{"id":"https://openalex.org/W2889737445","doi":null,"title":"The Spectrum of the Fisher Information Matrix of a Single-Hidden-Layer Neural Network","display_name":"The Spectrum of the Fisher Information Matrix of a Single-Hidden-Layer Neural Network","publication_year":2018,"publication_date":"2018-01-01","ids":{"openalex":"https://openalex.org/W2889737445","mag":"2889737445"},"language":"en","primary_location":{"id":"mag:2889737445","doi":null,"is_oa":false,"landing_page_url":"https://papers.nips.cc/paper/2018/file/18bb68e2b38e4a8ce7cf4f6b2625768c-Paper.pdf","pdf_url":null,"source":{"id":"https://openalex.org/S4363606243","display_name":"neural information processing systems","issn_l":"1049-5258","issn":["1049-5258"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core","norway-2"],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":"Neural Information Processing Systems","raw_type":null},"volume":null,"type":"article","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5010225522","display_name":"Jeffrey Pennington","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jeffrey Pennington","raw_affiliation_strings":["Google (United States), Mountain View, United States"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google (United States), Mountain View, United States","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5049287396","display_name":"Pratik Worah","orcid":"https://orcid.org/0000-0002-2739-1246","observed_orcids":["https://orcid.org/0000-0002-2739-1246"]},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Pratik Worah","raw_affiliation_strings":["Google (United States), Mountain View, United States"],"raw_orcid":"https://orcid.org/0000-0002-2739-1246","affiliations":[{"raw_affiliation_string":"Google (United States), Mountain View, United States","institution_ids":["https://openalex.org/I1291425158"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I1291425158"],"apc_list":null,"apc_paid":null,"fwci":4.5981,"has_fulltext":false,"cited_by_count":48,"citation_normalized_percentile":{"value":0.95700456,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":88,"max":99},"biblio":{"volume":"31","issue":null,"first_page":"5410","last_page":"5419"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.8628000020980835,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.8628000020980835,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.1307000070810318,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12814","display_name":"Gaussian Processes and Bayesian Inference","score":0.0010000000474974513,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/fisher-information-matrix","display_name":"Fisher information matrix","score":0.9919999837875366},{"id":"https://openalex.org/keywords/eigenvalue-spectrum","display_name":"eigenvalue spectrum","score":0.7160000205039978},{"id":"https://openalex.org/keywords/single-hidden-layer-feedforward-neural-network","display_name":"single hidden layer neural network","score":0.9490000009536743},{"id":"https://openalex.org/keywords/infinite-width-limit","display_name":"infinite-width limit","score":0.796999990940094},{"id":"https://openalex.org/keywords/first-order-methods","display_name":"first-order methods","score":0.7250000238418579},{"id":"https://openalex.org/keywords/stochastic-gradient-descent","display_name":"stochastic gradient descent","score":0.7699999809265137},{"id":"https://openalex.org/keywords/gaussian-data","display_name":"Gaussian data","score":0.5860000252723694}],"concepts":[{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.6651301980018616},{"id":"https://openalex.org/C8642999","wikidata":"https://www.wikidata.org/wiki/Q4171168","display_name":"Hyperparameter","level":2,"score":0.6573060154914856},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.6476473808288574},{"id":"https://openalex.org/C158693339","wikidata":"https://www.wikidata.org/wiki/Q190524","display_name":"Eigenvalues and eigenvectors","level":2,"score":0.5945080518722534},{"id":"https://openalex.org/C158622935","wikidata":"https://www.wikidata.org/wiki/Q660848","display_name":"Nonlinear system","level":2,"score":0.5607859492301941},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5500397682189941},{"id":"https://openalex.org/C29406490","wikidata":"https://www.wikidata.org/wiki/Q1420659","display_name":"Fisher information","level":2,"score":0.5410820841789246},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.5254232883453369},{"id":"https://openalex.org/C163716315","wikidata":"https://www.wikidata.org/wiki/Q901177","display_name":"Gaussian","level":2,"score":0.5219050049781799},{"id":"https://openalex.org/C195065555","wikidata":"https://www.wikidata.org/wiki/Q214881","display_name":"Curvature","level":2,"score":0.4879305064678192},{"id":"https://openalex.org/C86582703","wikidata":"https://www.wikidata.org/wiki/Q7617824","display_name":"Stochastic neural network","level":4,"score":0.45883244276046753},{"id":"https://openalex.org/C106487976","wikidata":"https://www.wikidata.org/wiki/Q685816","display_name":"Matrix (chemical