 65.374Q-50.337 65.493-50.337 66.122Q-50.337 66.221-50.373 66.385Q-50.409 66.549-50.409 66.638Q-50.409 66.997-50.170 66.997Q-49.927 66.997-49.780 66.725Q-49.633 66.454-49.555 66.122Q-49.531 66.061-49.476 66.061L-49.367 66.061Q-49.332 66.061-49.310 66.086Q-49.288 66.112-49.288 66.143Q-49.288 66.156-49.295 66.170Q-49.394 66.570-49.616 66.895Q-49.838 67.219-50.183 67.219Q-50.512 67.219-50.732 67.021Q-50.952 66.823-50.952 66.502Q-50.952 66.443-50.932 66.324Q-50.911 66.204-50.911 66.143Q-50.911 65.876-51.202 65.729Q-51.493 65.582-51.786 65.582L-52.132 66.977Q-52.162 67.079-52.256 67.149Q-52.350 67.219-52.453 67.219Q-52.552 67.219-52.624 67.156Q-52.696 67.093-52.696 66.990\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(59.941 -131.227)\">\u003Cpath d=\"M-48.185 67.812Q-47.931 68.036-47.296 68.036Q-47.074 68.036-46.909 67.937Q-46.744 67.838-46.660 67.663Q-46.576 67.487-46.576 67.272Q-46.576 67.050-46.663 66.878Q-46.749 66.706-46.915 66.606Q-47.081 66.505-47.301 66.505L-47.726 66.505Q-47.780 66.493-47.792 66.442L-47.792 66.381Q-47.780 66.337-47.726 66.320L-47.367 66.296Q-47.181 66.296-47.025 66.184Q-46.869 66.073-46.782 65.894Q-46.696 65.714-46.696 65.536Q-46.696 65.378-46.776 65.262Q-46.857 65.146-46.991 65.088Q-47.125 65.031-47.291 65.031Q-47.806 65.031-48.041 65.260Q-47.933 65.282-47.867 65.362Q-47.802 65.441-47.802 65.551Q-47.802 65.675-47.887 65.761Q-47.972 65.846-48.102 65.846Q-48.224 65.846-48.309 65.761Q-48.395 65.675-48.395 65.551Q-48.395 65.295-48.223 65.130Q-48.051 64.965-47.797 64.893Q-47.543 64.821-47.291 64.821Q-47.059 64.821-46.790 64.897Q-46.520 64.972-46.333 65.133Q-46.146 65.295-46.146 65.536Q-46.146 65.854-46.365 66.077Q-46.583 66.300-46.910 66.401Q-46.740 66.435-46.577 66.507Q-46.415 66.579-46.276 66.690Q-46.137 66.801-46.054 66.947Q-45.971 67.094-45.971 67.272Q-45.971 67.509-46.089 67.694Q-46.207 67.880-46.410 68.008Q-46.613 68.136-46.845 68.199Q-47.076 68.261-47.291 68.261Q-47.582 68.261-47.872 68.196Q-48.163 68.131-48.367 67.962Q-48.571 67.792-48.571 67.502Q-48.571 67.367-48.480 67.277Q-48.390 67.187-48.251 67.187Q-48.163 67.187-48.091 67.228Q-48.019 67.270-47.977 67.342Q-47.936 67.414-47.936 67.502Q-47.936 67.619-48.004 67.703Q-48.073 67.787-48.185 67.812\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.150\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"none\">\u003Cg transform=\"translate(85.549 -131.227)\">\u003Cpath d=\"M-52.696 66.990Q-52.696 66.949-52.689 66.925L-51.698 62.930Q-51.660 62.834-51.660 62.742Q-51.660 62.650-52.094 62.650Q-52.180 62.622-52.180 62.537L-52.152 62.427Q-52.145 62.386-52.074 62.369L-51.093 62.294Q-51.048 62.294-51.019 62.320Q-50.990 62.345-50.990 62.397L-51.711 65.275Q-51.499 65.186-51.351 65.061Q-51.202 64.936-50.864 64.632Q-50.525 64.328-50.284 64.196Q-50.043 64.065-49.784 64.065Q-49.565 64.065-49.416 64.210Q-49.267 64.355-49.267 64.574Q-49.267 64.769-49.384 64.914Q-49.500 65.059-49.688 65.059Q-49.804 65.059-49.886 64.986Q-49.968 64.912-49.968 64.793Q-49.968 64.639-49.855 64.509Q-49.743 64.379-49.589 64.379Q-49.671 64.287-49.801 64.287Q-50.060 64.287-50.310 64.456Q-50.559 64.625-50.911 64.946Q-51.264 65.268-51.438 65.374Q-50.337 65.493-50.337 66.122Q-50.337 66.221-50.373 66.385Q-50.409 66.549-50.409 66.638Q-50.409 66.997-50.170 66.997Q-49.927 66.997-49.780 66.725Q-49.633 66.454-49.555 66.122Q-49.531 66.061-49.476 66.061L-49.367 66.061Q-49.332 66.061-49.310 66.086Q-49.288 66.112-49.288 66.143Q-49.288 66.156-49.295 66.170Q-49.394 66.570-49.616 66.895Q-49.838 67.219-50.183 67.219Q-50.512 67.219-50.732 67.021Q-50.952 66.823-50.952 66.502Q-50.952 66.443-50.932 66.324Q-50.911 66.204-50.911 66.143Q-50.911 65.876-51.202 65.729Q-51.493 65.582-51.786 65.582L-52.132 66.977Q-52.162 67.079-52.256 67.149Q-52.350 67.219-52.453 67.219Q-52.552 67.219-52.624 67.156Q-52.696 67.093-52.696 66.990\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(85.549 -131.227)\">\u003Cpath d=\"M-47.001 67.336L-48.656 67.336L-48.656 67.096L-46.791 64.855Q-46.764 64.821-46.710 64.821L-46.581 64.821Q-46.552 64.821-46.526 64.844Q-46.500 64.867-46.500 64.901L-46.500 67.096L-45.885 67.096L-45.885 67.336L-46.500 67.336L-46.500 67.792Q-46.500 67.912-45.890 67.912L-45.890 68.151L-47.611 68.151L-47.611 67.912Q-47.001 67.912-47.001 67.792L-47.001 67.336M-46.962 65.375L-48.390 67.096L-46.962 67.096\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.150\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"none\">\u003Cg transform=\"translate(111.156 -131.227)\">\u003Cpath d=\"M-52.696 66.990Q-52.696 66.949-52.689 66.925L-51.698 62.930Q-51.660 62.834-51.660 62.742Q-51.660 62.650-52.094 62.650Q-52.180 62.622-52.180 62.537L-52.152 62.427Q-52.145 62.386-52.074 62.369L-51.093 62.294Q-51.048 62.294-51.019 62.320Q-50.990 62.345-50.990 62.397L-51.711 65.275Q-51.499 65.186-51.351 65.061Q-51.202 64.936-50.864 64.632Q-50.525 64.328-50.284 64.196Q-50.043 64.065-49.784 64.065Q-49.565 64.065-49.416 64.210Q-49.267 64.355-49.267 64.574Q-49.267 64.769-49.384 64.914Q-49.500 65.059-49.688 65.059Q-49.804 65.059-49.886 64.986Q-49.968 64.912-49.968 64.793Q-49.968 64.639-49.855 64.509Q-49.743 64.379-49.589 64.379Q-49.671 64.287-49.801 64.287Q-50.060 64.287-50.310 64.456Q-50.559 64.625-50.911 64.946Q-51.264 65.268-51.438 65.374Q-50.337 65.493-50.337 66.122Q-50.337 66.221-50.373 66.385Q-50.409 66.549-50.409 66.638Q-50.409 66.997-50.170 66.997Q-49.927 66.997-49.780 66.725Q-49.633 66.454-49.555 66.122Q-49.531 66.061-49.476 66.061L-49.367 66.061Q-49.332 66.061-49.310 66.086Q-49.288 66.112-49.288 66.143Q-49.288 66.156-49.295 66.170Q-49.394 66.570-49.616 66.895Q-49.838 67.219-50.183 67.219Q-50.512 67.219-50.732 67.021Q-50.952 66.823-50.952 66.502Q-50.952 66.443-50.932 66.324Q-50.911 66.204-50.911 66.143Q-50.911 65.876-51.202 65.729Q-51.493 65.582-51.786 65.582L-52.132 66.977Q-52.162 67.079-52.256 67.149Q-52.350 67.219-52.453 67.219Q-52.552 67.219-52.624 67.156Q-52.696 67.093-52.696 66.990\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(111.156 -131.227)\">\u003Cpath d=\"M-48.207 67.641Q-48.102 67.831-47.867 67.934Q-47.633 68.036-47.391 68.036Q-46.979 68.036-46.777 67.799Q-46.576 67.563-46.576 67.145Q-46.576 66.911-46.627 66.707Q-46.679 66.503-46.818 66.370Q-46.957 66.237-47.201 66.237Q-47.758 66.237-48.026 66.601Q-48.055 66.630-48.087 66.630L-48.146 66.630Q-48.173 66.630-48.197 66.606Q-48.221 66.581-48.221 66.557L-48.221 64.882Q-48.204 64.821-48.151 64.821Q-48.146 64.821-48.126 64.826Q-47.679 64.982-47.245 64.982Q-46.808 64.982-46.361 64.826Q-46.342 64.821-46.337 64.821Q-46.283 64.821-46.266 64.882L-46.266 64.941Q-46.268 64.963-46.281 64.982Q-46.515 65.221-46.816 65.348Q-47.118 65.475-47.450 65.475Q-47.741 65.475-47.980 65.422L-47.980 66.296Q-47.689 66.051-47.201 66.051Q-46.976 66.051-46.759 66.134Q-46.542 66.217-46.381 66.361Q-46.220 66.505-46.123 66.707Q-46.027 66.908-46.027 67.145Q-46.027 67.397-46.146 67.609Q-46.266 67.821-46.466 67.965Q-46.666 68.109-46.909 68.185Q-47.152 68.261-47.391 68.261Q-47.670 68.261-47.929 68.153Q-48.187 68.046-48.352 67.837Q-48.517 67.629-48.517 67.350Q-48.517 67.228-48.431 67.145Q-48.346 67.062-48.226 67.062Q-48.104 67.062-48.020 67.144Q-47.936 67.226-47.936 67.350Q-47.936 67.465-48.014 67.553Q-48.092 67.641-48.207 67.641\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.150\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"none\">\u003Cg transform=\"translate(-12.381 -114.406)\">\u003Cpath d=\"M-51.845 68.405L-51.814 68.293Q-51.807 68.245-51.739 68.228Q-51.393 68.228-51.270 68.187Q-51.226 68.166-51.205 68.120Q-51.185 68.074-51.158 67.985L-50.877 66.857Q-51.308 67.219-51.732 67.219Q-52.056 67.219-52.301 67.062Q-52.545 66.905-52.677 66.640Q-52.808 66.375-52.808 66.050Q-52.808 65.582-52.555 65.119Q-52.303 64.656-51.879 64.360Q-51.455 64.065-50.983 64.065Q-50.747 64.065-50.554 64.184Q-50.361 64.304-50.248 64.505Q-50.177 64.393-49.994 64.229Q-49.811 64.065-49.729 64.065Q-49.691 64.065-49.660 64.099Q-49.630 64.133-49.630 64.171L-50.597 68.033Q-50.618 68.094-50.618 68.159Q-50.618 68.228-50.177 68.228Q-50.095 68.252-50.095 68.340L-50.122 68.453Q-50.129 68.491-50.197 68.508L-51.759 68.508Q-51.790 68.508-51.817 68.472Q-51.845 68.436-51.845 68.405M-51.718 66.997Q-51.445 66.997-51.193 66.813Q-50.942 66.628-50.751 66.361L-50.382 64.881Q-50.419 64.721-50.500 64.584Q-50.580 64.447-50.706 64.367Q-50.833 64.287-50.997 64.287Q-51.205 64.287-51.392 64.411Q-51.578 64.536-51.716 64.728Q-51.855 64.919-51.940 65.121Q-52.053 65.415-52.137 65.760Q-52.221 66.105-52.221 66.355Q-52.221 66.611-52.092 66.804Q-51.964 66.997-51.718 66.997\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-12.381 -114.406)\">\u003Cpath d=\"M-46.776 68.151L-48.817 68.151L-48.817 67.912Q-48.036 67.912-48.036 67.792L-48.036 65.246Q-48.214 65.321-48.418 65.351Q-48.622 65.380-48.851 65.380L-48.851 65.141Q-48.515 65.141-48.231 65.072Q-47.948 65.004-47.738 64.821L-47.633 64.821Q-47.606 64.821-47.582 64.845Q-47.557 64.870-47.557 64.897L-47.557 67.792Q-47.557 67.912-46.776 67.912\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.150\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"none\">\u003Cg transform=\"translate(-12.381 -88.8)\">\u003Cpath d=\"M-51.845 68.405L-51.814 68.293Q-51.807 68.245-51.739 68.228Q-51.393 68.228-51.270 68.187Q-51.226 68.166-51.205 68.120Q-51.185 68.074-51.158 67.985L-50.877 66.857Q-51.308 67.219-51.732 67.219Q-52.056 67.219-52.301 67.062Q-52.545 66.905-52.677 66.640Q-52.808 66.375-52.808 66.050Q-52.808 65.582-52.555 65.119Q-52.303 64.656-51.879 64.360Q-51.455 64.065-50.983 64.065Q-50.747 64.065-50.554 64.184Q-50.361 64.304-50.248 64.505Q-50.177 64.393-49.994 64.229Q-49.811 64.065-49.729 64.065Q-49.691 