{"id":"https://openalex.org/W2946840143","doi":"https://doi.org/10.48550/arxiv.1905.13210","title":"Generalization Bounds of Stochastic Gradient Descent for Wide and Deep Neural Networks","display_name":"Generalization Bounds of Stochastic Gradient Descent for Wide and Deep Neural Networks","alternate_titles":[],"publication_year":2019,"publication_date":"2019-05-30","ids":{"openalex":"https://openalex.org/W2946840143","doi":"https://doi.org/10.48550/arxiv.1905.13210","mag":"2946840143"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:1905.13210","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/1905.13210","pdf_url":"https://arxiv.org/pdf/1905.13210","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"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":"text"},"parent_works":[],"lineage":["https://openalex.org/W2946840143"],"child_works_count":0,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/1905.13210","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5101989491","display_name":"Yuan Jie Cao","orcid":"https://orcid.org/0000-0002-2495-4701","observed_orcids":["https://orcid.org/0000-0002-2495-4701"]},"institutions":[{"id":"https://openalex.org/I161318765","display_name":"University of California, Los Angeles","ror":"https://ror.org/046rm7j60","country_code":"US","type":"education","lineage":["https://openalex.org/I161318765"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Cao, Yuan","raw_affiliation_strings":["Department of Computer Science , University of California , Los Angeles , CA 90095 , USA;","University of California-Los Angeles"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science , University of California , Los Angeles , CA 90095 , USA;","institution_ids":["https://openalex.org/I161318765"]},{"raw_affiliation_string":"University of California-Los Angeles","institution_ids":["https://openalex.org/I161318765"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5051448391","display_name":"Quanquan Gu","orcid":"https://orcid.org/0000-0001-9830-793X","observed_orcids":["https://orcid.org/0000-0001-9830-793X"]},"institutions":[{"id":"https://openalex.org/I161318765","display_name":"University of California, Los Angeles","ror":"https://ror.org/046rm7j60","country_code":"US","type":"education","lineage":["https://openalex.org/I161318765"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Gu, Quanquan","raw_affiliation_strings":["Department of Computer Science , University of California , Los Angeles , CA 90095 , USA;","University of California-Los Angeles"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science , University of California , Los Angeles , CA 90095 , USA;","institution_ids":["https://openalex.org/I161318765"]},{"raw_affiliation_string":"University of California-Los Angeles","institution_ids":["https://openalex.org/I161318765"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I161318765"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":419,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":"32","issue":null,"first_page":"10835","last_page":"10845"},"is_retracted":false,"is_pulled":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9872999787330627,"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":0.9872999787330627,"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":0.007899999618530273,"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":0.0015999999595806003,"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/generalization-bounds","display_name":"generalization bounds","score":0.9269999861717224},{"id":"https://openalex.org/keywords/stochastic-gradient-descent","display_name":"stochastic gradient descent","score":0.9860000014305115},{"id":"https://openalex.org/keywords/overparameterized-neural-networks","display_name":"overparameterized neural networks","score":0.9350000023841858},{"id":"https://openalex.org/keywords/wide-neural-networks","display_name":"wide neural networks","score":0.8370000123977661},{"id":"https://openalex.org/keywords/relu-neural-networks","display_name":"ReLU networks","score":0.8669999837875366},{"id":"https://openalex.org/keywords/neural-tangent-kernel","display_name":"neural tangent kernel","score":0.8930000066757202},{"id":"https://openalex.org/keywords/random-feature-models","display_name":"random feature models","score":0.6769999861717224},{"id":"https://openalex.org/keywords/0-1-loss","display_name":"0-1 loss","score":0.8769999742507935},{"id":"https://openalex.org/keywords/random-initialization","display_name":"random initialization","score":0.7139999866485596},{"id":"https://openalex.org/keywords/training-loss","display_name":"training loss","score":0.40299999713897705}],"concepts":[{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.7997664213180542},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.6138373613357544},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.5786486268043518},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.5773792266845703},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.524183452129364},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.5199097990989685},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.47538435459136963},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2520921230316162},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.05693089962005615}],"mesh":[],"locations_count":3,"locations":[{"id":"pmh:oai:arXiv.org:1905.13210","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/1905.13210","pdf_url":"https://arxiv.org/pdf/1905.13210","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"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":"text"},{"id":"doi:10.48550/arxiv.1905.13210","doi":"https://doi.org/10.48550/arxiv.1905.13210","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.1905.13210","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},{"id":"mag:2970332347","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/pdf/1905.13210.pdf","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":"arXiv (Cornell University)","raw_type":null}],"best_oa_location":{"id":"pmh:oai:arXiv.org:1905.13210","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/1905.13210","pdf_url":"https://arxiv.org/pdf/1905.13210","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"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":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G5325791252","display_name":"III: Small: Collaborative Research: High-Dimensional Machine Learning Methods for Personalized Cancer Genomics","funder_award_id":"1903202","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G6024092204","display_name":"CAREER:  Scaling Up Knowledge Discovery in High-Dimensional Data Via Nonconvex Statistical Optimization","funder_award_id":"1906169","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W2946840143.pdf","grobid_xml":"https://content.openalex.org/works/W2946840143.grobid-xml"},"referenced_works_count":43,"referenced_works":["https://openalex.org/W1677182931","https://openalex.org/W1720114023","https://openalex.org/W2109339818","https://openalex.org/W2112796928","https://openalex.org/W2115065944","https://openalex.org/W2123395972","https://openalex.org/W2144902422","https://openalex.org/W2160815625","https://openalex.org/W2163605009","https://openalex.org/W2257979135","https://openalex.org/W2593958421","https://openalex.org/W2604117713","https://openalex.org/W2709553318","https://openalex.org/W2741952635","https://openalex.org/W2765861987","https://openalex.org/W2777138330","https://openalex.org/W2785749378","https://openalex.org/W2807299122","https://openalex.org/W2808157231","https://openalex.org/W2809090039","https://openalex.org/W2886067286","https://openalex.org/W2894604724","https://openalex.org/W2899748887","https://openalex.org/W2900103278","https://openalex.org/W2900959181","https://openalex.org/W2911867426","https://openalex.org/W2912260645","https://openalex.org/W2912322140","https://openalex.org/W2913473169","https://openalex.org/W2926417365","https://openalex.org/W2938647293","https://openalex.org/W2942052807","https://openalex.org/W2947989684","https://openalex.org/W2950220847","https://openalex.org/W2962698540","https://openalex.org/W2970176397","https://openalex.org/W2996822312","https://openalex.org/W3009072036","https://openalex.org/W3021189130","https://openalex.org/W3147600416","https://openalex.org/W3153737196","https://openalex.org/W4236362309","https://openalex.org/W4299649195"],"related_works":["https://openalex.org/W2927724204","https://openalex.org/W2996822312","https://openalex.org/W4281633082","https://openalex.org/W3092390915","https://openalex.org/W7163906927","https://openalex.org/W4319987045","https://openalex.org/W3199714525","https://openalex.org/W7163898681","https://openalex.org/W4400601118","https://openalex.org/W4226215924"],"abstract_inverted_index":{"We":[0,36],"study":[1],"the":[2,12,16,30,39,62,72,109,117,150],"training":[3,33,63],"and":[4,55,125,149],"generalization":[5,105,130,146],"of":[6,21,32,43,65,111,116],"deep":[7],"neural":[8,81,135,151],"networks":[9],"(DNNs)":[10],"in":[11,108,156],"over-parameterized":[13,134],"regime,":[14],"where":[15],"network":[17,48,73,118],"width":[18],"(i.e.,":[19],"number":[20,31],"hidden":[22],"nodes":[23],"per":[24],"layer)":[25],"is":[26,114,122],"much":[27],"larger":[28],"than":[29,127],"data":[34,88],"points.":[35],"show":[37],"that,":[38],"expected":[40],"$0$-$1$":[41],"loss":[42,64],"a":[44,66,80,104,141],"wide":[45],"enough":[46],"ReLU":[47],"trained":[49],"with":[50,97],"stochastic":[51],"gradient":[52,74],"descent":[53],"(SGD)":[54],"random":[56,67,83],"initialization":[57],"can":[58,91],"be":[59,92],"bounded":[60],"by":[61,71,94],"feature":[68,84],"model":[69,96],"induced":[70],"at":[75],"initialization,":[76],"which":[77],"we":[78,139],"call":[79],"tangent":[82,152],"(NTRF)":[85],"model.":[86],"For":[87],"distributions":[89],"that":[90,113],"classified":[93],"NTRF":[95],"sufficiently":[98],"small":[99],"error,":[100],"our":[101,145],"result":[102,121],"yields":[103],"error":[106,131,147],"bound":[107,148],"order":[110],"$\\tilde{\\mathcal{O}}(n^{-1/2})$":[112],"independent":[115],"width.":[119],"Our":[120],"more":[123],"general":[124],"sharper":[126],"many":[128],"existing":[129],"bounds":[132],"for":[133],"networks.":[136],"In":[137],"addition,":[138],"establish":[140],"strong":[142],"connection":[143],"between":[144],"kernel":[153],"(NTK)":[154],"proposed":[155],"recent":[157],"work.":[158]},"cited_by_api_url":"https://api.openalex.org/works?filter=cites:W2946840143","counts_by_year":[{"year":2026,"cited_by_count":40},{"year":2025,"cited_by_count":46},{"year":2024,"cited_by_count":56},{"year":2023,"cited_by_count":59},{"year":2022,"cited_by_count":72},{"year":2021,"cited_by_count":67},{"year":2020,"cited_by_count":60},{"year":2019,"cited_by_count":19}],"updated_date":"2026-11-20T13:11:24.684918","created_date":"2019-06-07T00:00:00"}