analysis)","level":2,"score":0.4223410487174988},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.4205879867076874},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4165147840976715},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.41312795877456665},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.4076055884361267},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.38029831647872925},{"id":"https://openalex.org/C184720557","wikidata":"https://www.wikidata.org/wiki/Q7825049","display_name":"Topology (electrical circuits)","level":2,"score":0.3512193560600281},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3448566198348999},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.20322191715240479},{"id":"https://openalex.org/C147168706","wikidata":"https://www.wikidata.org/wiki/Q1457734","display_name":"Recurrent neural network","level":3,"score":0.1548786461353302},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.10469228029251099},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.1007651686668396},{"id":"https://openalex.org/C159985019","wikidata":"https://www.wikidata.org/wiki/Q181790","display_name":"Composite material","level":1,"score":0.0},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.0},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","level":1,"score":0.0},{"id":"https://openalex.org/C114614502","wikidata":"https://www.wikidata.org/wiki/Q76592","display_name":"Combinatorics","level":1,"score":0.0},{"id":"https://openalex.org/C120665830","wikidata":"https://www.wikidata.org/wiki/Q14620","display_name":"Optics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"mag:2889737445","doi":null,"is_oa":false,"landing_page_url":"https://papers.nips.cc/paper/2018/file/18bb68e2b38e4a8ce7cf4f6b2625768c-Paper.pdf","pdf_url":null,"source":{"id":"https://openalex.org/S4363606243","display_name":"neural information processing systems","issn_l":"1049-5258","issn":["1049-5258"],"is_oa":false,"is_in_doaj":false,"is_core":true,"listed_in":["cwts-core","norway-2"],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":"Neural Information Processing Systems","raw_type":null}],"best_oa_location":null,"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":7,"referenced_works":["https://openalex.org/W1969810729","https://openalex.org/W2035311609","https://openalex.org/W2092997395","https://openalex.org/W2160815625","https://openalex.org/W2163605009","https://openalex.org/W2525778437","https://openalex.org/W2752851182"],"related_works":["https://openalex.org/W3034876183","https://openalex.org/W2979581378","https://openalex.org/W2912713668","https://openalex.org/W4289713189","https://openalex.org/W6929594390","https://openalex.org/W2963470399","https://openalex.org/W4283814645","https://openalex.org/W7065486322","https://openalex.org/W4226174562","https://openalex.org/W7168484903"],"abstract_inverted_index":{"An":[0],"important":[1,78],"factor":[2],"contributing":[3],"to":[4,15,75,152],"the":[5,12,30,38,42,83,86,108,111,143,145,154],"success":[6],"of":[7,32,41,71,80,85,113,156],"deep":[8],"learning":[9],"has":[10],"been":[11],"remarkable":[13],"ability":[14],"optimize":[16],"large":[17],"neural":[18,95],"networks":[19,120,126,130],"using":[20],"simple":[21],"first-order":[22,157],"optimization":[23,158],"algorithms":[24],"like":[25],"stochastic":[26],"gradient":[27],"descent.":[28],"While":[29],"efficiency":[31,155],"such":[33],"methods":[34],"depends":[35,54],"crucially":[36],"on":[37,55,92],"local":[39],"curvature":[40],"loss":[43],"surface,":[44],"very":[45],"little":[46],"is":[47],"actually":[48],"known":[49],"about":[50],"how":[51],"this":[52,61],"geometry":[53],"network":[56,96],"architecture":[57],"and":[58,100,102,127,137],"hyperparameters.":[59],"In":[60],"work,":[62],"we":[63],"extend":[64],"a":[65,93],"recently-developed":[66],"framework":[67],"for":[68,107],"studying":[69],"spectra":[70],"nonlinear":[72,125,129],"random":[73],"matrices":[74],"characterize":[76],"an":[77,104],"measure":[79],"curvature,":[81],"namely":[82],"eigenvalues":[84],"Fisher":[87],"information":[88],"matrix.":[89],"We":[90,116,134],"focus":[91],"single-hidden-layer":[94],"with":[97],"Gaussian":[98],"data":[99],"weights":[101],"provide":[103],"exact":[105],"expression":[106],"spectrum":[109,146],"in":[110],"limit":[112],"infinite":[114],"width.":[115],"find":[117],"that":[118,128,140],"linear":[119],"suffer":[121],"worse":[122],"conditioning":[123],"than":[124],"are":[131],"generically":[132],"non-degenerate.":[133],"also":[135],"predict":[136],"demonstrate":[138],"empirically":[139],"by":[141],"adjusting":[142],"nonlinearity,":[144],"can":[147],"be":[148],"tuned":[149],"so":[150],"as":[151],"improve":[153],"methods.":[159]},"counts_by_year":[{"year":2025,"cited_by_count":2},{"year":2024,"cited_by_count":1},{"year":2023,"cited_by_count":3},{"year":2022,"cited_by_count":4},{"year":2021,"cited_by_count":17},{"year":2020,"cited_by_count":9},{"year":2019,"cited_by_count":11},{"year":2018,"cited_by_count":1}],"updated_date":"2026-11-04T10:54:05.112534","created_date":"2018-09-27T00:00:00"}