64.065-49.660 64.099Q-49.630 64.133-49.630 64.171L-50.597 68.033Q-50.618 68.094-50.618 68.159Q-50.618 68.228-50.177 68.228Q-50.095 68.252-50.095 68.340L-50.122 68.453Q-50.129 68.491-50.197 68.508L-51.759 68.508Q-51.790 68.508-51.817 68.472Q-51.845 68.436-51.845 68.405M-51.718 66.997Q-51.445 66.997-51.193 66.813Q-50.942 66.628-50.751 66.361L-50.382 64.881Q-50.419 64.721-50.500 64.584Q-50.580 64.447-50.706 64.367Q-50.833 64.287-50.997 64.287Q-51.205 64.287-51.392 64.411Q-51.578 64.536-51.716 64.728Q-51.855 64.919-51.940 65.121Q-52.053 65.415-52.137 65.760Q-52.221 66.105-52.221 66.355Q-52.221 66.611-52.092 66.804Q-51.964 66.997-51.718 66.997\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-12.381 -88.8)\">\u003Cpath d=\"M-46.776 68.151L-49.103 68.151L-49.103 67.970Q-49.100 67.958-49.081 67.931L-48.041 67.055Q-47.733 66.796-47.579 66.652Q-47.426 66.508-47.296 66.291Q-47.167 66.073-47.167 65.832Q-47.167 65.590-47.294 65.414Q-47.421 65.238-47.625 65.149Q-47.828 65.060-48.068 65.060Q-48.275 65.060-48.471 65.148Q-48.666 65.236-48.771 65.402Q-48.651 65.402-48.574 65.494Q-48.497 65.585-48.497 65.700Q-48.497 65.827-48.584 65.916Q-48.671 66.005-48.798 66.005Q-48.927 66.005-49.015 65.915Q-49.103 65.824-49.103 65.700Q-49.103 65.419-48.930 65.220Q-48.756 65.021-48.485 64.921Q-48.214 64.821-47.941 64.821Q-47.616 64.821-47.311 64.928Q-47.006 65.036-46.809 65.263Q-46.613 65.490-46.613 65.827Q-46.613 66.064-46.726 66.256Q-46.840 66.449-47 66.587Q-47.160 66.725-47.442 66.913Q-47.724 67.101-47.802 67.160L-48.468 67.651L-48.012 67.651Q-47.579 67.651-47.285 67.644Q-46.991 67.638-46.976 67.626Q-46.898 67.528-46.842 67.167L-46.613 67.167\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.150\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"none\">\u003Cg transform=\"translate(-12.381 -63.192)\">\u003Cpath d=\"M-51.845 68.405L-51.814 68.293Q-51.807 68.245-51.739 68.228Q-51.393 68.228-51.270 68.187Q-51.226 68.166-51.205 68.120Q-51.185 68.074-51.158 67.985L-50.877 66.857Q-51.308 67.219-51.732 67.219Q-52.056 67.219-52.301 67.062Q-52.545 66.905-52.677 66.640Q-52.808 66.375-52.808 66.050Q-52.808 65.582-52.555 65.119Q-52.303 64.656-51.879 64.360Q-51.455 64.065-50.983 64.065Q-50.747 64.065-50.554 64.184Q-50.361 64.304-50.248 64.505Q-50.177 64.393-49.994 64.229Q-49.811 64.065-49.729 64.065Q-49.691 64.065-49.660 64.099Q-49.630 64.133-49.630 64.171L-50.597 68.033Q-50.618 68.094-50.618 68.159Q-50.618 68.228-50.177 68.228Q-50.095 68.252-50.095 68.340L-50.122 68.453Q-50.129 68.491-50.197 68.508L-51.759 68.508Q-51.790 68.508-51.817 68.472Q-51.845 68.436-51.845 68.405M-51.718 66.997Q-51.445 66.997-51.193 66.813Q-50.942 66.628-50.751 66.361L-50.382 64.881Q-50.419 64.721-50.500 64.584Q-50.580 64.447-50.706 64.367Q-50.833 64.287-50.997 64.287Q-51.205 64.287-51.392 64.411Q-51.578 64.536-51.716 64.728Q-51.855 64.919-51.940 65.121Q-52.053 65.415-52.137 65.760Q-52.221 66.105-52.221 66.355Q-52.221 66.611-52.092 66.804Q-51.964 66.997-51.718 66.997\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-12.381 -63.192)\">\u003Cpath d=\"M-48.771 67.812Q-48.517 68.036-47.882 68.036Q-47.660 68.036-47.495 67.937Q-47.330 67.838-47.246 67.663Q-47.162 67.487-47.162 67.272Q-47.162 67.050-47.249 66.878Q-47.335 66.706-47.501 66.606Q-47.667 66.505-47.887 66.505L-48.312 66.505Q-48.366 66.493-48.378 66.442L-48.378 66.381Q-48.366 66.337-48.312 66.320L-47.953 66.296Q-47.767 66.296-47.611 66.184Q-47.455 66.073-47.368 65.894Q-47.282 65.714-47.282 65.536Q-47.282 65.378-47.362 65.262Q-47.443 65.146-47.577 65.088Q-47.711 65.031-47.877 65.031Q-48.392 65.031-48.627 65.260Q-48.519 65.282-48.453 65.362Q-48.388 65.441-48.388 65.551Q-48.388 65.675-48.473 65.761Q-48.558 65.846-48.688 65.846Q-48.810 65.846-48.895 65.761Q-48.981 65.675-48.981 65.551Q-48.981 65.295-48.809 65.130Q-48.637 64.965-48.383 64.893Q-48.129 64.821-47.877 64.821Q-47.645 64.821-47.376 64.897Q-47.106 64.972-46.919 65.133Q-46.732 65.295-46.732 65.536Q-46.732 65.854-46.951 66.077Q-47.169 66.300-47.496 66.401Q-47.326 66.435-47.163 66.507Q-47.001 66.579-46.862 66.690Q-46.723 66.801-46.640 66.947Q-46.557 67.094-46.557 67.272Q-46.557 67.509-46.675 67.694Q-46.793 67.880-46.996 68.008Q-47.199 68.136-47.431 68.199Q-47.662 68.261-47.877 68.261Q-48.168 68.261-48.458 68.196Q-48.749 68.131-48.953 67.962Q-49.157 67.792-49.157 67.502Q-49.157 67.367-49.066 67.277Q-48.976 67.187-48.837 67.187Q-48.749 67.187-48.677 67.228Q-48.605 67.270-48.563 67.342Q-48.522 67.414-48.522 67.502Q-48.522 67.619-48.590 67.703Q-48.659 67.787-48.771 67.812\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.150\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"none\">\u003Cg transform=\"translate(-12.381 -37.585)\">\u003Cpath d=\"M-51.845 68.405L-51.814 68.293Q-51.807 68.245-51.739 68.228Q-51.393 68.228-51.270 68.187Q-51.226 68.166-51.205 68.120Q-51.185 68.074-51.158 67.985L-50.877 66.857Q-51.308 67.219-51.732 67.219Q-52.056 67.219-52.301 67.062Q-52.545 66.905-52.677 66.640Q-52.808 66.375-52.808 66.050Q-52.808 65.582-52.555 65.119Q-52.303 64.656-51.879 64.360Q-51.455 64.065-50.983 64.065Q-50.747 64.065-50.554 64.184Q-50.361 64.304-50.248 64.505Q-50.177 64.393-49.994 64.229Q-49.811 64.065-49.729 64.065Q-49.691 64.065-49.660 64.099Q-49.630 64.133-49.630 64.171L-50.597 68.033Q-50.618 68.094-50.618 68.159Q-50.618 68.228-50.177 68.228Q-50.095 68.252-50.095 68.340L-50.122 68.453Q-50.129 68.491-50.197 68.508L-51.759 68.508Q-51.790 68.508-51.817 68.472Q-51.845 68.436-51.845 68.405M-51.718 66.997Q-51.445 66.997-51.193 66.813Q-50.942 66.628-50.751 66.361L-50.382 64.881Q-50.419 64.721-50.500 64.584Q-50.580 64.447-50.706 64.367Q-50.833 64.287-50.997 64.287Q-51.205 64.287-51.392 64.411Q-51.578 64.536-51.716 64.728Q-51.855 64.919-51.940 65.121Q-52.053 65.415-52.137 65.760Q-52.221 66.105-52.221 66.355Q-52.221 66.611-52.092 66.804Q-51.964 66.997-51.718 66.997\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-12.381 -37.585)\">\u003Cpath d=\"M-47.587 67.336L-49.242 67.336L-49.242 67.096L-47.377 64.855Q-47.350 64.821-47.296 64.821L-47.167 64.821Q-47.138 64.821-47.112 64.844Q-47.086 64.867-47.086 64.901L-47.086 67.096L-46.471 67.096L-46.471 67.336L-47.086 67.336L-47.086 67.792Q-47.086 67.912-46.476 67.912L-46.476 68.151L-48.197 68.151L-48.197 67.912Q-47.587 67.912-47.587 67.792L-47.587 67.336M-47.548 65.375L-48.976 67.096L-47.548 67.096\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.150\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"none\">\u003Cg transform=\"translate(-12.381 -11.977)\">\u003Cpath d=\"M-51.845 68.405L-51.814 68.293Q-51.807 68.245-51.739 68.228Q-51.393 68.228-51.270 68.187Q-51.226 68.166-51.205 68.120Q-51.185 68.074-51.158 67.985L-50.877 66.857Q-51.308 67.219-51.732 67.219Q-52.056 67.219-52.301 67.062Q-52.545 66.905-52.677 66.640Q-52.808 66.375-52.808 66.050Q-52.808 65.582-52.555 65.119Q-52.303 64.656-51.879 64.360Q-51.455 64.065-50.983 64.065Q-50.747 64.065-50.554 64.184Q-50.361 64.304-50.248 64.505Q-50.177 64.393-49.994 64.229Q-49.811 64.065-49.729 64.065Q-49.691 64.065-49.660 64.099Q-49.630 64.133-49.630 64.171L-50.597 68.033Q-50.618 68.094-50.618 68.159Q-50.618 68.228-50.177 68.228Q-50.095 68.252-50.095 68.340L-50.122 68.453Q-50.129 68.491-50.197 68.508L-51.759 68.508Q-51.790 68.508-51.817 68.472Q-51.845 68.436-51.845 68.405M-51.718 66.997Q-51.445 66.997-51.193 66.813Q-50.942 66.628-50.751 66.361L-50.382 64.881Q-50.419 64.721-50.500 64.584Q-50.580 64.447-50.706 64.367Q-50.833 64.287-50.997 64.287Q-51.205 64.287-51.392 64.411Q-51.578 64.536-51.716 64.728Q-51.855 64.919-51.940 65.121Q-52.053 65.415-52.137 65.760Q-52.221 66.105-52.221 66.355Q-52.221 66.611-52.092 66.804Q-51.964 66.997-51.718 66.997\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(-12.381 -11.977)\">\u003Cpath d=\"M-48.793 67.641Q-48.688 67.831-48.453 67.934Q-48.219 68.036-47.977 68.036Q-47.565 68.036-47.363 67.799Q-47.162 67.563-47.162 67.145Q-47.162 66.911-47.213 66.707Q-47.265 66.503-47.404 66.370Q-47.543 66.237-47.787 66.237Q-48.344 66.237-48.612 66.601Q-48.641 66.630-48.673 66.630L-48.732 66.630Q-48.759 66.630-48.783 66.606Q-48.807 66.581-48.807 66.557L-48.807 64.882Q-48.790 64.821-48.737 64.821Q-48.732 64.821-48.712 64.826Q-48.265 64.982-47.831 64.982Q-47.394 64.982-46.947 64.826Q-46.928 64.821-46.923 64.821Q-46.869 64.821-46.852 64.882L-46.852 64.941Q-46.854 64.963-46.867 64.982Q-47.101 65.221-47.402 65.348Q-47.704 65.475-48.036 65.475Q-48.327 65.475-48.566 65.422L-48.566 66.296Q-48.275 66.051-47.787 66.051Q-47.562 66.051-47.345 66.134Q-47.128 66.217-46.967 66.361Q-46.806 66.505-46.709 66.707Q-46.613 66.908-46.613 67.145Q-46.613 67.397-46.732 67.609Q-46.852 67.821-47.052 67.965Q-47.252 68.109-47.495 68.185Q-47.738 68.261-47.977 68.261Q-48.256 68.261-48.515 68.153Q-48.773 68.046-48.938 67.837Q-49.103 67.629-49.103 67.350Q-49.103 67.228-49.017 67.145Q-48.932 67.062-48.812 67.062Q-48.690 67.062-48.606 67.144Q-48.522 67.226-48.522 67.350Q-48.522 67.465-48.600 67.553Q-48.678 67.641-48.793 67.641\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.150\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cpath fill=\"var(--tk-soft-accent)\" stroke=\"var(--tk-accent)\" d=\"M-53.222-35.278v-25.607h25.607v25.607Zm25.607-25.607\"\u002F>\u003Cpath fill=\"none\" d=\"M-27.615-35.278v-25.607h25.607v25.607ZM-23.006-39.887l16.39-16.389M-23.006-56.276l16.39 16.39\"\u002F>\u003Cpath fill=\"none\" d=\"M-2.008-35.278v-25.607H23.6v25.607ZM2.601-39.887l16.39-16.389M2.601-56.276l16.39 16.39M23.6-35.278v-25.607h25.607v25.607ZM28.209-39.887l16.389-16.389M28.209-56.276l16.389 16.39M49.207-35.278v-25.607h25.607v25.607ZM53.816-39.887l16.39-16.389M53.816-56.276l16.39 