{"id":"https://openalex.org/W2885208219","doi":"https://doi.org/10.1109/tsp.2019.2904921","title":"Learning ReLU Networks on Linearly Separable Data: Algorithm, Optimality, and Generalization","display_name":"Learning ReLU Networks on Linearly Separable Data: Algorithm, Optimality, and Generalization","publication_year":2019,"publication_date":"2019-03-20","ids":{"openalex":"https://openalex.org/W2885208219","doi":"https://doi.org/10.1109/tsp.2019.2904921","mag":"2885208219"},"language":"en","primary_location":{"id":"doi:10.1109/tsp.2019.2904921","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tsp.2019.2904921","pdf_url":null,"source":{"id":"https://openalex.org/S168680287","display_name":"IEEE Transactions on Signal Processing","issn_l":"1053-587X","issn":["1053-587X","1941-0476"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Signal Processing","raw_type":"journal-article"},"type":"article","indexed_in":["arxiv","crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/1808.04685","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100367403","display_name":"Gang Wang","orcid":"https://orcid.org/0000-0002-7266-2412"},"institutions":[{"id":"https://openalex.org/I130238516","display_name":"University of Minnesota","ror":"https://ror.org/017zqws13","country_code":"US","type":"education","lineage":["https://openalex.org/I130238516"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Gang Wang","raw_affiliation_strings":["Digital Technology Center and the Department of Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN, USA"],"affiliations":[{"raw_affiliation_string":"Digital Technology Center and the Department of Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN, USA","institution_ids":["https://openalex.org/I130238516"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5026758314","display_name":"Georgios B. Giannakis","orcid":"https://orcid.org/0000-0002-0196-0260"},"institutions":[{"id":"https://openalex.org/I130238516","display_name":"University of Minnesota","ror":"https://ror.org/017zqws13","country_code":"US","type":"education","lineage":["https://openalex.org/I130238516"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Georgios B. Giannakis","raw_affiliation_strings":["Digital Technology Center and the Department of Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN, USA"],"affiliations":[{"raw_affiliation_string":"Digital Technology Center and the Department of Electrical and Computer Engineering, University of Minnesota, Minneapolis, MN, USA","institution_ids":["https://openalex.org/I130238516"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5100333005","display_name":"Jie Chen","orcid":"https://orcid.org/0000-0003-2449-9793"},"institutions":[{"id":"https://openalex.org/I116953780","display_name":"Tongji University","ror":"https://ror.org/03rc6as71","country_code":"CN","type":"education","lineage":["https://openalex.org/I116953780"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jie Chen","raw_affiliation_strings":["Tongji University, Shanghai, China"],"affiliations":[{"raw_affiliation_string":"Tongji University, Shanghai, China","institution_ids":["https://openalex.org/I116953780"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5100367403"],"corresponding_institution_ids":["https://openalex.org/I130238516"],"apc_list":null,"apc_paid":null,"fwci":13.8875,"has_fulltext":false,"cited_by_count":150,"citation_normalized_percentile":{"value":0.99050907,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":98,"max":100},"biblio":{"volume":"67","issue":"9","first_page":"2357","last_page":"2370"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12676","display_name":"Machine Learning and ELM","score":0.9998000264167786,"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/T12676","display_name":"Machine Learning and ELM","score":0.9998000264167786,"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":0.9994999766349792,"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":0.9990000128746033,"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/maxima-and-minima","display_name":"Maxima and minima","score":0.6436432600021362},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.5888090133666992},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.571487545967102},{"id":"https://openalex.org/keywords/stochastic-gradient-descent","display_name":"Stochastic gradient descent","score":0.5342874526977539},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.5217293500900269},{"id":"https://openalex.org/keywords/initialization","display_name":"Initialization","score":0.4724177420139313},{"id":"https://openalex.org/keywords/hinge-loss","display_name":"Hinge loss","score":0.4638403058052063},{"id":"https://openalex.org/keywords/upper-and-lower-bounds","display_name":"Upper and lower bounds","score":0.41057324409484863},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3493621349334717},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.33400410413742065},{"id":"https://openalex.org/keywords/support-vector-machine","display_name":"Support vector machine","score":0.14496928453445435}],"concepts":[{"id":"https://openalex.org/C186633575","wikidata":"https://www.wikidata.org/wiki/Q845060","display_name":"Maxima and