{"id":"https://openalex.org/W2970674435","doi":null,"title":"MetaInit: Initializing learning by learning to initialize","display_name":"MetaInit: Initializing learning by learning to initialize","publication_year":2019,"publication_date":"2019-01-01","ids":{"openalex":"https://openalex.org/W2970674435","mag":"2970674435"},"language":"en","primary_location":{"id":"mag:2970674435","doi":null,"is_oa":false,"landing_page_url":"https://papers.nips.cc/paper/9427-metainit-initializing-learning-by-learning-to-initialize.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/A5020785659","display_name":"Yann Dauphin","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":"Yann Dauphin","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/A5070645440","display_name":"Samuel S. Schoenholz","orcid":"https://orcid.org/0000-0003-1909-9226","observed_orcids":["https://orcid.org/0000-0003-1909-9226"]},"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":"Samuel S. Schoenholz","raw_affiliation_strings":["Google (United States), Mountain View, United States"],"raw_orcid":"https://orcid.org/0000-0003-1909-9226","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":1.4086,"has_fulltext":false,"cited_by_count":25,"citation_normalized_percentile":{"value":0.84340895,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":97,"max":99},"biblio":{"volume":"32","issue":null,"first_page":"12624","last_page":"12636"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.6855000257492065,"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":0.6855000257492065,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.10459999740123749,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.09009999781847,"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/neural-network-initialization","display_name":"neural network initialization","score":0.7080000042915344},{"id":"https://openalex.org/keywords/meta-learning","display_name":"meta-learning","score":0.9169999957084656},{"id":"https://openalex.org/keywords/weight-initialization","display_name":"weight initialization","score":0.5220000147819519},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"gradient descent","score":0.48399999737739563},{"id":"https://openalex.org/keywords/second-order-effects","display_name":"second-order effects","score":0.6100000143051147},{"id":"https://openalex.org/keywords/resnet","display_name":"ResNet","score":0.7059999704360962},{"id":"https://openalex.org/keywords/skip-connections","display_name":"skip connections","score":0.628000020980835}],"concepts":[{"id":"https://openalex.org/C114466953","wikidata":"https://www.wikidata.org/wiki/Q6034165","display_name":"Initialization","level":2,"score":0.8718904852867126},{"id":"https://openalex.org/C136886441","wikidata":"https://www.wikidata.org/wiki/Q926129","display_name":"Normalization (sociology)","level":2,"score":0.8153624534606934},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7158883810043335},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.671682596206665},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.6713539958000183},{"id":"https://openalex.org/C2944601119","wikidata":"https://www.wikidata.org/wiki/Q43744058","display_name":"Residual neural network","level":3,"score":0.6158252358436584},{"id":"https://openalex.org/C155512373","wikidata":"https://www.wikidata.org/wiki/Q287450","display_name":"Residual","level":2,"score":0.566022515296936},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.5389142632484436},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.4665296673774719},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.464335560798645},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.453429639339447},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3118526041507721},{"id":"https://openalex.org/C144024400","wikidata":"https://www.wikidata.org/wiki/Q21201","display_name":"Sociology","level":0,"score":0.0},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.0},{"id":"https://openalex.org/C19165224","wikidata":"https://www.wikidata.org/wiki/Q23404","display_name":"Anthropology","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"mag:2970674435","doi":null,"is_oa":false,"landing_page_url":"https://papers.nips.cc/paper/9427-metainit-initializing-learning-by-learning-to-initialize.