16.39\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" stroke=\"var(--tk-accent)\" d=\"M-53.222-9.67v-25.608h25.607V-9.67ZM-27.615-9.67v-25.608h25.607V-9.67Zm25.607-25.608\"\u002F>\u003Cpath fill=\"none\" d=\"M-2.008-9.67v-25.608H23.6V-9.67ZM2.601-14.28l16.39-16.389M2.601-30.669l16.39 16.39M23.6-9.67v-25.608h25.607V-9.67ZM28.209-14.28l16.389-16.389M28.209-30.669l16.389 16.39M49.207-9.67v-25.608h25.607V-9.67ZM53.816-14.28l16.39-16.389M53.816-30.669l16.39 16.39\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" stroke=\"var(--tk-accent)\" d=\"M-53.222 15.937V-9.671h25.607v25.608ZM-27.615 15.937V-9.671h25.607v25.608ZM-2.008-9.671\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" stroke=\"var(--tk-accent)\" d=\"M-2.008 15.937V-9.671H23.6v25.608ZM23.6-9.671\"\u002F>\u003Cpath fill=\"none\" d=\"M23.6 15.937V-9.671h25.607v25.608ZM28.209 11.328l16.389-16.39M28.209-5.061l16.389 16.389M49.207 15.937V-9.671h25.607v25.608ZM53.816 11.328l16.39-16.39M53.816-5.061l16.39 16.389\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" stroke=\"var(--tk-accent)\" d=\"M-53.222 41.544V15.937h25.607v25.607ZM-27.615 41.544V15.937h25.607v25.607Zm25.607-25.607\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" stroke=\"var(--tk-accent)\" d=\"M-2.008 41.544V15.937H23.6v25.607ZM23.6 41.544V15.937h25.607v25.607Zm25.607-25.607\"\u002F>\u003Cpath fill=\"none\" d=\"M49.207 41.544V15.937h25.607v25.607ZM53.816 36.935l16.39-16.39M53.816 20.546l16.39 16.389\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" stroke=\"var(--tk-accent)\" d=\"M-53.222 67.151V41.544h25.607v25.607ZM-27.615 67.151V41.544h25.607v25.607Zm25.607-25.607\"\u002F>\u003Cpath fill=\"var(--tk-soft-accent)\" stroke=\"var(--tk-accent)\" d=\"M-2.008 67.151V41.544H23.6v25.607ZM23.6 67.151V41.544h25.607v25.607ZM49.207 67.151V41.544h25.607v25.607Zm25.607-25.607\"\u002F>\u003C\u002Fg>\u003C\u002Fsvg>\u003Cfigcaption class=\"tikz-cap\">The causal mask. Below and on the diagonal (blue) a token may attend; above it (crossed) the score is set to \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6667em;vertical-align:-0.0833em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">−\u003C\u002Fspan>\u003Cspan class=\"mord\">∞\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan> so the softmax weight is \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6444em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">0\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>. Row \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6595em;\">\u003C\u002Fspan>\u003Cspan class=\"mord mathnormal\">i\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan> sees only tokens \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6595em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">1\u003C\u002Fspan>\u003Cspan class=\"mspace\" style=\"margin-right:0.1667em;\">\u003C\u002Fspan>\u003Cspan class=\"minner\">…\u003C\u002Fspan>\u003Cspan class=\"mspace\" style=\"margin-right:0.1667em;\">\u003C\u002Fspan>\u003Cspan class=\"mord mathnormal\">i\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>.\u003C\u002Ffigcaption>","\u003Csvg style=\"width:100%;max-width:466.975px;height:auto\" xmlns=\"http:\u002F\u002Fwww.w3.org\u002F2000\u002Fsvg\" viewBox=\"-75 -75 350.232 82.305\">\u003Cg stroke=\"currentColor\" style=\"stroke-miterlimit:10;stroke-width:.4\">\u003Cg stroke=\"var(--tk-accent)\">\u003Cpath fill=\"none\" d=\"M-45.486-38.79c0-5.5-4.459-9.959-9.959-9.959s-9.958 4.459-9.958 9.959 4.458 9.958 9.958 9.958 9.959-4.458 9.959-9.958Zm-9.959 0\"\u002F>\u003Cg transform=\"translate(33.44 2.578)\">\u003Cpath d=\"M-87.417-38.790L-90.210-38.790L-90.210-39.087Q-89.148-39.087-89.148-39.349L-89.148-43.517Q-89.577-43.302-90.257-43.302L-90.257-43.599Q-89.238-43.599-88.722-44.110L-88.577-44.110Q-88.503-44.091-88.484-44.013L-88.484-39.349Q-88.484-39.087-87.417-39.087\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"var(--tk-accent)\">\u003Cpath fill=\"none\" d=\"M-9.92-38.79c0-5.5-4.459-9.959-9.959-9.959s-9.958 4.459-9.958 9.959 4.458 9.958 9.958 9.958 9.958-4.458 9.958-9.958Zm-9.959 0\"\u002F>\u003Cg transform=\"translate(69.007 2.578)\">\u003Cpath d=\"M-87.425-38.790L-90.585-38.790L-90.585-38.997Q-90.585-39.024-90.562-39.056L-89.210-40.454Q-88.831-40.841-88.583-41.130Q-88.335-41.419-88.161-41.776Q-87.988-42.134-87.988-42.524Q-87.988-42.872-88.120-43.165Q-88.253-43.458-88.507-43.636Q-88.761-43.813-89.116-43.813Q-89.476-43.813-89.767-43.618Q-90.058-43.423-90.202-43.095L-90.148-43.095Q-89.964-43.095-89.839-42.974Q-89.714-42.852-89.714-42.661Q-89.714-42.481-89.839-42.352Q-89.964-42.224-90.148-42.224Q-90.327-42.224-90.456-42.352Q-90.585-42.481-90.585-42.661Q-90.585-43.063-90.365-43.399Q-90.144-43.735-89.779-43.923Q-89.413-44.110-89.011-44.110Q-88.531-44.110-88.115-43.923Q-87.698-43.735-87.447-43.374Q-87.195-43.013-87.195-42.524Q-87.195-42.165-87.349-41.862Q-87.503-41.560-87.755-41.300Q-88.007-41.040-88.357-40.755Q-88.706-40.470-88.874-40.317L-89.804-39.477L-89.089-39.477Q-87.714-39.477-87.675-39.517Q-87.605-39.595-87.562-39.780Q-87.519-39.966-87.476-40.255L-87.195-40.255\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"var(--tk-accent)\">\u003Cpath fill=\"none\" d=\"M25.645-38.79c0-5.5-4.458-9.959-9.958-9.959S5.729-44.29 5.729-38.79s4.458 9.958 9.958 9.958 9.958-4.458 9.958-9.958Zm-9.958 0\"\u002F>\u003Cg transform=\"translate(104.573 2.578)\">\u003Cpath d=\"M-90.218-39.423Q-90.027-39.149-89.671-39.022Q-89.316-38.895-88.933-38.895Q-88.597-38.895-88.388-39.081Q-88.179-39.267-88.083-39.560Q-87.988-39.852-87.988-40.165Q-87.988-40.489-88.085-40.784Q-88.183-41.079-88.396-41.263Q-88.609-41.446-88.941-41.446L-89.507-41.446Q-89.538-41.446-89.568-41.476Q-89.597-41.505-89.597-41.532L-89.597-41.614Q-89.597-41.649-89.568-41.675Q-89.538-41.700-89.507-41.700L-89.027-41.735Q-88.741-41.735-88.544-41.940Q-88.347-42.145-88.251-42.440Q-88.156-42.735-88.156-43.013Q-88.156-43.392-88.355-43.630Q-88.554-43.868-88.933-43.868Q-89.253-43.868-89.542-43.761Q-89.831-43.653-89.995-43.431Q-89.816-43.431-89.693-43.304Q-89.570-43.177-89.570-43.005Q-89.570-42.833-89.695-42.708Q-89.820-42.583-89.995-42.583Q-90.167-42.583-90.292-42.708Q-90.417-42.833-90.417-43.005Q-90.417-43.372-90.193-43.620Q-89.968-43.868-89.628-43.989Q-89.288-44.110-88.933-44.110Q-88.585-44.110-88.222-43.989Q-87.859-43.868-87.611-43.618Q-87.363-43.368-87.363-43.013Q-87.363-42.528-87.681-42.145Q-87.999-41.763-88.476-41.591Q-87.925-41.481-87.525-41.095Q-87.124-40.708-87.124-40.173Q-87.124-39.716-87.388-39.360Q-87.652-39.005-88.073-38.813Q-88.495-38.622-88.933-38.622Q-89.343-38.622-89.736-38.757Q-90.128-38.892-90.394-39.177Q-90.659-39.462-90.659-39.880Q-90.659-40.075-90.527-40.204Q-90.394-40.333-90.202-40.333Q-90.077-40.333-89.974-40.274Q-89.870-40.216-89.808-40.110Q-89.745-40.005-89.745-39.880Q-89.745-39.685-89.880-39.554Q-90.015-39.423-90.218-39.423\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cg stroke=\"var(--tk-accent)\">\u003Cpath fill=\"none\" d=\"M61.211-38.79c0-5.5-4.458-9.959-9.958-9.959s-9.959 4.459-9.959 9.959 4.459 9.958 9.959 9.958 9.958-4.458 9.958-9.958Zm-9.958 0\"\u002F>\u003Cg transform=\"translate(140.139 2.578)\">\u003Cpath d=\"M-88.531-40.102L-90.773-40.102L-90.773-40.399L-88.202-44.056Q-88.163-44.110-88.101-44.110L-87.956-44.110Q-87.906-44.110-87.874-44.079Q-87.843-44.048-87.843-43.997L-87.843-40.399L-87.011-40.399L-87.011-40.102L-87.843-40.102L-87.843-39.349Q-87.843-39.087-87.019-39.087L-87.019-38.790L-89.355-38.790L-89.355-39.087Q-88.531-39.087-88.531-39.349L-88.531-40.102M-88.476-43.204L-90.445-40.399L-88.476-40.399\" fill=\"var(--tk-accent)\" stroke=\"var(--tk-accent)\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cpath fill=\"none\" stroke=\"var(--tk-accent)\" d=\"M-47.44-45.045c6.01-4.695 13.547-4.695 19.556 0\"\u002F>\u003Cpath fill=\"none\" stroke=\"var(--tk-accent)\" d=\"M-47.896-45.588c16.239-14.621 39.795-14.621 56.034 0\"\u002F>\u003Cpath fill=\"none\" stroke=\"var(--tk-accent)\" d=\"M-48.388-46.098c25.08-25.972 67.504-25.972 92.584 0\"\u002F>\u003Cpath fill=\"none\" stroke=\"var(--tk-accent)\" d=\"M-11.874-45.045c6.01-4.695 13.547-4.695 19.556 0\"\u002F>\u003Cpath fill=\"none\" stroke=\"var(--tk-accent)\" d=\"M-12.33-45.588c16.239-14.621 39.795-14.621 56.034 0\"\u002F>\u003Cpath fill=\"none\" stroke=\"var(--tk-accent)\" d=\"M23.692-45.045c6.01-4.695 13.547-4.695 19.556 0\"\u002F>\u003Cg stroke=\"none\" font-size=\"8\">\u003Cg transform=\"translate(55.357 37.493)\">\u003Cpath