{"id":"https://openalex.org/W4378945239","doi":"https://doi.org/10.48550/arxiv.2305.18887","title":"How Does Information Bottleneck Help Deep Learning?","display_name":"How Does Information Bottleneck Help Deep Learning?","publication_year":2023,"publication_date":"2023-05-30","ids":{"openalex":"https://openalex.org/W4378945239","doi":"https://doi.org/10.48550/arxiv.2305.18887"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2305.18887","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2305.18887","pdf_url":"https://arxiv.org/pdf/2305.18887","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"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2305.18887","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5003184366","display_name":"Kenji Kawaguchi","orcid":"https://orcid.org/0000-0003-1839-7504","observed_orcids":["https://orcid.org/0000-0003-1839-7504","https://orcid.org/0000-0002-0791-0870","https://orcid.org/0000-0002-5361-9793"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kawaguchi, Kenji","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006488129","display_name":"Zhun Deng","orcid":"https://orcid.org/0009-0008-7623-7705","observed_orcids":["https://orcid.org/0009-0008-7623-7705","https://orcid.org/0009-0002-3785-8513"]},"institutions":[{"id":"https://openalex.org/I78577930","display_name":"Columbia University","ror":"https://ror.org/00hj8s172","country_code":"US","type":"education","lineage":["https://openalex.org/I78577930"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Deng, Zhun","raw_affiliation_strings":["Columbia University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Columbia University","institution_ids":["https://openalex.org/I78577930"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102846500","display_name":"Xu Ji","orcid":"https://orcid.org/0009-0002-0849-7333","observed_orcids":["https://orcid.org/0009-0002-0849-7333"]},"institutions":[{"id":"https://openalex.org/I4210164802","display_name":"Mila - Quebec Artificial Intelligence Institute","ror":"https://ror.org/05c22rx21","country_code":"CA","type":"facility","lineage":["https://openalex.org/I4210164802"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Ji, Xu","raw_affiliation_strings":["Mila"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Mila","institution_ids":["https://openalex.org/I4210164802"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5007414666","display_name":"Jiaoyang Huang","orcid":"https://orcid.org/0000-0001-8266-1301","observed_orcids":["https://orcid.org/0000-0001-8266-1301"]},"institutions":[{"id":"https://openalex.org/I79576946","display_name":"University of Pennsylvania","ror":"https://ror.org/00b30xv10","country_code":"US","type":"education","lineage":["https://openalex.org/I79576946"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Huang, Jiaoyang","raw_affiliation_strings":["University of Pennsylvania"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Pennsylvania","institution_ids":["https://openalex.org/I79576946"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":17,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.7294999957084656,"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":0.7294999957084656,"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/T12072","display_name":"Machine Learning and Algorithms","score":0.10859999805688858,"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":0.10750000178813934,"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/information-bottleneck","display_name":"information bottleneck","score":0.9980000257492065},{"id":"https://openalex.org/keywords/generalization-error","display_name":"generalization error","score":0.8600000143051147},{"id":"https://openalex.org/keywords/generalization-bounds","display_name":"generalization bounds","score":0.9160000085830688},{"id":"https://openalex.org/keywords/deep-learning","display_name":"deep learning","score":0.7649999856948853},{"id":"https://openalex.org/keywords/learning-theory","display_name":"learning theory","score":0.6650000214576721}],"concepts":[{"id":"https://openalex.org/C60008888","wikidata":"https://www.wikidata.org/wiki/Q6031013","display_name":"Information bottleneck