minima","level":2,"score":0.6436432600021362},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5888090133666992},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.571487545967102},{"id":"https://openalex.org/C206688291","wikidata":"https://www.wikidata.org/wiki/Q7617819","display_name":"Stochastic gradient descent","level":3,"score":0.5342874526977539},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.5217293500900269},{"id":"https://openalex.org/C114466953","wikidata":"https://www.wikidata.org/wiki/Q6034165","display_name":"Initialization","level":2,"score":0.4724177420139313},{"id":"https://openalex.org/C39891107","wikidata":"https://www.wikidata.org/wiki/Q5767098","display_name":"Hinge loss","level":3,"score":0.4638403058052063},{"id":"https://openalex.org/C77553402","wikidata":"https://www.wikidata.org/wiki/Q13222579","display_name":"Upper and lower bounds","level":2,"score":0.41057324409484863},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3493621349334717},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.33400410413742065},{"id":"https://openalex.org/C12267149","wikidata":"https://www.wikidata.org/wiki/Q282453","display_name":"Support vector machine","level":2,"score":0.14496928453445435},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1109/tsp.2019.2904921","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tsp.2019.2904921","pdf_url":null,"source":{"id":"https://openalex.org/S168680287","display_name":"IEEE Transactions on Signal Processing","issn_l":"1053-587X","issn":["1053-587X","1941-0476"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Signal Processing","raw_type":"journal-article"},{"id":"pmh:oai:arXiv.org:1808.04685","is_oa":true,"landing_page_url":"http://arxiv.org/abs/1808.04685","pdf_url":"https://arxiv.org/pdf/1808.04685","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:1808.04685","is_oa":true,"landing_page_url":"http://arxiv.org/abs/1808.04685","pdf_url":"https://arxiv.org/pdf/1808.04685","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":[],"awards":[{"id":"https://openalex.org/G1515802232","display_name":null,"funder_award_id":"1711471","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G3789437314","display_name":null,"funder_award_id":"1505970","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G4232543837","display_name":null,"funder_award_id":"1514056","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G4963119340","display_name":null,"funder_award_id":"1500713","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G570175137","display_name":null,"funder_award_id":"61621063","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"},{"id":"https://openalex.org/G8812163771","display_name":null,"funder_award_id":"U1509215","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":120,"referenced_works":["https://openalex.org/W65382156","https://openalex.org/W607505555","https://openalex.org/W1533861849","https://openalex.org/W1551360398","https://openalex.org/W1665214252","https://openalex.org/W1686810756","https://openalex.org/W1697075315","https://openalex.org/W2022740958","https://openalex.org/W2037699108","https://openalex.org/W2040870580","https://openalex.org/W2098437527","https://openalex.org/W2098874108","https://openalex.org/W2101762657","https://openalex.org/W2122575255","https://openalex.org/W2124136621","https://openalex.org/W2127265926","https://openalex.org/W2128865943","https://openalex.org/W2147366477","https://openalex.org/W2163605009","https://openalex.org/W2242818861","https://openalex.org/W2423689290","https://openalex.org/W2579923771","https://openalex.org/W2591714514","https://openalex.org/W2604117713","https://openalex.org/W2618530766","https://openalex.org/W2625063094","https://openalex.org/W2709553318","https://openalex.org/W2766196653","https://openalex.org/W2773177849","https://openalex.org/W2774044172","https://openalex.org/W2777256551","https://openalex.org/W2787132992","https://openalex.org/W2789199296","https://openalex.org/W2803955134","https://openalex.org/W2804822090","https://openalex.org/W2806252860","https://openalex.org/W2807930578","https://openalex.org/W2808157231","https://openalex.org/W2808958252","https://openalex.org/W2809670082","https://openalex.org/W2886067286","https://openalex.org/W2892723008","https://openalex.org/W2896834587","https://openalex.org/W2900620189","https://openalex.org/W2907931557","https://openalex.org/W2908206262","https://openalex.org/W2911742574","https://openalex.org/W2914605961","https://openalex.org/W2919115771","https://openalex.org/W2949148114","https://openalex.org/W2949804919","https://openalex.org/W2950618860","https://openalex.org/W2952104325","https://openalex.org/W2952318479","https://openalex.org/W2952410051","https://openalex.org/W2962754331","https://openalex.org/W2962767131","https://openalex.org/W2962990163","https://openalex.org/W2963100491","https://openalex.org/W2963383839","https://openalex.org/W2963417959","https://openalex.org/W2963446085","https://openalex.org/W2963588390","https://openalex.org/W2963651774","https://openalex.org/W2963695615","https://openalex.org/W2963716183","https://openalex.org/W2963739978","https://openalex.org/W2963744427","https://openalex.org/W2963982496","https://openalex.org/W2964072429","https://openalex.org/W2964088238","https://