d=\"M-87.328-48.489L-87.328-49.403Q-87.301-49.610-87.090-49.634L-86.922-49.634Q-86.758-49.610-86.699-49.450Q-86.496-48.810-85.769-48.810Q-85.562-48.810-85.334-48.845Q-85.105-48.880-84.937-48.995Q-84.769-49.110-84.769-49.313Q-84.769-49.524-84.992-49.638Q-85.215-49.751-85.488-49.794L-86.187-49.907Q-87.328-50.118-87.328-50.841Q-87.328-51.130-87.184-51.319Q-87.039-51.509-86.799-51.616Q-86.559-51.724-86.303-51.763Q-86.047-51.802-85.769-51.802Q-85.519-51.802-85.326-51.772Q-85.133-51.743-84.969-51.665Q-84.891-51.782-84.762-51.802L-84.684-51.802Q-84.586-51.790-84.523-51.727Q-84.461-51.665-84.449-51.571L-84.449-50.864Q-84.461-50.770-84.523-50.704Q-84.586-50.638-84.684-50.626L-84.852-50.626Q-84.945-50.638-85.012-50.704Q-85.078-50.770-85.090-50.864Q-85.090-51.243-85.785-51.243Q-86.133-51.243-86.451-51.161Q-86.769-51.079-86.769-50.833Q-86.769-50.567-86.098-50.458L-85.394-50.337Q-84.910-50.255-84.560-50.007Q-84.211-49.759-84.211-49.313Q-84.211-48.923-84.447-48.681Q-84.684-48.438-85.033-48.345Q-85.383-48.251-85.769-48.251Q-86.348-48.251-86.746-48.505Q-86.816-48.380-86.865-48.323Q-86.914-48.267-87.019-48.251L-87.090-48.251Q-87.305-48.274-87.328-48.489M-80.156-49.778L-82.598-49.778Q-82.543-49.501-82.346-49.278Q-82.148-49.056-81.871-48.933Q-81.594-48.810-81.309-48.810Q-80.836-48.810-80.613-49.099Q-80.605-49.110-80.549-49.216Q-80.492-49.321-80.443-49.364Q-80.394-49.407-80.301-49.419L-80.156-49.419Q-79.965-49.399-79.906-49.185L-79.906-49.130Q-79.973-48.829-80.203-48.632Q-80.434-48.435-80.746-48.343Q-81.059-48.251-81.363-48.251Q-81.848-48.251-82.287-48.479Q-82.727-48.708-82.994-49.108Q-83.262-49.509-83.262-50.001L-83.262-50.060Q-83.262-50.528-83.016-50.931Q-82.769-51.333-82.361-51.567Q-81.953-51.802-81.484-51.802Q-80.980-51.802-80.627-51.579Q-80.273-51.356-80.090-50.968Q-79.906-50.579-79.906-50.075L-79.906-50.017Q-79.965-49.802-80.156-49.778M-82.590-50.329L-80.562-50.329Q-80.609-50.739-80.848-50.991Q-81.086-51.243-81.484-51.243Q-81.879-51.243-82.185-50.981Q-82.492-50.720-82.590-50.329M-78.973-48.528L-78.973-48.618Q-78.922-48.825-78.727-48.849L-77.621-48.849L-77.621-52.618L-78.727-52.618Q-78.922-52.642-78.973-52.856L-78.973-52.946Q-78.922-53.153-78.727-53.177L-77.230-53.177Q-77.039-53.153-76.980-52.946L-76.980-48.849L-75.879-48.849Q-75.680-48.825-75.629-48.618L-75.629-48.528Q-75.680-48.313-75.879-48.290L-78.727-48.290Q-78.922-48.313-78.973-48.528M-74.855-48.528L-74.855-48.618Q-74.805-48.825-74.609-48.849L-73.727-48.849L-73.727-51.177L-74.582-51.177Q-74.781-51.200-74.832-51.419L-74.832-51.505Q-74.781-51.716-74.582-51.739L-73.727-51.739L-73.727-52.192Q-73.727-52.657-73.320-52.938Q-72.914-53.220-72.434-53.220Q-72.121-53.220-71.877-53.099Q-71.633-52.977-71.633-52.696Q-71.633-52.532-71.742-52.415Q-71.852-52.298-72.016-52.298Q-72.164-52.298-72.285-52.403Q-72.406-52.509-72.406-52.657L-72.488-52.657Q-72.641-52.657-72.777-52.597Q-72.914-52.536-73-52.421Q-73.086-52.306-73.086-52.161L-73.086-51.739L-72.055-51.739Q-71.859-51.720-71.809-51.505L-71.809-51.419Q-71.859-51.200-72.055-51.177L-73.086-51.177L-73.086-48.849L-72.207-48.849Q-72.012-48.825-71.961-48.618L-71.961-48.528Q-72.012-48.313-72.207-48.290L-74.609-48.290Q-74.805-48.313-74.855-48.528M-67.434-50.427L-70.176-50.427Q-70.301-50.438-70.387-50.524Q-70.473-50.610-70.473-50.739Q-70.473-50.868-70.387-50.950Q-70.301-51.032-70.176-51.044L-67.434-51.044Q-67.309-51.032-67.227-50.950Q-67.144-50.868-67.144-50.739Q-67.144-50.610-67.227-50.524Q-67.309-50.438-67.434-50.427M-66.277-49.403Q-66.277-49.849-65.863-50.106Q-65.449-50.364-64.908-50.464Q-64.367-50.563-63.859-50.571Q-63.859-50.786-63.994-50.938Q-64.129-51.091-64.336-51.167Q-64.543-51.243-64.754-51.243Q-65.098-51.243-65.258-51.220L-65.258-51.161Q-65.258-50.993-65.377-50.878Q-65.496-50.763-65.660-50.763Q-65.836-50.763-65.951-50.886Q-66.066-51.009-66.066-51.177Q-66.066-51.583-65.685-51.692Q-65.305-51.802-64.746-51.802Q-64.477-51.802-64.209-51.724Q-63.941-51.645-63.717-51.495Q-63.492-51.345-63.355-51.124Q-63.219-50.903-63.219-50.626L-63.219-48.907Q-63.219-48.849-62.691-48.849Q-62.496-48.829-62.445-48.618L-62.445-48.528Q-62.496-48.313-62.691-48.290L-62.836-48.290Q-63.180-48.290-63.408-48.337Q-63.637-48.384-63.781-48.571Q-64.242-48.251-64.949-48.251Q-65.285-48.251-65.590-48.392Q-65.894-48.532-66.086-48.794Q-66.277-49.056-66.277-49.403M-65.637-49.395Q-65.637-49.122-65.394-48.966Q-65.152-48.810-64.867-48.810Q-64.648-48.810-64.416-48.868Q-64.184-48.927-64.021-49.065Q-63.859-49.204-63.859-49.427L-63.859-50.017Q-64.141-50.017-64.557-49.960Q-64.973-49.903-65.305-49.765Q-65.637-49.626-65.637-49.395M-61.238-49.395L-61.238-51.177L-61.988-51.177Q-62.187-51.200-62.238-51.419L-62.238-51.505Q-62.187-51.716-61.988-51.739L-61.238-51.739L-61.238-52.489Q-61.187-52.696-60.988-52.724L-60.844-52.724Q-60.648-52.696-60.598-52.489L-60.598-51.739L-59.238-51.739Q-59.047-51.720-58.988-51.505L-58.988-51.419Q-59.043-51.200-59.238-51.177L-60.598-51.177L-60.598-49.427Q-60.598-48.810-60.023-48.810Q-59.773-48.810-59.609-48.995Q-59.445-49.181-59.445-49.427Q-59.445-49.520-59.373-49.591Q-59.301-49.661-59.199-49.673L-59.055-49.673Q-58.855-49.649-58.805-49.442L-58.805-49.395Q-58.805-49.071-58.988-48.808Q-59.172-48.544-59.465-48.397Q-59.758-48.251-60.078-48.251Q-60.590-48.251-60.914-48.561Q-61.238-48.872-61.238-49.395M-56.992-49.395L-56.992-51.177L-57.742-51.177Q-57.941-51.200-57.992-51.419L-57.992-51.505Q-57.941-51.716-57.742-51.739L-56.992-51.739L-56.992-52.489Q-56.941-52.696-56.742-52.724L-56.598-52.724Q-56.402-52.696-56.352-52.489L-56.352-51.739L-54.992-51.739Q-54.801-51.720-54.742-51.505L-54.742-51.419Q-54.797-51.200-54.992-51.177L-56.352-51.177L-56.352-49.427Q-56.352-48.810-55.777-48.810Q-55.527-48.810-55.363-48.995Q-55.199-49.181-55.199-49.427Q-55.199-49.520-55.127-49.591Q-55.055-49.661-54.953-49.673L-54.809-49.673Q-54.609-49.649-54.559-49.442L-54.559-49.395Q-54.559-49.071-54.742-48.808Q-54.926-48.544-55.219-48.397Q-55.512-48.251-55.832-48.251Q-56.344-48.251-56.668-48.561Q-56.992-48.872-56.992-49.395M-50.434-49.778L-52.875-49.778Q-52.820-49.501-52.623-49.278Q-52.426-49.056-52.148-48.933Q-51.871-48.810-51.586-48.810Q-51.113-48.810-50.891-49.099Q-50.883-49.110-50.826-49.216Q-50.769-49.321-50.721-49.364Q-50.672-49.407-50.578-49.419L-50.434-49.419Q-50.242-49.399-50.184-49.185L-50.184-49.130Q-50.250-48.829-50.480-48.632Q-50.711-48.435-51.023-48.343Q-51.336-48.251-51.641-48.251Q-52.125-48.251-52.564-48.479Q-53.004-48.708-53.271-49.108Q-53.539-49.509-53.539-50.001L-53.539-50.060Q-53.539-50.528-53.293-50.931Q-53.047-51.333-52.639-51.567Q-52.230-51.802-51.762-51.802Q-51.258-51.802-50.904-51.579Q-50.551-51.356-50.367-50.968Q-50.184-50.579-50.184-50.075L-50.184-50.017Q-50.242-49.802-50.434-49.778M-52.867-50.329L-50.840-50.329Q-50.887-50.739-51.125-50.991Q-51.363-51.243-51.762-51.243Q-52.156-51.243-52.463-50.981Q-52.769-50.720-52.867-50.329M-49.629-48.528L-49.629-48.618Q-49.586-48.825-49.379-48.849L-48.957-48.849L-48.957-51.177L-49.379-51.177Q-49.586-51.200-49.629-51.419L-49.629-51.505Q-49.582-51.716-49.379-51.739L-48.562-51.739Q-48.367-51.716-48.316-51.505L-48.316-51.419L-48.324-51.395Q-48.098-51.575-47.824-51.677Q-47.551-51.778-47.258-51.778Q-46.910-51.778-46.672-51.638Q-46.434-51.497-46.318-51.239Q-46.203-50.981-46.203-50.626L-46.203-48.849L-45.777-48.849Q-45.570-48.825-45.531-48.618L-45.531-48.528Q-45.570-48.313-45.777-48.290L-47.172-48.290Q-47.367-48.313-47.418-48.528L-47.418-48.618Q-47.367-48.829-47.172-48.849L-46.844-48.849L-46.844-50.595Q-46.844-50.903-46.934-51.061Q-47.023-51.220-47.316-51.220Q-47.586-51.220-47.814-51.089Q-48.043-50.958-48.180-50.729Q-48.316-50.501-48.316-50.235L-48.316-48.849L-47.891-48.849Q-47.684-48.825-47.644-48.618L-47.644-48.528Q-47.684-48.313-47.891-48.290L-49.379-48.290Q-49.586-48.313-49.629-48.528M-44.254-49.395L-44.254-51.177L-45.004-51.177Q-45.203-51.200-45.254-51.419L-45.254-51.505Q-45.203-51.716-45.004-51.739L-44.254-51.739L-44.254-52.489Q-44.203-52.696-44.004-52.724L-43.859-52.724Q-43.664-52.696-43.613-52.489L-43.613-51.739L-42.254-51.739Q-42.062-51.720-42.004-51.505L-42.004-51.419Q-42.059-51.200-42.254-51.177L-43.613-51.177L-43.613-49.427Q-43.613-48.810-43.039-48.810Q-42.789-48.810-42.625-48.995Q-42.461-49.181-42.461-49.427Q-42.461-49.520-42.389-49.591Q-42.316-49.661-42.215-49.673L-42.070-49.673Q-41.871-49.649-41.820-49.442L-41.820-49.395Q-41.820-49.071-42.004-48.808Q-42.187-48.544-42.480-48.397Q-42.773-48.251-43.094-48.251Q-43.605-48.251-43.930-48.561Q-44.254-48.872-44.254-49.395M-40.590-48.528L-40.590-48.618Q-40.539-48.825-40.344-48.849L-39.305-48.849L-39.305-51.177L-40.277-51.177Q-40.477-51.200-40.527-51.419L-40.527-51.505Q-40.477-51.716-40.277-51.739L-38.910-51.739Q-38.715-51.720-38.664-51.505L-38.664-48.849L-37.750-48.849Q-37.555-48.825-37.504-48.618L-37.504-48.528Q-37.555-48.313-37.750-48.290L-40.344-48.290Q-40.539-48.313-40.590-48.528M-39.559-52.716L-39.559-52.770Q-39.559-52.942-39.422-53.063Q-39.285-53.185-39.109-53.185Q-38.937-53.185-38.801-53.063Q-38.664-52.942-38.664-52.770L-38.664-52.716Q-38.664-52.540-38.801-52.419Q-38.937-52.298-39.109-52.298Q-39.285-52.298-39.422-52.419Q-39.559-52.540-39.559-52.716M-34.840-48.251Q-35.312-48.251-35.697-48.495Q-36.082-48.739-36.305-49.149Q-36.527-49.560-36.527-50.017Q-36.527-50.360-36.402-50.683Q-36.277-51.005-36.047-51.259Q-35.816-51.513-35.510-51.657Q-35.203-51.802-34.840-51.802Q-34.477-51.802-34.164-51.655Q-33.852-51.509-33.629-51.263Q-33.406-51.017-33.279-50.696Q-33.152-50.376-33.152-50.017Q-33.152-49.560-33.377-49.147Q-33.602-48.735-33.986-48.493Q-34.371-48.251-34.840-48.251M-34.840-48.810Q-34.375-48.810-34.084-49.204Q-33.793-49.599-33.793-50.083Q-33.793-50.376-33.928-50.644Q-34.062-50.911-34.303-51.077Q-34.543-51.243-34.840-51.243Q-35.144-51.243-35.383-51.077Q-35.621-50.911-35.756-50.644Q-35.891-50.376-35.891-50.083Q-35.891-49.602-35.598-49.206Q-35.305-48.810-34.840-48.810M-32.644-48.528L-32.644-48.618Q-32.602-48.825-32.394-48.849L-31.973-48.849L-31.973-51.177L-32.394-51.177Q-32.602-51.200-32.644-51.419L-32.644-51.505Q-32.598-51.716-32.394-51.739L-31.578-51.739Q-31.383-51.716-31.332-51.505L-31.332-51.419L-31.340-51.395Q-31.113-51.575-30.840-51.677Q-30.566-51.778-30.273-51.778Q-29.926-51.778-29.687-51.638Q-29.449-51.497-29.334-51.239Q-29.219-50.981-29.219-50.626L-29.219-48.849L-28.793-48.849Q-28.586-48.825-28.547-48.618L-28.547-48.528Q-28.586-48.313-28.793-48.290L-30.187-48.290Q-30.383-48.313-30.434-48.528L-30.434-48.618Q-30.383-48.829-30.187-48.849L-29.859-48.849L-29.859-50.595Q-29.859-50.903-29.949-51.061Q-30.039-51.220-30.332-51.220Q-30.602-51.220-30.830-51.089Q-31.059-50.958-31.195-50.729Q-31.332-50.501-31.332-50.235L-31.332-48.849L-30.906-48.849Q-30.699-48.825-30.660-48.618L-30.660-48.528Q-30.699-48.313-30.906-48.290L-32.394-48.290Q-32.602-48.313-32.644-48.528\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(55.357 37.493)\">\u003Cpath