{"id":"https://openalex.org/W4404468869","doi":"https://doi.org/10.1162/neco_a_01725","title":"Generalization Guarantees of Gradient Descent for Shallow Neural Networks","display_name":"Generalization Guarantees of Gradient Descent for Shallow Neural Networks","publication_year":2024,"publication_date":"2024-11-18","ids":{"openalex":"https://openalex.org/W4404468869","doi":"https://doi.org/10.1162/neco_a_01725","pmid":"https://pubmed.ncbi.nlm.nih.gov/39556516"},"language":"en","primary_location":{"id":"doi:10.1162/neco_a_01725","is_oa":false,"landing_page_url":"https://doi.org/10.1162/neco_a_01725","pdf_url":null,"source":{"id":"https://openalex.org/S207023548","display_name":"Neural Computation","issn_l":"0899-7667","issn":["0899-7667","1530-888X"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310315718","host_organization_name":"The MIT Press","host_organization_lineage":["https://openalex.org/P4310315718"],"host_organization_lineage_names":["The MIT Press"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Neural Computation","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","pubmed"],"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/A5103081079","display_name":"Puyu Wang","orcid":"https://orcid.org/0000-0003-3104-3081"},"institutions":[{"id":"https://openalex.org/I141568987","display_name":"Hong Kong Baptist University","ror":"https://ror.org/0145fw131","country_code":"HK","type":"education","lineage":["https://openalex.org/I141568987"]}],"countries":["HK"],"is_corresponding":true,"raw_author_name":"Puyu Wang","raw_affiliation_strings":["Hong Kong Baptist University, Hong Kong wangpuyu1026@gmail.com"],"affiliations":[{"raw_affiliation_string":"Hong Kong Baptist University, Hong Kong wangpuyu1026@gmail.com","institution_ids":["https://openalex.org/I141568987"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5046468616","display_name":"Yunwen Lei","orcid":"https://orcid.org/0000-0002-5383-467X"},"institutions":[{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Yunwen Lei","raw_affiliation_strings":["University of Hong Kong, Hong Kong leiyw@hku.hk"],"affiliations":[{"raw_affiliation_string":"University of Hong Kong, Hong Kong leiyw@hku.hk","institution_ids":["https://openalex.org/I889458895"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100401482","display_name":"Di Wang","orcid":"https://orcid.org/0000-0003-4908-0243"},"institutions":[{"id":"https://openalex.org/I71920554","display_name":"King Abdullah University of Science and Technology","ror":"https://ror.org/01q3tbs38","country_code":"SA","type":"education","lineage":["https://openalex.org/I71920554"]}],"countries":["SA"],"is_corresponding":false,"raw_author_name":"Di Wang","raw_affiliation_strings":["King Abdullah University of Science and Technology 23955, Saudi Arabia di.wang@kaust.edu.sa"],"affiliations":[{"raw_affiliation_string":"King Abdullah University of Science and Technology 23955, Saudi Arabia di.wang@kaust.edu.sa","institution_ids":["https://openalex.org/I71920554"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048960543","display_name":"Yiming Ying","orcid":null},"institutions":[{"id":"https://openalex.org/I129604602","display_name":"University of Sydney","ror":"https://ror.org/0384j8v12","country_code":"AU","type":"education","lineage":["https://openalex.org/I129604602"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Yiming Ying","raw_affiliation_strings":["University of Sydney, Camperdown, NSW 2050, Australia yiming.ying@sydney.edu.au"],"affiliations":[{"raw_affiliation_string":"University of Sydney, Camperdown, NSW 2050, Australia yiming.ying@sydney.edu.au","institution_ids":["https://openalex.org/I129604602"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5006802589","display_name":"Ding\u2010Xuan Zhou","orcid":"https://orcid.org/0000-0003-0224-9216"},"institutions":[{"id":"https://openalex.org/I129604602","display_name":"University of Sydney","ror":"https://ror.org/0384j8v12","country_code":"AU","type":"education","lineage":["https://openalex.org/I129604602"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Ding-Xuan Zhou","raw_affiliation_strings":["University of Sydney, Camperdown, NSW 2050, Australia dingxuan.zhou@sydney.edu.au"],"affiliations":[{"raw_affiliation_string":"University of Sydney, Camperdown, NSW 2050, Australia dingxuan.zhou@sydney.edu.au","institution_ids":["https://openalex.org/I129604602"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5103081079"],"corresponding_institution_ids":["https://openalex.org/I141568987"],"apc_list":null,"apc_paid":null,"fwci":0.7252,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.77283317,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":95,"max":96},"biblio":{"volume":"37","issue":"2","first_page":"344","last_page":"402"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.9976000189781189,"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/T10320","display_name":"Neural