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":30,"referenced_works":["https://openalex.org/W6908809","https://openalex.org/W71081281","https://openalex.org/W1677182931","https://openalex.org/W1868018859","https://openalex.org/W1980287119","https://openalex.org/W2108598243","https://openalex.org/W2112796928","https://openalex.org/W2125930537","https://openalex.org/W2146502635","https://openalex.org/W2155894447","https://openalex.org/W2176412452","https://openalex.org/W2194775991","https://openalex.org/W2302255633","https://openalex.org/W2525778437","https://openalex.org/W2591954064","https://openalex.org/W2604763608","https://openalex.org/W2626778328","https://openalex.org/W2808197691","https://openalex.org/W2808460953","https://openalex.org/W2809090039","https://openalex.org/W2899771611","https://openalex.org/W2916158460","https://openalex.org/W2917041815","https://openalex.org/W2919115771","https://openalex.org/W2936995161","https://openalex.org/W2949117887","https://openalex.org/W2963208657","https://openalex.org/W2963570896","https://openalex.org/W2963685250","https://openalex.org/W2963970792"],"related_works":["https://openalex.org/W7109286769","https://openalex.org/W4283733629","https://openalex.org/W2783494467","https://openalex.org/W6990681061","https://openalex.org/W4288335814","https://openalex.org/W3207971231","https://openalex.org/W3035543253","https://openalex.org/W2755231872","https://openalex.org/W3130602232","https://openalex.org/W7165241369"],"abstract_inverted_index":{"Deep":[0],"learning":[1],"models":[2,194],"frequently":[3],"trade":[4],"handcrafted":[5],"features":[6,9],"for":[7,60,182],"deep":[8,26],"learned":[10],"with":[11,87,109,164],"much":[12],"less":[13],"human":[14],"intervention":[15],"using":[16,63,120],"gradient":[17,76,103,121],"descent.":[18],"While":[19],"this":[20,45,94,116,178],"paradigm":[21],"has":[22],"been":[23],"enormously":[24],"successful,":[25],"networks":[27,139,159,183],"are":[28],"often":[29],"difficult":[30],"to":[31,123,157,192],"train":[32,158],"and":[33,137,140,160,189],"performance":[34,162],"can":[35,106,145,190],"depend":[36],"crucially":[37],"on":[38,69,135,187,195],"the":[39,58,102,125,128,143,165],"initial":[40,129],"choice":[41],"of":[42,127,151],"parameters.":[43],"In":[44,173],"work,":[46],"we":[47,100,175],"introduce":[48],"an":[49],"algorithm":[50,144],"called":[51],"MetaInit":[52,114,154],"as":[53],"a":[54,70,97,149],"step":[55],"towards":[56],"automating":[57],"search":[59],"good":[61,73],"initializations":[62,74],"meta-learning.":[64],"Our":[65],"approach":[66,179],"is":[67],"based":[68],"hypothesis":[71],"that":[72,83,99,142,177],"make":[75],"descent":[77,122],"easier":[78],"by":[79,119],"starting":[80],"in":[81],"regions":[82],"look":[84],"locally":[85],"linear":[86],"minimal":[88],"second":[89],"order":[90],"effects.":[91],"We":[92,132],"formalize":[93],"notion":[95],"via":[96],"quantity":[98,117],"call":[101],"quotient,":[104],"which":[105],"be":[107],"computed":[108],"any":[110],"architecture":[111],"or":[112,170],"dataset.":[113],"minimizes":[115],"efficiently":[118],"tune":[124],"norms":[126],"weight":[130],"matrices.":[131],"conduct":[133],"experiments":[134],"plain":[136],"residual":[138,171],"show":[141],"automatically":[146],"recover":[147],"from":[148],"class":[150],"bad":[152],"initializations.":[153],"allows":[155],"us":[156],"achieve":[161],"competitive":[163],"state-of-the-art":[166],"without":[167,184],"batch":[168],"normalization":[169,181],"connections.":[172],"particular,":[174],"find":[176],"outperforms":[180],"skip":[185],"connections":[186],"CIFAR-10":[188],"scale":[191],"Resnet-50":[193],"Imagenet.":[196]},"counts_by_year":[{"year":2024,"cited_by_count":6},{"year":2023,"cited_by_count":5},{"year":2021,"cited_by_count":6},{"year":2020,"cited_by_count":8}],"updated_date":"2026-11-04T10:54:05.112534","created_date":"2019-09-05T00:00:00"}