d=\"M-88.394-36.798Q-89.007-37.255-89.409-37.890Q-89.812-38.524-90.007-39.270Q-90.202-40.017-90.202-40.790Q-90.202-41.563-90.007-42.310Q-89.812-43.056-89.409-43.690Q-89.007-44.325-88.394-44.782Q-88.382-44.786-88.374-44.788Q-88.366-44.790-88.355-44.790L-88.277-44.790Q-88.238-44.790-88.212-44.763Q-88.187-44.735-88.187-44.692Q-88.187-44.642-88.218-44.622Q-88.726-44.169-89.048-43.546Q-89.370-42.923-89.511-42.227Q-89.652-41.532-89.652-40.790Q-89.652-40.056-89.513-39.356Q-89.374-38.657-89.050-38.032Q-88.726-37.407-88.218-36.958Q-88.187-36.938-88.187-36.888Q-88.187-36.845-88.212-36.817Q-88.238-36.790-88.277-36.790L-88.355-36.790Q-88.363-36.794-88.372-36.796Q-88.382-36.798-88.394-36.798M-87.370-39.622Q-87.370-40.106-86.968-40.401Q-86.566-40.696-86.015-40.815Q-85.464-40.935-84.972-40.935L-84.972-41.224Q-84.972-41.450-85.087-41.657Q-85.202-41.864-85.400-41.983Q-85.597-42.102-85.827-42.102Q-86.253-42.102-86.538-41.997Q-86.468-41.970-86.421-41.915Q-86.374-41.860-86.349-41.790Q-86.323-41.720-86.323-41.645Q-86.323-41.540-86.374-41.448Q-86.425-41.356-86.517-41.306Q-86.609-41.255-86.714-41.255Q-86.820-41.255-86.911-41.306Q-87.003-41.356-87.054-41.448Q-87.105-41.540-87.105-41.645Q-87.105-42.063-86.716-42.210Q-86.327-42.356-85.827-42.356Q-85.495-42.356-85.142-42.226Q-84.788-42.095-84.560-41.841Q-84.331-41.587-84.331-41.239L-84.331-39.438Q-84.331-39.306-84.259-39.196Q-84.187-39.087-84.058-39.087Q-83.933-39.087-83.865-39.192Q-83.796-39.298-83.796-39.438L-83.796-39.950L-83.515-39.950L-83.515-39.438Q-83.515-39.235-83.632-39.077Q-83.749-38.919-83.931-38.835Q-84.113-38.751-84.316-38.751Q-84.546-38.751-84.698-38.923Q-84.851-39.095-84.882-39.325Q-85.042-39.044-85.351-38.878Q-85.659-38.712-86.011-38.712Q-86.523-38.712-86.947-38.935Q-87.370-39.157-87.370-39.622M-86.683-39.622Q-86.683-39.337-86.456-39.151Q-86.230-38.966-85.937-38.966Q-85.691-38.966-85.466-39.083Q-85.241-39.200-85.107-39.403Q-84.972-39.606-84.972-39.860L-84.972-40.692Q-85.238-40.692-85.523-40.638Q-85.808-40.583-86.079-40.454Q-86.351-40.325-86.517-40.118Q-86.683-39.911-86.683-39.622M-81.308-38.790L-83.140-38.790L-83.140-39.087Q-82.866-39.087-82.698-39.134Q-82.531-39.181-82.531-39.349L-82.531-43.509Q-82.531-43.724-82.593-43.819Q-82.656-43.915-82.775-43.936Q-82.894-43.958-83.140-43.958L-83.140-44.255L-81.917-44.341L-81.917-39.349Q-81.917-39.181-81.749-39.134Q-81.581-39.087-81.308-39.087L-81.308-38.790M-78.948-38.790L-80.781-38.790L-80.781-39.087Q-80.507-39.087-80.339-39.134Q-80.171-39.181-80.171-39.349L-80.171-43.509Q-80.171-43.724-80.234-43.819Q-80.296-43.915-80.415-43.936Q-80.534-43.958-80.781-43.958L-80.781-44.255L-79.558-44.341L-79.558-39.349Q-79.558-39.181-79.390-39.134Q-79.222-39.087-78.948-39.087L-78.948-38.790M-76.390-40.239L-78.644-40.239L-78.644-40.790L-76.390-40.790L-76.390-40.239M-73.788-37.239L-75.644-37.239L-75.644-37.532Q-75.374-37.532-75.206-37.577Q-75.038-37.622-75.038-37.798L-75.038-41.622Q-75.038-41.829-75.195-41.882Q-75.351-41.935-75.644-41.935L-75.644-42.231L-74.421-42.317L-74.421-41.852Q-74.191-42.075-73.876-42.196Q-73.562-42.317-73.222-42.317Q-72.749-42.317-72.345-42.071Q-71.941-41.825-71.708-41.409Q-71.476-40.993-71.476-40.517Q-71.476-40.142-71.624-39.813Q-71.773-39.485-72.042-39.233Q-72.312-38.981-72.656-38.847Q-72.999-38.712-73.359-38.712Q-73.648-38.712-73.919-38.833Q-74.191-38.954-74.398-39.165L-74.398-37.798Q-74.398-37.622-74.230-37.577Q-74.062-37.532-73.788-37.532L-73.788-37.239M-74.398-41.454L-74.398-39.614Q-74.245-39.325-73.984-39.145Q-73.722-38.966-73.413-38.966Q-73.128-38.966-72.906-39.104Q-72.683-39.243-72.531-39.474Q-72.378-39.704-72.300-39.976Q-72.222-40.247-72.222-40.517Q-72.222-40.849-72.347-41.206Q-72.472-41.563-72.720-41.800Q-72.968-42.036-73.316-42.036Q-73.640-42.036-73.935-41.880Q-74.230-41.724-74.398-41.454M-70.855-39.622Q-70.855-40.106-70.452-40.401Q-70.050-40.696-69.499-40.815Q-68.948-40.935-68.456-40.935L-68.456-41.224Q-68.456-41.450-68.572-41.657Q-68.687-41.864-68.884-41.983Q-69.081-42.102-69.312-42.102Q-69.738-42.102-70.023-41.997Q-69.952-41.970-69.906-41.915Q-69.859-41.860-69.833-41.790Q-69.808-41.720-69.808-41.645Q-69.808-41.540-69.859-41.448Q-69.909-41.356-70.001-41.306Q-70.093-41.255-70.198-41.255Q-70.304-41.255-70.396-41.306Q-70.488-41.356-70.538-41.448Q-70.589-41.540-70.589-41.645Q-70.589-42.063-70.200-42.210Q-69.812-42.356-69.312-42.356Q-68.980-42.356-68.626-42.226Q-68.273-42.095-68.044-41.841Q-67.816-41.587-67.816-41.239L-67.816-39.438Q-67.816-39.306-67.743-39.196Q-67.671-39.087-67.542-39.087Q-67.417-39.087-67.349-39.192Q-67.281-39.298-67.281-39.438L-67.281-39.950L-66.999-39.950L-66.999-39.438Q-66.999-39.235-67.116-39.077Q-67.234-38.919-67.415-38.835Q-67.597-38.751-67.800-38.751Q-68.031-38.751-68.183-38.923Q-68.335-39.095-68.366-39.325Q-68.527-39.044-68.835-38.878Q-69.144-38.712-69.495-38.712Q-70.007-38.712-70.431-38.935Q-70.855-39.157-70.855-39.622M-70.167-39.622Q-70.167-39.337-69.941-39.151Q-69.714-38.966-69.421-38.966Q-69.175-38.966-68.950-39.083Q-68.726-39.200-68.591-39.403Q-68.456-39.606-68.456-39.860L-68.456-40.692Q-68.722-40.692-69.007-40.638Q-69.292-40.583-69.564-40.454Q-69.835-40.325-70.001-40.118Q-70.167-39.911-70.167-39.622M-64.847-38.790L-66.624-38.790L-66.624-39.087Q-66.351-39.087-66.183-39.134Q-66.015-39.181-66.015-39.349L-66.015-41.485Q-66.015-41.700-66.072-41.796Q-66.128-41.892-66.241-41.913Q-66.355-41.935-66.601-41.935L-66.601-42.231L-65.402-42.317L-65.402-39.349Q-65.402-39.181-65.255-39.134Q-65.109-39.087-64.847-39.087L-64.847-38.790M-66.288-43.712Q-66.288-43.903-66.154-44.034Q-66.019-44.165-65.823-44.165Q-65.702-44.165-65.599-44.102Q-65.495-44.040-65.433-43.936Q-65.370-43.833-65.370-43.712Q-65.370-43.517-65.501-43.382Q-65.632-43.247-65.823-43.247Q-66.023-43.247-66.156-43.380Q-66.288-43.513-66.288-43.712M-62.339-38.790L-64.320-38.790L-64.320-39.087Q-64.050-39.087-63.882-39.132Q-63.714-39.177-63.714-39.349L-63.714-41.485Q-63.714-41.700-63.777-41.796Q-63.839-41.892-63.956-41.913Q-64.073-41.935-64.320-41.935L-64.320-42.231L-63.152-42.317L-63.152-41.532Q-63.073-41.743-62.921-41.929Q-62.769-42.114-62.570-42.216Q-62.370-42.317-62.144-42.317Q-61.898-42.317-61.706-42.173Q-61.515-42.028-61.515-41.798Q-61.515-41.642-61.620-41.532Q-61.726-41.423-61.882-41.423Q-62.038-41.423-62.148-41.532Q-62.257-41.642-62.257-41.798Q-62.257-41.958-62.152-42.063Q-62.476-42.063-62.691-41.835Q-62.906-41.606-63.001-41.267Q-63.097-40.927-63.097-40.622L-63.097-39.349Q-63.097-39.181-62.870-39.134Q-62.644-39.087-62.339-39.087L-62.339-38.790M-60.991-38.798L-60.991-40.020Q-60.991-40.048-60.960-40.079Q-60.929-40.110-60.906-40.110L-60.800-40.110Q-60.730-40.110-60.714-40.048Q-60.652-39.727-60.513-39.487Q-60.374-39.247-60.142-39.106Q-59.909-38.966-59.601-38.966Q-59.363-38.966-59.154-39.026Q-58.945-39.087-58.808-39.235Q-58.671-39.384-58.671-39.630Q-58.671-39.884-58.882-40.050Q-59.093-40.216-59.363-40.270L-59.984-40.384Q-60.390-40.462-60.691-40.718Q-60.991-40.974-60.991-41.349Q-60.991-41.716-60.790-41.938Q-60.589-42.161-60.265-42.259Q-59.941-42.356-59.601-42.356Q-59.136-42.356-58.839-42.149L-58.616-42.333Q-58.593-42.356-58.562-42.356L-58.511-42.356Q-58.480-42.356-58.452-42.329Q-58.425-42.302-58.425-42.270L-58.425-41.286Q-58.425-41.255-58.450-41.226Q-58.476-41.196-58.511-41.196L-58.616-41.196Q-58.652-41.196-58.679-41.224Q-58.706-41.251-58.706-41.286Q-58.706-41.685-58.958-41.905Q-59.210-42.126-59.609-42.126Q-59.964-42.126-60.247-42.003Q-60.531-41.880-60.531-41.575Q-60.531-41.356-60.329-41.224Q-60.128-41.091-59.882-41.048L-59.257-40.935Q-58.827-40.845-58.519-40.548Q-58.210-40.251-58.210-39.837Q-58.210-39.267-58.609-38.989Q-59.007-38.712-59.601-38.712Q-60.152-38.712-60.503-39.048L-60.800-38.735Q-60.823-38.712-60.859-38.712L-60.906-38.712Q-60.929-38.712-60.960-38.743Q-60.991-38.774-60.991-38.798M-57.097-37.384Q-57.097-37.407-57.066-37.454Q-56.773-37.716-56.607-38.083Q-56.441-38.450-56.441-38.837L-56.441-38.895Q-56.570-38.790-56.738-38.790Q-56.929-38.790-57.066-38.923Q-57.202-39.056-57.202-39.255Q-57.202-39.446-57.066-39.579Q-56.929-39.712-56.738-39.712Q-56.437-39.712-56.312-39.442Q-56.187-39.173-56.187-38.837Q-56.187-38.388-56.368-37.974Q-56.550-37.560-56.890-37.263Q-56.913-37.239-56.952-37.239Q-56.999-37.239-57.048-37.284Q-57.097-37.329-57.097-37.384\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(55.357 37.493)\">\u003Cpath