Networks and Applications","score":0.9976000189781189,"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/T12676","display_name":"Machine Learning and ELM","score":0.9958999752998352,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.9955000281333923,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/normalization","display_name":"Normalization (sociology)","score":0.7815358638763428},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.7734723091125488},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.7335694432258606},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.7245843410491943},{"id":"https://openalex.org/keywords/gradient-descent","display_name":"Gradient descent","score":0.6874917149543762},{"id":"https://openalex.org/keywords/layer","display_name":"Layer (electronics)","score":0.5591282844543457},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.536827027797699},{"id":"https://openalex.org/keywords/backpropagation","display_name":"Backpropagation","score":0.5104817152023315},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.4521104395389557},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.4388166666030884},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.3965952396392822},{"id":"https://openalex.org/keywords/applied-mathematics","display_name":"Applied mathematics","score":0.3924207389354706},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3542190194129944},{"id":"https://openalex.org/keywords/mathematical-analysis","display_name":"Mathematical analysis","score":0.17471396923065186},{"id":"https://openalex.org/keywords/geometry","display_name":"Geometry","score":0.14340978860855103},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.11935779452323914},{"id":"https://openalex.org/keywords/materials-science","display_name":"Materials science","score":0.10248076915740967},{"id":"https://openalex.org/keywords/nanotechnology","display_name":"Nanotechnology","score":0.06937989592552185}],"concepts":[{"id":"https://openalex.org/C136886441","wikidata":"https://www.wikidata.org/wiki/Q926129","display_name":"Normalization (sociology)","level":2,"score":0.7815358638763428},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.7734723091125488},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.7335694432258606},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.7245843410491943},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.6874917149543762},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.5591282844543457},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.536827027797699},{"id":"https://openalex.org/C155032097","wikidata":"https://www.wikidata.org/wiki/Q798503","display_name":"Backpropagation","level":3,"score":0.5104817152023315},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4521104395389557},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.4388166666030884},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3965952396392822},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.3924207389354706},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3542190194129944},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.17471396923065186},{"id":"https://openalex.org/C2524010","wikidata":"https://www.wikidata.org/wiki/Q8087","display_name":"Geometry","level":1,"score":0.14340978860855103},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.11935779452323914},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.10248076915740967},{"id":"https://openalex.org/C171250308","wikidata":"https://www.wikidata.org/wiki/Q11468","display_name":"Nanotechnology","level":1,"score":0.06937989592552185},{"id":"https://openalex.org/C19165224","wikidata":"https://www.wikidata.org/wiki/Q23404","display_name":"Anthropology","level":1,"score":0.0},{"id":"https://openalex.org/C144024400","wikidata":"https://www.wikidata.org/wiki/Q21201","display_name":"Sociology","level":0,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1162/neco_a_01725","is_oa":false,"landing_page_url":"https://doi.org/10.1162/neco_a_01725","pdf_url":null,"source":{"id":"https://openalex.org/S207023548","display_name":"Neural Computation","issn_l":"0899-7667","issn":["0899-7667","1530-888X"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310315718","host_organization_name":"The MIT Press","host_organization_lineage":["https://openalex.org/P4310315718"],"host_organization_lineage_names":["The MIT Press"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Neural Computation","raw_type":"journal-article"},{"id":"pmid:39556516","is_oa":false,"landing_page_url":"https://pubmed.ncbi.nlm.nih.gov/39556516","pdf_url":null,"source":{"id":"https://openalex.org/S4306525036","display_name":"PubMed","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1299303238","host_organization_name":"National