method","level":3,"score":0.9269613027572632},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.8878816366195679},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.7246874570846558},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.704213559627533},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5372635722160339},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4713689088821411},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.4676121473312378},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.4610530436038971},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.450636088848114},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.1535913348197937},{"id":"https://openalex.org/C152139883","wikidata":"https://www.wikidata.org/wiki/Q252973","display_name":"Mutual information","level":2,"score":0.14447468519210815},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.0},{"id":"https://openalex.org/C104317684","wikidata":"https://www.wikidata.org/wiki/Q7187","display_name":"Gene","level":2,"score":0.0},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"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/C55493867","wikidata":"https://www.wikidata.org/wiki/Q7094","display_name":"Biochemistry","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2305.18887","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2305.18887","pdf_url":"https://arxiv.org/pdf/2305.18887","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.2305.18887","doi":"https://doi.org/10.48550/arxiv.2305.18887","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2305.18887","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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2305.18887","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2305.18887","pdf_url":"https://arxiv.org/pdf/2305.18887","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":[],"sustainable_development_goals_aurora":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4378945239.pdf","grobid_xml":"https://content.openalex.org/works/W4378945239.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2622284819","https://openalex.org/W1504394672","https://openalex.org/W3089381707","https://openalex.org/W4367625101","https://openalex.org/W4285254085","https://openalex.org/W3034190530","https://openalex.org/W4385764429","https://openalex.org/W2741297526","https://openalex.org/W4295728955","https://openalex.org/W3129794609"],"abstract_inverted_index":{"Numerous":[0],"deep":[1,63,88],"learning":[2,53,64,116],"algorithms":[3],"have":[4],"been":[5,43],"inspired":[6],"by":[7,65],"and":[8,115,151],"understood":[9],"via":[10],"the":[11,50,57,94,101,128,134,137,155,166,172,178],"notion":[12],"of":[13,59,103,113,130,136,157,168,180],"information":[14,18,25,40,60,68,78,131,139,158,169],"bottleneck,":[15,170],"where":[16],"unnecessary":[17,138],"is":[19,26,36,80,92,191],"(often":[20],"implicitly)":[21],"minimized":[22],"while":[23],"task-relevant":[24],"maximized.":[27],"However,":[28],"a":[29,111,146],"rigorous":[30,52],"argument":[31],"for":[32,55,149],"justifying":[33,56],"why":[34],"it":[35,91],"desirable":[37],"to":[38,70,83,125],"control":[39,84],"bottlenecks":[41],"has":[42],"elusive.":[44],"In":[45,118],"this":[46],"paper,":[47],"we":[48],"provide":[49],"first":[51],"theory":[54,74],"benefit":[58],"bottleneck":[61,69,79],"in":[62,87],"mathematically":[66],"relating":[67],"generalization":[71,85,121,162],"errors.":[72],"Our":[73,160,189],"proves":[75],"that":[76,175],"controlling":[77],"one":[81],"way":[82],"errors":[86,122],"learning,":[89],"although":[90],"not":[93],"only":[95],"or":[96,187],"necessary":[97],"way.":[98],"We":[99],"investigate":[100],"merit":[102],"our":[104],"new":[105,161],"mathematical":[106],"findings":[107],"with":[108,127,165,177],"experiments":[109],"across":[110],"range":[112],"architectures":[114],"settings.":[117],"many":[119],"cases,":[120],"are":[123],"shown":[124],"correlate":[126],"degree":[129,167],"bottleneck:":[132],"i.e.,":[133],"amount":[135],"at":[140],"hidden":[141],"layers.":[142],"This":[143],"paper":[144],"provides":[145],"theoretical":[147],"foundation":[148],"current":[150],"future":[152],"methods":[153],"through":[154],"lens":[156],"bottleneck.":[159],"bounds":[163,174],"scale":[164,176],"unlike":[171],"previous":[173],"number":[179],"parameters,":[181],"VC":[182],"dimension,":[183],"Rademacher":[184],"complexity,":[185],"stability":[186],"robustness.":[188],"code":[190],"publicly":[192],"available":[193],"at:":[194],"https://github.com/xu-ji/information-bottleneck":[195]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":6},{"year":2024,"cited_by_count":9},{"year":2023,"cited_by_count":1}],"updated_date":"2026-10-16T09:59:22.235173","created_date":"2025-10-10T00:00:00"}