d=\"M-50.585-37.239L-52.440-37.239L-52.440-37.532Q-52.171-37.532-52.003-37.577Q-51.835-37.622-51.835-37.798L-51.835-41.622Q-51.835-41.829-51.991-41.882Q-52.147-41.935-52.440-41.935L-52.440-42.231L-51.218-42.317L-51.218-41.852Q-50.987-42.075-50.673-42.196Q-50.358-42.317-50.019-42.317Q-49.546-42.317-49.142-42.071Q-48.737-41.825-48.505-41.409Q-48.272-40.993-48.272-40.517Q-48.272-40.142-48.421-39.813Q-48.569-39.485-48.839-39.233Q-49.108-38.981-49.452-38.847Q-49.796-38.712-50.155-38.712Q-50.444-38.712-50.716-38.833Q-50.987-38.954-51.194-39.165L-51.194-37.798Q-51.194-37.622-51.026-37.577Q-50.858-37.532-50.585-37.532L-50.585-37.239M-51.194-41.454L-51.194-39.614Q-51.042-39.325-50.780-39.145Q-50.519-38.966-50.210-38.966Q-49.925-38.966-49.702-39.104Q-49.479-39.243-49.327-39.474Q-49.175-39.704-49.097-39.976Q-49.019-40.247-49.019-40.517Q-49.019-40.849-49.144-41.206Q-49.269-41.563-49.517-41.800Q-49.765-42.036-50.112-42.036Q-50.436-42.036-50.731-41.880Q-51.026-41.724-51.194-41.454M-47.651-39.622Q-47.651-40.106-47.249-40.401Q-46.847-40.696-46.296-40.815Q-45.745-40.935-45.253-40.935L-45.253-41.224Q-45.253-41.450-45.368-41.657Q-45.483-41.864-45.681-41.983Q-45.878-42.102-46.108-42.102Q-46.534-42.102-46.819-41.997Q-46.749-41.970-46.702-41.915Q-46.655-41.860-46.630-41.790Q-46.604-41.720-46.604-41.645Q-46.604-41.540-46.655-41.448Q-46.706-41.356-46.798-41.306Q-46.890-41.255-46.995-41.255Q-47.101-41.255-47.192-41.306Q-47.284-41.356-47.335-41.448Q-47.386-41.540-47.386-41.645Q-47.386-42.063-46.997-42.210Q-46.608-42.356-46.108-42.356Q-45.776-42.356-45.423-42.226Q-45.069-42.095-44.841-41.841Q-44.612-41.587-44.612-41.239L-44.612-39.438Q-44.612-39.306-44.540-39.196Q-44.468-39.087-44.339-39.087Q-44.214-39.087-44.145-39.192Q-44.077-39.298-44.077-39.438L-44.077-39.950L-43.796-39.950L-43.796-39.438Q-43.796-39.235-43.913-39.077Q-44.030-38.919-44.212-38.835Q-44.394-38.751-44.597-38.751Q-44.827-38.751-44.979-38.923Q-45.132-39.095-45.163-39.325Q-45.323-39.044-45.632-38.878Q-45.940-38.712-46.292-38.712Q-46.804-38.712-47.227-38.935Q-47.651-39.157-47.651-39.622M-46.964-39.622Q-46.964-39.337-46.737-39.151Q-46.511-38.966-46.218-38.966Q-45.972-38.966-45.747-39.083Q-45.522-39.200-45.388-39.403Q-45.253-39.606-45.253-39.860L-45.253-40.692Q-45.519-40.692-45.804-40.638Q-46.089-40.583-46.360-40.454Q-46.632-40.325-46.798-40.118Q-46.964-39.911-46.964-39.622M-42.878-39.751L-42.878-41.942L-43.581-41.942L-43.581-42.196Q-43.226-42.196-42.983-42.429Q-42.741-42.661-42.630-43.009Q-42.519-43.356-42.519-43.712L-42.237-43.712L-42.237-42.239L-41.061-42.239L-41.061-41.942L-42.237-41.942L-42.237-39.767Q-42.237-39.446-42.118-39.218Q-41.999-38.989-41.718-38.989Q-41.538-38.989-41.421-39.112Q-41.304-39.235-41.251-39.415Q-41.198-39.595-41.198-39.767L-41.198-40.239L-40.917-40.239L-40.917-39.751Q-40.917-39.497-41.022-39.257Q-41.128-39.017-41.325-38.864Q-41.522-38.712-41.780-38.712Q-42.097-38.712-42.349-38.835Q-42.601-38.958-42.739-39.192Q-42.878-39.427-42.878-39.751M-38.269-38.790L-40.124-38.790L-40.124-39.087Q-39.851-39.087-39.683-39.134Q-39.515-39.181-39.515-39.349L-39.515-43.509Q-39.515-43.724-39.577-43.819Q-39.640-43.915-39.759-43.936Q-39.878-43.958-40.124-43.958L-40.124-44.255L-38.901-44.341L-38.901-41.638Q-38.776-41.849-38.589-41.999Q-38.401-42.149-38.175-42.233Q-37.948-42.317-37.702-42.317Q-36.534-42.317-36.534-41.239L-36.534-39.349Q-36.534-39.181-36.364-39.134Q-36.194-39.087-35.925-39.087L-35.925-38.790L-37.780-38.790L-37.780-39.087Q-37.507-39.087-37.339-39.134Q-37.171-39.181-37.171-39.349L-37.171-41.224Q-37.171-41.606-37.292-41.835Q-37.413-42.063-37.765-42.063Q-38.077-42.063-38.331-41.901Q-38.585-41.739-38.731-41.470Q-38.878-41.200-38.878-40.903L-38.878-39.349Q-38.878-39.181-38.708-39.134Q-38.538-39.087-38.269-39.087\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(55.357 37.493)\">\u003Cpath d=\"M-29.278-38.790L-32.071-38.790L-32.071-39.087Q-31.009-39.087-31.009-39.349L-31.009-43.517Q-31.438-43.302-32.118-43.302L-32.118-43.599Q-31.099-43.599-30.583-44.110L-30.438-44.110Q-30.364-44.091-30.345-44.013L-30.345-39.349Q-30.345-39.087-29.278-39.087L-29.278-38.790M-27.985-36.790L-28.067-36.790Q-28.102-36.790-28.128-36.819Q-28.153-36.849-28.153-36.888Q-28.153-36.938-28.122-36.958Q-27.735-37.294-27.452-37.743Q-27.169-38.192-27.003-38.692Q-26.837-39.192-26.763-39.710Q-26.688-40.227-26.688-40.790Q-26.688-41.360-26.763-41.876Q-26.837-42.392-27.003-42.888Q-27.169-43.384-27.448-43.831Q-27.727-44.278-28.122-44.622Q-28.153-44.642-28.153-44.692Q-28.153-44.731-28.128-44.761Q-28.102-44.790-28.067-44.790L-27.985-44.790Q-27.974-44.790-27.964-44.788Q-27.954-44.786-27.946-44.782Q-27.333-44.325-26.931-43.690Q-26.528-43.056-26.333-42.310Q-26.138-41.563-26.138-40.790Q-26.138-40.017-26.333-39.270Q-26.528-38.524-26.931-37.890Q-27.333-37.255-27.946-36.798Q-27.958-36.798-27.966-36.796Q-27.974-36.794-27.985-36.790\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"M165.064-38.79c0-5.5-4.459-9.959-9.959-9.959s-9.958 4.459-9.958 9.959 4.458 9.958 9.958 9.958 9.959-4.458 9.959-9.958Zm-9.959 0\"\u002F>\u003Cg transform=\"translate(243.991 2.578)\">\u003Cpath d=\"M-87.417-38.790L-90.210-38.790L-90.210-39.087Q-89.148-39.087-89.148-39.349L-89.148-43.517Q-89.577-43.302-90.257-43.302L-90.257-43.599Q-89.238-43.599-88.722-44.110L-88.577-44.110Q-88.503-44.091-88.484-44.013L-88.484-39.349Q-88.484-39.087-87.417-39.087\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"M200.63-38.79c0-5.5-4.459-9.959-9.959-9.959s-9.958 4.459-9.958 9.959 4.458 9.958 9.958 9.958 9.959-4.458 9.959-9.958Zm-9.959 0\"\u002F>\u003Cg transform=\"translate(279.557 2.578)\">\u003Cpath d=\"M-87.425-38.790L-90.585-38.790L-90.585-38.997Q-90.585-39.024-90.562-39.056L-89.210-40.454Q-88.831-40.841-88.583-41.130Q-88.335-41.419-88.161-41.776Q-87.988-42.134-87.988-42.524Q-87.988-42.872-88.120-43.165Q-88.253-43.458-88.507-43.636Q-88.761-43.813-89.116-43.813Q-89.476-43.813-89.767-43.618Q-90.058-43.423-90.202-43.095L-90.148-43.095Q-89.964-43.095-89.839-42.974Q-89.714-42.852-89.714-42.661Q-89.714-42.481-89.839-42.352Q-89.964-42.224-90.148-42.224Q-90.327-42.224-90.456-42.352Q-90.585-42.481-90.585-42.661Q-90.585-43.063-90.365-43.399Q-90.144-43.735-89.779-43.923Q-89.413-44.110-89.011-44.110Q-88.531-44.110-88.115-43.923Q-87.698-43.735-87.447-43.374Q-87.195-43.013-87.195-42.524Q-87.195-42.165-87.349-41.862Q-87.503-41.560-87.755-41.300Q-88.007-41.040-88.357-40.755Q-88.706-40.470-88.874-40.317L-89.804-39.477L-89.089-39.477Q-87.714-39.477-87.675-39.517Q-87.605-39.595-87.562-39.780Q-87.519-39.966-87.476-40.255L-87.195-40.255\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"M236.196-38.79c0-5.5-4.459-9.959-9.959-9.959s-9.958 4.459-9.958 9.959 4.458 9.958 9.958 9.958 9.959-4.458 9.959-9.958Zm-9.959 0\"\u002F>\u003Cg transform=\"translate(315.123 2.578)\">\u003Cpath d=\"M-90.218-39.423Q-90.027-39.149-89.671-39.022Q-89.316-38.895-88.933-38.895Q-88.597-38.895-88.388-39.081Q-88.179-39.267-88.083-39.560Q-87.988-39.852-87.988-40.165Q-87.988-40.489-88.085-40.784Q-88.183-41.079-88.396-41.263Q-88.609-41.446-88.941-41.446L-89.507-41.446Q-89.538-41.446-89.568-41.476Q-89.597-41.505-89.597-41.532L-89.597-41.614Q-89.597-41.649-89.568-41.675Q-89.538-41.700-89.507-41.700L-89.027-41.735Q-88.741-41.735-88.544-41.940Q-88.347-42.145-88.251-42.440Q-88.156-42.735-88.156-43.013Q-88.156-43.392-88.355-43.630Q-88.554-43.868-88.933-43.868Q-89.253-43.868-89.542-43.761Q-89.831-43.653-89.995-43.431Q-89.816-43.431-89.693-43.304Q-89.570-43.177-89.570-43.005Q-89.570-42.833-89.695-42.708Q-89.820-42.583-89.995-42.583Q-90.167-42.583-90.292-42.708Q-90.417-42.833-90.417-43.005Q-90.417-43.372-90.193-43.620Q-89.968-43.868-89.628-43.989Q-89.288-44.110-88.933-44.110Q-88.585-44.110-88.222-43.989Q-87.859-43.868-87.611-43.618Q-87.363-43.368-87.363-43.013Q-87.363-42.528-87.681-42.145Q-87.999-41.763-88.476-41.591Q-87.925-41.481-87.525-41.095Q-87.124-40.708-87.124-40.173Q-87.124-39.716-87.388-39.360Q-87.652-39.005-88.073-38.813Q-88.495-38.622-88.933-38.622Q-89.343-38.622-89.736-38.757Q-90.128-38.892-90.394-39.177Q-90.659-39.462-90.659-39.880Q-90.659-40.075-90.527-40.204Q-90.394-40.333-90.202-40.333Q-90.077-40.333-89.974-40.274Q-89.870-40.216-89.808-40.110Q-89.745-40.005-89.745-39.880Q-89.745-39.685-89.880-39.554Q-90.015-39.423-90.218-39.423\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cpath fill=\"none\" d=\"M271.762-38.79c0-5.5-4.459-9.959-9.959-9.959s-9.958 4.459-9.958 9.959 4.458 9.958 9.958 9.958 9.959-4.458 9.959-9.958Zm-9.959 0\"\u002F>\u003Cg transform=\"translate(350.689 2.578)\">\u003Cpath d=\"M-88.531-40.102L-90.773-40.102L-90.773-40.399L-88.202-44.056Q-88.163-44.110-88.101-44.110L-87.956-44.110Q-87.906-44.110-87.874-44.079Q-87.843-44.048-87.843-43.997L-87.843-40.399L-87.011-40.399L-87.011-40.102L-87.843-40.102L-87.843-39.349Q-87.843-39.087-87.019-39.087L-87.019-38.790L-89.355-38.790L-89.355-39.087Q-88.531-39.087-88.531-39.349L-88.531-40.102M-88.476-43.204L-90.445-40.399L-88.476-40.399\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.240\"\u002F>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M165.264-38.79h12.649\"\u002F>\u003Cpath stroke=\"none\" d=\"m180.513-38.79-4.16-2.08 1.56 2.08-1.56 2.08\"\u002F>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M200.83-38.79h12.649\"\u002F>\u003Cpath stroke=\"none\" d=\"m216.079-38.79-4.16-2.08 1.56 2.08-1.56 2.08\"\u002F>\u003C\u002Fg>\u003Cg style=\"stroke-width:.8\">\u003Cpath fill=\"none\" d=\"M236.396-38.79h12.649\"\u002F>\u003Cpath stroke=\"none\" d=\"m251.645-38.79-4.16-2.08 1.56 2.08-1.56 2.08\"\u002F>\u003C\u002Fg>\u003Cg stroke=\"none\" font-size=\"7\">\u003Cg transform=\"translate(272.147 35.93)\">\u003Cpath