openalex.org/W2964106499","https://openalex.org/W2964262188","https://openalex.org/W2966289053","https://openalex.org/W2970618525","https://openalex.org/W2975778934","https://openalex.org/W3001454063","https://openalex.org/W3021189130","https://openalex.org/W3021623424","https://openalex.org/W3120740533","https://openalex.org/W4236362309","https://openalex.org/W4285719527","https://openalex.org/W4288400754","https://openalex.org/W4288593476","https://openalex.org/W4288626239","https://openalex.org/W4293448676","https://openalex.org/W4294087498","https://openalex.org/W4297738147","https://openalex.org/W4299785371","https://openalex.org/W4299804536","https://openalex.org/W4310348270","https://openalex.org/W6602644328","https://openalex.org/W6608025168","https://openalex.org/W6637242042","https://openalex.org/W6637373629","https://openalex.org/W6675100857","https://openalex.org/W6678917918","https://openalex.org/W6681915100","https://openalex.org/W6684191040","https://openalex.org/W6713348437","https://openalex.org/W6717556742","https://openalex.org/W6718212895","https://openalex.org/W6735544424","https://openalex.org/W6739166439","https://openalex.org/W6744989570","https://openalex.org/W6746339620","https://openalex.org/W6747113094","https://openalex.org/W6747847215","https://openalex.org/W6751437525","https://openalex.org/W6751860949","https://openalex.org/W6752360885","https://openalex.org/W6754634692","https://openalex.org/W6757310519","https://openalex.org/W6757628700","https://openalex.org/W6758158593","https://openalex.org/W6759039803","https://openalex.org/W6759176382","https://openalex.org/W6765617343","https://openalex.org/W6766511095","https://openalex.org/W6891931234"],"related_works":["https://openalex.org/W2596692027","https://openalex.org/W2148025067","https://openalex.org/W4295724548","https://openalex.org/W2549189808","https://openalex.org/W3197169288","https://openalex.org/W3118431319","https://openalex.org/W3004700632","https://openalex.org/W3177888558","https://openalex.org/W3128659392","https://openalex.org/W2963588390"],"abstract_inverted_index":{"Neural":[0],"networks":[1,26,182],"with":[2,184],"rectified":[3],"linear":[4],"unit":[5],"(ReLU)":[6],"activation":[7],"functions":[8],"(a.k.a.":[9],"ReLU":[10,25,59,104,181,208,215],"networks)":[11],"have":[12],"achieved":[13],"great":[14],"empirical":[15],"success":[16],"in":[17,63,123,200,217],"various":[18],"domains.":[19],"Nonetheless,":[20],"existing":[21],"results":[22],"for":[23,180],"learning":[24,56,206],"either":[27],"pose":[28],"assumptions":[29,135],"on":[30,136,166],"the":[31,40,53,69,82,111,128,137,146,167,185,201,232,235,238,242],"underlying":[32],"data":[33,70,138,227],"distribution":[34],"being,":[35],"e.g.,":[36],"Gaussian,":[37],"or":[38,142],"require":[39],"network":[41,60,105,209,216],"size":[42,45],"and/or":[43],"training":[44],"to":[46,106,150,172],"be":[47,173],"sufficiently":[48],"large.":[49],"In":[50],"this":[51,88],"context,":[52],"problem":[54],"of":[55,84,113,130,145,169,219,234,241],"a":[57,64,75,91,151,163,195,207,213],"two-layer":[58],"is":[61,79,127,154],"approached":[62],"binary":[65],"classification":[66],"setting,":[67],"where":[68],"are":[71,178],"linearly":[72],"separable":[73],"and":[74,120,162,228,237],"hinge":[76],"loss":[77],"criterion":[78],"adopted.":[80],"Leveraging":[81],"power":[83],"random":[85],"noise":[86],"perturbation,":[87],"paper":[89],"presents":[90],"novel":[92,186],"stochastic":[93],"gradient":[94],"descent":[95],"(SGD)":[96],"algorithm,":[97],"which":[98],"can":[99],"provably":[100],"train":[101],"any":[102],"single-hidden-layer":[103],"attain":[107],"global":[108,152],"optimality,":[109],"despite":[110],"presence":[112],"infinitely":[114],"many":[115],"bad":[116],"local":[117],"minima,":[118],"maxima,":[119],"saddle":[121],"points":[122],"general.":[124],"This":[125],"result":[126],"first":[129],"its":[131],"kind,":[132],"requiring":[133],"no":[134],"distribution,":[139],"training/network":[140],"size,":[141],"initialization.":[143],"Convergence":[144],"resultant":[147],"iterative":[148],"algorithm":[149,236],"minimum":[153],"analyzed":[155],"by":[156],"establishing":[157],"both":[158,225],"an":[159],"upper":[160],"bound":[161,165],"lower":[164],"number":[168],"non-zero":[170],"updates":[171],"performed.":[174],"Moreover,":[175],"generalization":[176],"guarantees":[177,193],"developed":[179],"trained":[183],"SGD":[187],"leveraging":[188],"classic":[189],"compression":[190],"bounds.":[191],"These":[192],"highlight":[194],"key":[196],"difference":[197],"(at":[198],"least":[199],"worst":[202],"case)":[203],"between":[204],"reliably":[205],"as":[210,212],"well":[211],"leaky":[214],"terms":[218],"sample":[220],"complexity.":[221],"Numerical":[222],"tests":[223],"using":[224],"synthetic":[226],"real":[229],"images":[230],"validate":[231],"effectiveness":[233],"practical":[239],"merits":[240],"theory.":[243]},"counts_by_year":[{"year":2025,"cited_by_count":14},{"year":2024,"cited_by_count":12},{"year":2023,"cited_by_count":22},{"year":2022,"cited_by_count":28},{"year":2021,"cited_by_count":31},{"year":2020,"cited_by_count":18},{"year":2019,"cited_by_count":19},{"year":2018,"cited_by_count":6}],"updated_date":"2026-03-20T23:20:44.827607","created_date":"2025-10-10T00:00:00"}