Institutes of Health","host_organization_lineage":["https://openalex.org/I1299303238"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Neural computation","raw_type":null},{"id":"pmh:oai:hub.hku.hk:10722/355113","is_oa":false,"landing_page_url":"https://hub.hku.hk/handle/10722/355113","pdf_url":null,"source":{"id":"https://openalex.org/S4377196271","display_name":"The HKU Scholars Hub (University of Hong Kong)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I889458895","host_organization_name":"University of Hong Kong","host_organization_lineage":["https://openalex.org/I889458895"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":36,"referenced_works":["https://openalex.org/W2160815625","https://openalex.org/W2257979135","https://openalex.org/W2566079294","https://openalex.org/W2618530766","https://openalex.org/W2963095610","https://openalex.org/W3191067499","https://openalex.org/W3203457432","https://openalex.org/W6637414576","https://openalex.org/W6638214083","https://openalex.org/W6680741563","https://openalex.org/W6735717691","https://openalex.org/W6739659843","https://openalex.org/W6741472820","https://openalex.org/W6741653254","https://openalex.org/W6744989570","https://openalex.org/W6745375793","https://openalex.org/W6747520023","https://openalex.org/W6752009368","https://openalex.org/W6752495264","https://openalex.org/W6753918066","https://openalex.org/W6755150206","https://openalex.org/W6756001544","https://openalex.org/W6756091659","https://openalex.org/W6756455746","https://openalex.org/W6758728975","https://openalex.org/W6758902322","https://openalex.org/W6761496057","https://openalex.org/W6763368363","https://openalex.org/W6764442487","https://openalex.org/W6778993914","https://openalex.org/W6788820006","https://openalex.org/W6789878931","https://openalex.org/W6838769558","https://openalex.org/W6842735619","https://openalex.org/W6850617383","https://openalex.org/W7034108470"],"related_works":["https://openalex.org/W2115605526","https://openalex.org/W3093883775","https://openalex.org/W1539246760","https://openalex.org/W2786746258","https://openalex.org/W4225893763","https://openalex.org/W2788727425","https://openalex.org/W2405196115","https://openalex.org/W2104893957","https://openalex.org/W4402471162","https://openalex.org/W2971074373"],"abstract_inverted_index":{"Significant":[0],"progress":[1],"has":[2,30,36],"been":[3],"made":[4],"recently":[5],"in":[6,111,148,219],"understanding":[7],"the":[8,20,27,39,50,53,61,102,126,140,172,183,188,200],"generalization":[9,78],"of":[10,26,41,52,80,104,128,134,144,176,215],"neural":[11],"networks":[12],"(NNs)":[13],"trained":[14,167],"by":[15,72,117,168],"gradient":[16],"descent":[17],"(GD)":[18],"using":[19],"algorithmic":[21],"stability":[22,76],"approach.":[23],"However,":[24],"most":[25],"existing":[28],"research":[29],"focused":[31],"on":[32,157],"one-hidden-layer":[33],"NNs":[34,166],"and":[35,77,84,151,164,222],"not":[37],"addressed":[38],"impact":[40],"different":[42],"network":[43,46,96,189],"scaling.":[44],"Here,":[45],"scaling":[47,184],"corresponds":[48],"to":[49,170,198],"normalization":[51],"layers.":[54],"In":[55,101],"this":[56],"article,":[57],"we":[58,138,179,209],"greatly":[59],"extend":[60],"previous":[62,99],"work":[63],"(Lei":[64],"et":[65],"al.,":[66],"2022;":[67],"Richards":[68],"&":[69],"Kuzborskij,":[70],"2021)":[71],"conducting":[73],"a":[74,119,131,206,211],"comprehensive":[75],"analysis":[79],"GD":[81,147,169,197,218],"for":[82,146,162,196,217],"two-layer":[83,88,150,221],"three-layer":[85,105,152,223],"NNs.":[86,153,224],"For":[87],"NNs,":[89,106],"our":[90,107,135],"results":[91],"are":[92],"established":[93],"under":[94,205],"general":[95,136],"scaling,":[97],"relaxing":[98],"conditions.":[100],"case":[103],"technical":[108],"contribution":[109],"lies":[110],"demonstrating":[112],"its":[113],"nearly":[114],"co-coercive":[115],"property":[116],"utilizing":[118],"novel":[120],"induction":[121],"strategy":[122],"that":[123,181],"thoroughly":[124],"explores":[125],"effects":[127],"overparameterization.":[129],"As":[130],"direct":[132],"application":[133],"findings,":[137],"derive":[139],"excess":[141],"risk":[142,174,213],"rate":[143,175,214],"O(1/n)":[145,216],"both":[149,220],"This":[154],"sheds":[155],"light":[156],"sufficient":[158],"or":[159,187],"necessary":[160],"conditions":[161],"underparameterized":[163],"overparameterized":[165],"attain":[171],"desired":[173,201],"O(1/n).":[177],"Moreover,":[178],"demonstrate":[180],"as":[182],"factor":[185],"increases":[186],"complexity":[190],"decreases,":[191],"less":[192],"overparameterization":[193],"is":[194],"required":[195],"achieve":[199],"error":[202],"rates.":[203],"Additionally,":[204],"low-noise":[207],"condition,":[208],"obtain":[210],"fast":[212]},"counts_by_year":[{"year":2025,"cited_by_count":2}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}