d=\"M-80.333-46.790L-82.069-46.790L-82.069-47.070Q-81.840-47.070-81.691-47.104Q-81.543-47.139-81.543-47.279L-81.543-49.128Q-81.543-49.398-81.650-49.459Q-81.758-49.521-82.069-49.521L-82.069-49.801L-81.040-49.876L-81.040-49.169Q-80.910-49.477-80.668-49.676Q-80.425-49.876-80.107-49.876Q-79.888-49.876-79.717-49.752Q-79.546-49.627-79.546-49.415Q-79.546-49.278-79.646-49.179Q-79.745-49.080-79.878-49.080Q-80.015-49.080-80.114-49.179Q-80.213-49.278-80.213-49.415Q-80.213-49.555-80.114-49.654Q-80.404-49.654-80.604-49.458Q-80.804-49.261-80.897-48.967Q-80.989-48.673-80.989-48.393L-80.989-47.279Q-80.989-47.070-80.333-47.070L-80.333-46.790M-79.003-48.325Q-79.003-48.646-78.878-48.935Q-78.753-49.224-78.528-49.447Q-78.302-49.671-78.007-49.791Q-77.711-49.911-77.393-49.911Q-77.065-49.911-76.803-49.811Q-76.542-49.712-76.366-49.530Q-76.190-49.347-76.096-49.089Q-76.002-48.831-76.002-48.499Q-76.002-48.407-76.084-48.386L-78.340-48.386L-78.340-48.325Q-78.340-47.737-78.056-47.354Q-77.772-46.971-77.205-46.971Q-76.884-46.971-76.616-47.164Q-76.347-47.357-76.258-47.672Q-76.251-47.713-76.176-47.727L-76.084-47.727Q-76.002-47.703-76.002-47.631Q-76.002-47.624-76.009-47.597Q-76.122-47.200-76.492-46.961Q-76.863-46.722-77.287-46.722Q-77.725-46.722-78.125-46.930Q-78.524-47.139-78.764-47.506Q-79.003-47.873-79.003-48.325M-78.333-48.595L-76.518-48.595Q-76.518-48.872-76.616-49.124Q-76.713-49.377-76.911-49.533Q-77.109-49.688-77.393-49.688Q-77.670-49.688-77.884-49.530Q-78.097-49.371-78.215-49.116Q-78.333-48.861-78.333-48.595M-75.414-48.301Q-75.414-48.629-75.279-48.930Q-75.144-49.230-74.908-49.451Q-74.672-49.671-74.368-49.791Q-74.064-49.911-73.739-49.911Q-73.233-49.911-72.885-49.808Q-72.536-49.706-72.536-49.330Q-72.536-49.183-72.634-49.082Q-72.731-48.981-72.878-48.981Q-73.032-48.981-73.131-49.080Q-73.230-49.179-73.230-49.330Q-73.230-49.518-73.090-49.610Q-73.292-49.661-73.732-49.661Q-74.088-49.661-74.317-49.465Q-74.546-49.268-74.647-48.959Q-74.748-48.649-74.748-48.301Q-74.748-47.952-74.621-47.646Q-74.495-47.340-74.240-47.156Q-73.985-46.971-73.630-46.971Q-73.408-46.971-73.223-47.055Q-73.039-47.139-72.904-47.294Q-72.769-47.450-72.710-47.658Q-72.697-47.713-72.642-47.713L-72.529-47.713Q-72.499-47.713-72.476-47.689Q-72.454-47.665-72.454-47.631L-72.454-47.610Q-72.540-47.323-72.728-47.125Q-72.916-46.927-73.180-46.824Q-73.445-46.722-73.739-46.722Q-74.170-46.722-74.558-46.928Q-74.946-47.135-75.180-47.498Q-75.414-47.860-75.414-48.301M-71.292-47.624L-71.292-49.128Q-71.292-49.398-71.400-49.459Q-71.507-49.521-71.818-49.521L-71.818-49.801L-70.711-49.876L-70.711-47.644L-70.711-47.624Q-70.711-47.344-70.660-47.200Q-70.608-47.057-70.467-47Q-70.325-46.944-70.038-46.944Q-69.785-46.944-69.580-47.084Q-69.375-47.224-69.258-47.450Q-69.142-47.675-69.142-47.925L-69.142-49.128Q-69.142-49.398-69.250-49.459Q-69.357-49.521-69.668-49.521L-69.668-49.801L-68.561-49.876L-68.561-47.463Q-68.561-47.272-68.508-47.190Q-68.455-47.108-68.354-47.089Q-68.253-47.070-68.038-47.070L-68.038-46.790L-69.115-46.722L-69.115-47.286Q-69.224-47.104-69.369-46.981Q-69.515-46.858-69.701-46.790Q-69.887-46.722-70.089-46.722Q-71.292-46.722-71.292-47.624M-65.700-46.790L-67.437-46.790L-67.437-47.070Q-67.208-47.070-67.059-47.104Q-66.910-47.139-66.910-47.279L-66.910-49.128Q-66.910-49.398-67.018-49.459Q-67.126-49.521-67.437-49.521L-67.437-49.801L-66.408-49.876L-66.408-49.169Q-66.278-49.477-66.035-49.676Q-65.793-49.876-65.475-49.876Q-65.256-49.876-65.085-49.752Q-64.914-49.627-64.914-49.415Q-64.914-49.278-65.013-49.179Q-65.112-49.080-65.246-49.080Q-65.382-49.080-65.481-49.179Q-65.581-49.278-65.581-49.415Q-65.581-49.555-65.481-49.654Q-65.772-49.654-65.972-49.458Q-66.172-49.261-66.264-48.967Q-66.356-48.673-66.356-48.393L-66.356-47.279Q-66.356-47.070-65.700-47.070L-65.700-46.790M-62.580-46.790L-64.316-46.790L-64.316-47.070Q-64.087-47.070-63.938-47.104Q-63.790-47.139-63.790-47.279L-63.790-49.128Q-63.790-49.398-63.897-49.459Q-64.005-49.521-64.316-49.521L-64.316-49.801L-63.287-49.876L-63.287-49.169Q-63.157-49.477-62.915-49.676Q-62.672-49.876-62.354-49.876Q-62.135-49.876-61.964-49.752Q-61.793-49.627-61.793-49.415Q-61.793-49.278-61.893-49.179Q-61.992-49.080-62.125-49.080Q-62.262-49.080-62.361-49.179Q-62.460-49.278-62.460-49.415Q-62.460-49.555-62.361-49.654Q-62.651-49.654-62.851-49.458Q-63.051-49.261-63.144-48.967Q-63.236-48.673-63.236-48.393L-63.236-47.279Q-63.236-47.070-62.580-47.070L-62.580-46.790M-61.250-48.325Q-61.250-48.646-61.125-48.935Q-61.001-49.224-60.775-49.447Q-60.549-49.671-60.254-49.791Q-59.958-49.911-59.640-49.911Q-59.312-49.911-59.051-49.811Q-58.789-49.712-58.613-49.530Q-58.437-49.347-58.343-49.089Q-58.249-48.831-58.249-48.499Q-58.249-48.407-58.331-48.386L-60.587-48.386L-60.587-48.325Q-60.587-47.737-60.303-47.354Q-60.020-46.971-59.452-46.971Q-59.131-46.971-58.863-47.164Q-58.594-47.357-58.505-47.672Q-58.499-47.713-58.423-47.727L-58.331-47.727Q-58.249-47.703-58.249-47.631Q-58.249-47.624-58.256-47.597Q-58.369-47.200-58.740-46.961Q-59.110-46.722-59.534-46.722Q-59.972-46.722-60.372-46.930Q-60.772-47.139-61.011-47.506Q-61.250-47.873-61.250-48.325M-60.580-48.595L-58.765-48.595Q-58.765-48.872-58.863-49.124Q-58.960-49.377-59.158-49.533Q-59.356-49.688-59.640-49.688Q-59.917-49.688-60.131-49.530Q-60.344-49.371-60.462-49.116Q-60.580-48.861-60.580-48.595M-55.980-46.790L-57.613-46.790L-57.613-47.070Q-57.384-47.070-57.236-47.104Q-57.087-47.139-57.087-47.279L-57.087-49.128Q-57.087-49.398-57.195-49.459Q-57.302-49.521-57.613-49.521L-57.613-49.801L-56.554-49.876L-56.554-49.227Q-56.383-49.535-56.079-49.706Q-55.774-49.876-55.429-49.876Q-54.923-49.876-54.640-49.653Q-54.356-49.429-54.356-48.933L-54.356-47.279Q-54.356-47.142-54.207-47.106Q-54.059-47.070-53.833-47.070L-53.833-46.790L-55.463-46.790L-55.463-47.070Q-55.234-47.070-55.086-47.104Q-54.937-47.139-54.937-47.279L-54.937-48.919Q-54.937-49.254-55.057-49.454Q-55.176-49.654-55.491-49.654Q-55.761-49.654-55.995-49.518Q-56.229-49.381-56.367-49.147Q-56.506-48.913-56.506-48.639L-56.506-47.279Q-56.506-47.142-56.355-47.106Q-56.205-47.070-55.980-47.070L-55.980-46.790M-53.245-48.301Q-53.245-48.629-53.110-48.930Q-52.975-49.230-52.739-49.451Q-52.503-49.671-52.199-49.791Q-51.895-49.911-51.570-49.911Q-51.064-49.911-50.716-49.808Q-50.367-49.706-50.367-49.330Q-50.367-49.183-50.465-49.082Q-50.562-48.981-50.709-48.981Q-50.863-48.981-50.962-49.080Q-51.061-49.179-51.061-49.330Q-51.061-49.518-50.921-49.610Q-51.123-49.661-51.564-49.661Q-51.919-49.661-52.148-49.465Q-52.377-49.268-52.478-48.959Q-52.579-48.649-52.579-48.301Q-52.579-47.952-52.452-47.646Q-52.326-47.340-52.071-47.156Q-51.816-46.971-51.461-46.971Q-51.239-46.971-51.054-47.055Q-50.870-47.139-50.735-47.294Q-50.600-47.450-50.542-47.658Q-50.528-47.713-50.473-47.713L-50.360-47.713Q-50.330-47.713-50.307-47.689Q-50.285-47.665-50.285-47.631L-50.285-47.610Q-50.371-47.323-50.559-47.125Q-50.747-46.927-51.012-46.824Q-51.276-46.722-51.570-46.722Q-52.001-46.722-52.389-46.928Q-52.777-47.135-53.011-47.498Q-53.245-47.860-53.245-48.301M-49.738-48.325Q-49.738-48.646-49.614-48.935Q-49.489-49.224-49.263-49.447Q-49.038-49.671-48.742-49.791Q-48.446-49.911-48.128-49.911Q-47.800-49.911-47.539-49.811Q-47.277-49.712-47.101-49.530Q-46.925-49.347-46.831-49.089Q-46.737-48.831-46.737-48.499Q-46.737-48.407-46.819-48.386L-49.075-48.386L-49.075-48.325Q-49.075-47.737-48.792-47.354Q-48.508-46.971-47.940-46.971Q-47.619-46.971-47.351-47.164Q-47.083-47.357-46.994-47.672Q-46.987-47.713-46.912-47.727L-46.819-47.727Q-46.737-47.703-46.737-47.631Q-46.737-47.624-46.744-47.597Q-46.857-47.200-47.228-46.961Q-47.599-46.722-48.022-46.722Q-48.460-46.722-48.860-46.930Q-49.260-47.139-49.499-47.506Q-49.738-47.873-49.738-48.325M-49.068-48.595L-47.253-48.595Q-47.253-48.872-47.351-49.124Q-47.448-49.377-47.647-49.533Q-47.845-49.688-48.128-49.688Q-48.405-49.688-48.619-49.530Q-48.833-49.371-48.950-49.116Q-49.068-48.861-49.068-48.595\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(272.147 35.93)\">\u003Cpath d=\"M-88.567-37.040Q-89.117-37.440-89.488-37.995Q-89.859-38.551-90.040-39.197Q-90.221-39.843-90.221-40.540Q-90.221-41.053-90.121-41.548Q-90.020-42.044-89.815-42.495Q-89.610-42.946-89.297-43.338Q-88.984-43.729-88.567-44.033Q-88.557-44.037-88.550-44.038Q-88.543-44.040-88.533-44.040L-88.465-44.040Q-88.430-44.040-88.408-44.016Q-88.386-43.992-88.386-43.955Q-88.386-43.910-88.413-43.893Q-88.762-43.592-89.015-43.208Q-89.268-42.823-89.420-42.382Q-89.572-41.941-89.644-41.485Q-89.716-41.029-89.716-40.540Q-89.716-39.539-89.406-38.652Q-89.097-37.765-88.413-37.180Q-88.386-37.163-88.386-37.119Q-88.386-37.081-88.408-37.057Q-88.430-37.033-88.465-37.033L-88.533-37.033Q-88.540-37.037-88.548-37.038Q-88.557-37.040-88.567-37.040M-87.576-40.301Q-87.576-40.629-87.441-40.930Q-87.306-41.230-87.070-41.451Q-86.834-41.671-86.530-41.791Q-86.226-41.911-85.901-41.911Q-85.395-41.911-85.047-41.808Q-84.698-41.706-84.698-41.330Q-84.698-41.183-84.795-41.082Q-84.893-40.981-85.040-40.981Q-85.194-40.981-85.293-41.080Q-85.392-41.179-85.392-41.330Q-85.392-41.518-85.252-41.610Q-85.453-41.661-85.894-41.661Q-86.250-41.661-86.479-41.465Q-86.708-41.268-86.809-40.959Q-86.909-40.649-86.909-40.301Q-86.909-39.952-86.783-39.646Q-86.657-39.340-86.402-39.156Q-86.147-38.971-85.792-38.971Q-85.570-38.971-85.385-39.055Q-85.200-39.139-85.065-39.294Q-84.930-39.450-84.872-39.658Q-84.859-39.713-84.804-39.713L-84.691-39.713Q-84.660-39.713-84.638-39.689Q-84.616-39.665-84.616-39.631L-84.616-39.610Q-84.701-39.323-84.889-39.125Q-85.077-38.927-85.342-38.824Q-85.607-38.722-85.901-38.722Q-86.332-38.722-86.720-38.928Q-87.108-39.135-87.342-39.498Q-87.576-39.860-87.576-40.301\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(272.147 35.93)\">\u003Cpath