{"id":"https://openalex.org/W7134273872","doi":"https://doi.org/10.48550/arxiv.2603.05582","title":"Bias In, Bias Out? Finding Unbiased Subnetworks in Vanilla Models","display_name":"Bias In, Bias Out? Finding Unbiased Subnetworks in Vanilla Models","publication_year":2026,"publication_date":"2026-03-05","ids":{"openalex":"https://openalex.org/W7134273872","doi":"https://doi.org/10.48550/arxiv.2603.05582"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2603.05582","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"type":"article","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5128537735","display_name":"Ivan Luiz De Moura Matos","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Matos, Ivan Luiz De Moura","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128394056","display_name":"Abdel Djalil Sad Saoud","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Saoud, Abdel Djalil Sad","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128479603","display_name":"Ekaterina Iakovleva","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Iakovleva, Ekaterina","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073140566","display_name":"Vito Paolo Pastore","orcid":"https://orcid.org/0000-0002-5827-5571"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pastore, Vito Paolo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5128442544","display_name":"Enzo Tartaglione","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tartaglione, Enzo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.21855494,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.5011000037193298,"subfield":{"id":"https://openalex.org/subfields/3311","display_name":"Safety Research"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},"topics":[{"id":"https://openalex.org/T10883","display_name":"Ethics and Social Impacts of AI","score":0.5011000037193298,"subfield":{"id":"https://openalex.org/subfields/3311","display_name":"Safety Research"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.30070000886917114,"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/T11704","display_name":"Mobile Crowdsensing and Crowdsourcing","score":0.0997999981045723,"subfield":{"id":"https://openalex.org/subfields/1706","display_name":"Computer Science Applications"},"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/subnetwork","display_name":"Subnetwork","score":0.8628000020980835},{"id":"https://openalex.org/keywords/debiasing","display_name":"Debiasing","score":0.8148999810218811},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.6753000020980835},{"id":"https://openalex.org/keywords/overfitting","display_name":"Overfitting","score":0.5626999735832214},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.45260000228881836},{"id":"https://openalex.org/keywords/adaptation","display_name":"Adaptation (eye)","score":0.4480000138282776},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.41929998993873596},{"id":"https://openalex.org/keywords/scheme","display_name":"Scheme (mathematics)","score":0.3772999942302704}],"concepts":[{"id":"https://openalex.org/C2780186347","wikidata":"https://www.wikidata.org/wiki/Q11414","display_name":"Subnetwork","level":2,"score":0.8628000020980835},{"id":"https://openalex.org/C2779458634","wikidata":"https://www.wikidata.org/wiki/Q24963715","display_name":"Debiasing","level":2,"score":0.8148999810218811},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.743399977684021},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.6753000020980835},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6560999751091003},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.582099974155426},{"id":"https://openalex.org/C22019652","wikidata":"https://www.wikidata.org/wiki/Q331309","display_name":"Overfitting","level":3,"score":0.5626999735832214},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.45260000228881836},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.4480000138282776},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.41929998993873596},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.3772999942302704},{"id":"https://openalex.org/C2778712577","wikidata":"https://www.wikidata.org/wiki/Q3505966","display_name":"Retraining","level":2,"score":0.3619999885559082},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.35280001163482666},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.34389999508857727},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.34150001406669617},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.32010000944137573},{"id":"https://openalex.org/C63584917","wikidata":"https://www.wikidata.org/wiki/Q333286","display_name":"Bounding overwatch","level":2,"score":0.29789999127388},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.27469998598098755},{"id":"https://openalex.org/C116834253","wikidata":"https://www.wikidata.org/wiki/Q2039217","display_name":"Identification (biology)","level":2,"score":0.26759999990463257}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2603.05582","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"doi:10.48550/arxiv.2603.05582","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.05582","pdf_url":null,"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":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:doi:10.48550/arxiv.2603.05582","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0],"issue":[1],"of":[2,13,18,123,153,158,164],"algorithmic":[3],"biases":[4],"in":[5,156],"deep":[6],"learning":[7,65],"has":[8],"led":[9],"to":[10,35,132],"the":[11,84,151,159,165],"development":[12],"various":[14],"debiasing":[15],"techniques,":[16],"many":[17],"which":[19],"perform":[20],"complex":[21],"training":[22,53],"procedures":[23],"or":[24,82,139],"dataset":[25],"manipulation.":[26],"However,":[27],"an":[28],"intriguing":[29],"question":[30],"arises:":[31],"is":[32],"it":[33],"possible":[34],"extract":[36],"fair":[37],"and":[38,69,98,109,161],"bias-agnostic":[39],"subnetworks":[40,72,92],"from":[41],"standard":[42],"vanilla-trained":[43],"models":[44],"without":[45,80,101],"relying":[46,104],"on":[47,106,147],"additional":[48],"data,":[49],"such":[50,91],"as":[51,130],"unbiased":[52],"set?":[54],"In":[55],"this":[56],"work,":[57],"we":[58],"introduce":[59],"Bias-Invariant":[60],"Subnetwork":[61],"Extraction":[62],"(BISE),":[63],"a":[64],"strategy":[66],"that":[67,73,90,135],"identifies":[68],"isolates":[70],"\"bias-free\"":[71],"already":[74],"exist":[75],"within":[76],"conventionally":[77],"trained":[78],"models,":[79],"retraining":[81],"finetuning":[83],"original":[85],"parameters.":[86,144],"Our":[87,113],"approach":[88,155],"demonstrates":[89],"can":[93,99],"be":[94],"extracted":[95],"via":[96,127],"pruning":[97],"operate":[100],"modification,":[102],"effectively":[103],"less":[105],"biased":[107],"features":[108],"maintaining":[110],"robust":[111],"performance.":[112],"findings":[114],"contribute":[115],"towards":[116],"efficient":[117],"bias":[118],"mitigation":[119],"through":[120],"structural":[121],"adaptation":[122],"pre-trained":[124],"neural":[125],"networks":[126],"parameter":[128],"removal,":[129],"opposed":[131],"costly":[133],"strategies":[134],"are":[136],"either":[137],"data-centric":[138],"involve":[140],"(re)training":[141],"all":[142],"model":[143],"Extensive":[145],"experiments":[146],"common":[148],"benchmarks":[149],"show":[150],"advantages":[152],"our":[154],"terms":[157],"performance":[160],"computational":[162],"efficiency":[163],"resulting":[166],"debiased":[167],"model.":[168]},"counts_by_year":[],"updated_date":"2026-07-15T18:14:33.161393","created_date":"2026-03-10T00:00:00"}