d=\"M-82.550-38.790L-84.184-38.790L-84.184-39.070Q-83.955-39.070-83.806-39.104Q-83.657-39.139-83.657-39.279L-83.657-42.898Q-83.657-43.168-83.765-43.230Q-83.873-43.291-84.184-43.291L-84.184-43.572L-83.104-43.647L-83.104-41.261Q-82.998-41.446-82.820-41.588Q-82.642-41.729-82.434-41.803Q-82.225-41.876-82-41.876Q-81.494-41.876-81.210-41.653Q-80.926-41.429-80.926-40.933L-80.926-39.279Q-80.926-39.142-80.778-39.106Q-80.629-39.070-80.403-39.070L-80.403-38.790L-82.034-38.790L-82.034-39.070Q-81.805-39.070-81.656-39.104Q-81.507-39.139-81.507-39.279L-81.507-40.919Q-81.507-41.254-81.627-41.454Q-81.747-41.654-82.061-41.654Q-82.331-41.654-82.565-41.518Q-82.799-41.381-82.938-41.147Q-83.076-40.913-83.076-40.639L-83.076-39.279Q-83.076-39.142-82.926-39.106Q-82.775-39.070-82.550-39.070L-82.550-38.790M-79.757-39.518Q-79.757-39.850-79.534-40.077Q-79.310-40.304-78.966-40.432Q-78.623-40.561-78.250-40.613Q-77.878-40.666-77.573-40.666L-77.573-40.919Q-77.573-41.124-77.681-41.304Q-77.789-41.483-77.970-41.586Q-78.151-41.688-78.359-41.688Q-78.766-41.688-79.002-41.596Q-78.913-41.559-78.867-41.475Q-78.821-41.391-78.821-41.289Q-78.821-41.193-78.867-41.114Q-78.913-41.036-78.994-40.991Q-79.074-40.947-79.163-40.947Q-79.313-40.947-79.414-41.044Q-79.515-41.142-79.515-41.289Q-79.515-41.911-78.359-41.911Q-78.148-41.911-77.898-41.847Q-77.649-41.784-77.447-41.665Q-77.245-41.545-77.119-41.360Q-76.992-41.176-76.992-40.933L-76.992-39.357Q-76.992-39.241-76.931-39.145Q-76.869-39.050-76.756-39.050Q-76.647-39.050-76.582-39.144Q-76.517-39.238-76.517-39.357L-76.517-39.805L-76.251-39.805L-76.251-39.357Q-76.251-39.087-76.478-38.922Q-76.705-38.756-76.985-38.756Q-77.194-38.756-77.331-38.910Q-77.467-39.063-77.491-39.279Q-77.638-39.012-77.920-38.867Q-78.202-38.722-78.527-38.722Q-78.804-38.722-79.088-38.797Q-79.371-38.872-79.564-39.051Q-79.757-39.231-79.757-39.518M-79.142-39.518Q-79.142-39.344-79.041-39.214Q-78.941-39.084-78.785-39.014Q-78.629-38.944-78.465-38.944Q-78.247-38.944-78.038-39.041Q-77.830-39.139-77.702-39.320Q-77.573-39.501-77.573-39.727L-77.573-40.455Q-77.898-40.455-78.264-40.364Q-78.629-40.273-78.886-40.061Q-79.142-39.850-79.142-39.518M-74.217-38.790L-75.769-38.790L-75.769-39.070Q-75.543-39.070-75.394-39.104Q-75.246-39.139-75.246-39.279L-75.246-41.128Q-75.246-41.316-75.294-41.400Q-75.341-41.483-75.439-41.502Q-75.536-41.521-75.748-41.521L-75.748-41.801L-74.692-41.876L-74.692-39.279Q-74.692-39.139-74.560-39.104Q-74.429-39.070-74.217-39.070L-74.217-38.790M-75.488-43.097Q-75.488-43.268-75.365-43.387Q-75.242-43.507-75.071-43.507Q-74.904-43.507-74.781-43.387Q-74.658-43.268-74.658-43.097Q-74.658-42.922-74.781-42.799Q-74.904-42.676-75.071-42.676Q-75.242-42.676-75.365-42.799Q-75.488-42.922-75.488-43.097M-71.889-38.790L-73.523-38.790L-73.523-39.070Q-73.294-39.070-73.145-39.104Q-72.997-39.139-72.997-39.279L-72.997-41.128Q-72.997-41.398-73.104-41.459Q-73.212-41.521-73.523-41.521L-73.523-41.801L-72.463-41.876L-72.463-41.227Q-72.293-41.535-71.988-41.706Q-71.684-41.876-71.339-41.876Q-70.833-41.876-70.549-41.653Q-70.266-41.429-70.266-40.933L-70.266-39.279Q-70.266-39.142-70.117-39.106Q-69.968-39.070-69.743-39.070L-69.743-38.790L-71.373-38.790L-71.373-39.070Q-71.144-39.070-70.995-39.104Q-70.847-39.139-70.847-39.279L-70.847-40.919Q-70.847-41.254-70.966-41.454Q-71.086-41.654-71.400-41.654Q-71.671-41.654-71.905-41.518Q-72.139-41.381-72.277-41.147Q-72.416-40.913-72.416-40.639L-72.416-39.279Q-72.416-39.142-72.265-39.106Q-72.115-39.070-71.889-39.070L-71.889-38.790M-68.656-37.560Q-68.656-37.594-68.629-37.621Q-68.359-37.850-68.210-38.173Q-68.061-38.496-68.061-38.852L-68.061-38.889Q-68.171-38.790-68.335-38.790Q-68.516-38.790-68.635-38.910Q-68.755-39.029-68.755-39.210Q-68.755-39.385-68.635-39.504Q-68.516-39.624-68.335-39.624Q-68.078-39.624-67.959-39.385Q-67.839-39.145-67.839-38.852Q-67.839-38.452-68.008-38.081Q-68.177-37.710-68.475-37.454Q-68.505-37.433-68.533-37.433Q-68.574-37.433-68.615-37.474Q-68.656-37.515-68.656-37.560\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(272.147 35.93)\">\u003Cpath d=\"M-62.546-37.433L-64.176-37.433L-64.176-37.713Q-63.947-37.713-63.798-37.748Q-63.650-37.782-63.650-37.922L-63.650-41.268Q-63.650-41.439-63.786-41.480Q-63.923-41.521-64.176-41.521L-64.176-41.801L-63.096-41.876L-63.096-41.470Q-62.874-41.671-62.587-41.774Q-62.299-41.876-61.992-41.876Q-61.565-41.876-61.201-41.663Q-60.837-41.449-60.623-41.085Q-60.409-40.721-60.409-40.301Q-60.409-39.856-60.649-39.492Q-60.888-39.128-61.281-38.925Q-61.674-38.722-62.118-38.722Q-62.385-38.722-62.633-38.822Q-62.880-38.923-63.068-39.104L-63.068-37.922Q-63.068-37.785-62.920-37.749Q-62.771-37.713-62.546-37.713L-62.546-37.433M-63.068-41.121L-63.068-39.511Q-62.935-39.258-62.692-39.101Q-62.450-38.944-62.173-38.944Q-61.845-38.944-61.592-39.145Q-61.339-39.347-61.206-39.665Q-61.072-39.983-61.072-40.301Q-61.072-40.530-61.137-40.759Q-61.202-40.988-61.330-41.186Q-61.459-41.384-61.653-41.504Q-61.848-41.623-62.081-41.623Q-62.375-41.623-62.643-41.494Q-62.911-41.364-63.068-41.121M-59.715-39.518Q-59.715-39.850-59.492-40.077Q-59.268-40.304-58.924-40.432Q-58.581-40.561-58.208-40.613Q-57.836-40.666-57.531-40.666L-57.531-40.919Q-57.531-41.124-57.639-41.304Q-57.747-41.483-57.928-41.586Q-58.109-41.688-58.317-41.688Q-58.724-41.688-58.960-41.596Q-58.871-41.559-58.825-41.475Q-58.779-41.391-58.779-41.289Q-58.779-41.193-58.825-41.114Q-58.871-41.036-58.952-40.991Q-59.032-40.947-59.121-40.947Q-59.271-40.947-59.372-41.044Q-59.473-41.142-59.473-41.289Q-59.473-41.911-58.317-41.911Q-58.106-41.911-57.856-41.847Q-57.607-41.784-57.405-41.665Q-57.203-41.545-57.077-41.360Q-56.950-41.176-56.950-40.933L-56.950-39.357Q-56.950-39.241-56.889-39.145Q-56.827-39.050-56.714-39.050Q-56.605-39.050-56.540-39.144Q-56.475-39.238-56.475-39.357L-56.475-39.805L-56.209-39.805L-56.209-39.357Q-56.209-39.087-56.436-38.922Q-56.663-38.756-56.943-38.756Q-57.152-38.756-57.289-38.910Q-57.425-39.063-57.449-39.279Q-57.596-39.012-57.878-38.867Q-58.160-38.722-58.485-38.722Q-58.762-38.722-59.046-38.797Q-59.329-38.872-59.522-39.051Q-59.715-39.231-59.715-39.518M-59.100-39.518Q-59.100-39.344-58.999-39.214Q-58.899-39.084-58.743-39.014Q-58.587-38.944-58.423-38.944Q-58.205-38.944-57.996-39.041Q-57.788-39.139-57.660-39.320Q-57.531-39.501-57.531-39.727L-57.531-40.455Q-57.856-40.455-58.222-40.364Q-58.587-40.273-58.844-40.061Q-59.100-39.850-59.100-39.518M-55.265-39.631L-55.265-41.528L-55.904-41.528L-55.904-41.750Q-55.587-41.750-55.369-41.960Q-55.152-42.170-55.052-42.480Q-54.951-42.789-54.951-43.097L-54.684-43.097L-54.684-41.808L-53.608-41.808L-53.608-41.528L-54.684-41.528L-54.684-39.644Q-54.684-39.368-54.580-39.169Q-54.476-38.971-54.216-38.971Q-54.059-38.971-53.953-39.075Q-53.847-39.180-53.797-39.333Q-53.748-39.487-53.748-39.644L-53.748-40.058L-53.481-40.058L-53.481-39.631Q-53.481-39.405-53.580-39.195Q-53.679-38.985-53.864-38.853Q-54.048-38.722-54.277-38.722Q-54.715-38.722-54.990-38.959Q-55.265-39.197-55.265-39.631M-50.989-38.790L-52.623-38.790L-52.623-39.070Q-52.394-39.070-52.245-39.104Q-52.097-39.139-52.097-39.279L-52.097-42.898Q-52.097-43.168-52.204-43.230Q-52.312-43.291-52.623-43.291L-52.623-43.572L-51.543-43.647L-51.543-41.261Q-51.437-41.446-51.259-41.588Q-51.082-41.729-50.873-41.803Q-50.665-41.876-50.439-41.876Q-49.933-41.876-49.650-41.653Q-49.366-41.429-49.366-40.933L-49.366-39.279Q-49.366-39.142-49.217-39.106Q-49.068-39.070-48.843-39.070L-48.843-38.790L-50.473-38.790L-50.473-39.070Q-50.244-39.070-50.096-39.104Q-49.947-39.139-49.947-39.279L-49.947-40.919Q-49.947-41.254-50.067-41.454Q-50.186-41.654-50.501-41.654Q-50.771-41.654-51.005-41.518Q-51.239-41.381-51.377-41.147Q-51.516-40.913-51.516-40.639L-51.516-39.279Q-51.516-39.142-51.365-39.106Q-51.215-39.070-50.989-39.070\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(272.147 35.93)\">\u003Cpath d=\"M-45.082-38.944Q-45.082-38.992-45.075-39.016L-44.563-41.080Q-44.529-41.207-44.529-41.323Q-44.529-41.463-44.582-41.559Q-44.635-41.654-44.758-41.654Q-44.980-41.654-45.081-41.427Q-45.181-41.200-45.291-40.779Q-45.301-40.714-45.363-40.714L-45.472-40.714Q-45.503-40.714-45.527-40.745Q-45.551-40.776-45.551-40.800L-45.551-40.827Q-45.438-41.261-45.257-41.569Q-45.075-41.876-44.744-41.876Q-44.559-41.876-44.380-41.801Q-44.200-41.726-44.088-41.586Q-43.975-41.446-43.975-41.254Q-43.828-41.439-43.647-41.579Q-43.466-41.719-43.252-41.798Q-43.038-41.876-42.806-41.876Q-42.396-41.876-42.139-41.680Q-41.883-41.483-41.883-41.087Q-41.883-40.803-42.011-40.405Q-42.139-40.007-42.338-39.518Q-42.413-39.323-42.413-39.176Q-42.413-38.944-42.245-38.944Q-41.969-38.944-41.775-39.221Q-41.582-39.498-41.504-39.819Q-41.480-39.880-41.428-39.880L-41.316-39.880Q-41.282-39.880-41.259-39.851Q-41.237-39.822-41.237-39.798Q-41.237-39.785-41.244-39.771Q-41.305-39.521-41.447-39.279Q-41.589-39.036-41.798-38.879Q-42.006-38.722-42.259-38.722Q-42.543-38.722-42.744-38.884Q-42.946-39.046-42.946-39.316Q-42.946-39.433-42.898-39.566Q-42.427-40.745-42.427-41.183Q-42.427-41.391-42.522-41.523Q-42.618-41.654-42.820-41.654Q-43.575-41.654-44.101-40.639L-44.515-38.978Q-44.539-38.872-44.633-38.797Q-44.727-38.722-44.843-38.722Q-44.942-38.722-45.012-38.783Q-45.082-38.845-45.082-38.944\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003Cg transform=\"translate(272.147 35.93)\">\u003Cpath d=\"M-40.285-37.033L-40.354-37.033Q-40.388-37.033-40.410-37.059Q-40.432-37.084-40.432-37.119Q-40.432-37.163-40.401-37.180Q-40.046-37.484-39.796-37.874Q-39.547-38.264-39.395-38.696Q-39.243-39.128-39.173-39.597Q-39.103-40.065-39.103-40.540Q-39.103-41.019-39.173-41.485Q-39.243-41.952-39.397-42.387Q-39.550-42.823-39.802-43.211Q-40.053-43.599-40.401-43.893Q-40.432-43.910-40.432-43.955Q-40.432-43.989-40.410-44.014Q-40.388-44.040-40.354-44.040L-40.285-44.040Q-40.275-44.040-40.266-44.038Q-40.258-44.037-40.248-44.033Q-39.704-43.633-39.332-43.080Q-38.959-42.526-38.778-41.880Q-38.597-41.234-38.597-40.540Q-38.597-39.839-38.778-39.192Q-38.959-38.544-39.333-37.990Q-39.708-37.436-40.248-37.040Q-40.258-37.040-40.266-37.038Q-40.275-37.037-40.285-37.033\" fill=\"currentColor\" stroke=\"currentColor\" class=\"tikz-text\" style=\"stroke-width:0.210\"\u002F>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fg>\u003C\u002Fsvg>\u003Cfigcaption class=\"tikz-cap\">Connectivity, not just cost. Self-attention (left) links every token pair directly at path length \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:0.6444em;\">\u003C\u002Fspan>\u003Cspan class=\"mord\">1\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>; a recurrence (right) leaves distant tokens \u003Cspan class=\"katex\">\u003Cspan class=\"katex-html\" aria-hidden=\"true\">\u003Cspan class=\"base\">\u003Cspan class=\"strut\" style=\"height:1em;vertical-align:-0.25em;\">\u003C\u002Fspan>\u003Cspan class=\"mord mathnormal\" style=\"margin-right:0.0278em;\">O\u003C\u002Fspan>\u003Cspan class=\"mopen\">(\u003C\u002Fspan>\u003Cspan class=\"mord mathnormal\">n\u003C\u002Fspan>\u003Cspan class=\"mclose\">)\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan>\u003C\u002Fspan> hops apart.\u003C\u002Ffigcaption>",1785117769373]