{"id":"https://openalex.org/W3198192082","doi":"https://doi.org/10.18653/v1/2021.emnlp-main.116","title":"Debiasing Methods in Natural Language Understanding Make Bias More Accessible","display_name":"Debiasing Methods in Natural Language Understanding Make Bias More Accessible","publication_year":2021,"publication_date":"2021-01-01","ids":{"openalex":"https://openalex.org/W3198192082","doi":"https://doi.org/10.18653/v1/2021.emnlp-main.116","mag":"3198192082"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2021.emnlp-main.116","doi":"https://doi.org/10.18653/v1/2021.emnlp-main.116","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2021.emnlp-main.116","pdf_url":"https://aclanthology.org/2021.emnlp-main.116.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://aclanthology.org/2021.emnlp-main.116.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Michael Mendelson","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I174306211","display_name":"Technion \u2013 Israel Institute of Technology","ror":"https://ror.org/03qryx823","country_code":"IL","type":"education","lineage":["https://openalex.org/I174306211"]}],"countries":["IL"],"is_corresponding":true,"raw_author_name":"Michael Mendelson","raw_affiliation_strings":["Faculty of Computer Science Technion -Israel Institute of Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Faculty of Computer Science Technion -Israel Institute of Technology","institution_ids":["https://openalex.org/I174306211"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5051184573","display_name":"Yonatan Belinkov","orcid":"https://orcid.org/0000-0002-6280-5045","observed_orcids":["https://orcid.org/0000-0002-6280-5045"]},"institutions":[{"id":"https://openalex.org/I174306211","display_name":"Technion \u2013 Israel Institute of Technology","ror":"https://ror.org/03qryx823","country_code":"IL","type":"education","lineage":["https://openalex.org/I174306211"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Yonatan Belinkov","raw_affiliation_strings":["Faculty of Computer Science Technion -Israel Institute of Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Faculty of Computer Science Technion -Israel Institute of Technology","institution_ids":["https://openalex.org/I174306211"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I174306211"],"apc_list":null,"apc_paid":null,"fwci":2.4809,"has_fulltext":true,"cited_by_count":14,"citation_normalized_percentile":{"value":0.91163855,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":89,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"1545","last_page":"1557"},"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":0.357699990272522,"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":0.357699990272522,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.25119999051094055,"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/T10028","display_name":"Topic Modeling","score":0.18029999732971191,"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/bias-mitigation","display_name":"bias mitigation","score":0.9589999914169312},{"id":"https://openalex.org/keywords/natural-language-understanding","display_name":"natural language understanding","score":0.8669999837875366},{"id":"https://openalex.org/keywords/language-models","display_name":"language models","score":0.6830000281333923},{"id":"https://openalex.org/keywords/probing","display_name":"probing","score":0.8450000286102295},{"id":"https://openalex.org/keywords/post-hoc-explanation","display_name":"post-hoc explanation","score":0.5910000205039978},{"id":"https://openalex.org/keywords/out-of-distribution-generalization","display_name":"out-of-distribution generalization","score":0.7770000100135803},{"id":"https://openalex.org/keywords/model-robustness","display_name":"model robustness","score":0.6240000128746033}],"concepts":[{"id":"https://openalex.org/C2779458634","wikidata":"https://www.wikidata.org/wiki/Q24963715","display_name":"Debiasing","level":2,"score":0.9725984334945679},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7979028820991516},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.7487313747406006},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.596817135810852},{"id":"https://openalex.org/C2779439875","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Natural language understanding","level":3,"score":0.5838631987571716},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5253838300704956},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4948708415031433},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.4654885530471802},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.449618935585022},{"id":"https://openalex.org/C527412718","wikidata":"https://www.wikidata.org/wiki/Q855395","display_name":"Interpretation (philosophy)","level":2,"score":0.44501036405563354},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4425920248031616},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.410356342792511},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3047565221786499},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.09739309549331665},{"id":"https://openalex.org/C188147891","wikidata":"https://www.wikidata.org/wiki/Q147638","display_name":"Cognitive science","level":1,"score":0.07463821768760681},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"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/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","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},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0}],"mesh":[],"locations_count":4,"locations":[{"id":"doi:10.18653/v1/2021.emnlp-main.116","doi":"https://doi.org/10.18653/v1/2021.emnlp-main.116","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2021.emnlp-main.116","pdf_url":"https://aclanthology.org/2021.emnlp-main.116.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2109.04095","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2109.04095","pdf_url":"https://arxiv.org/pdf/2109.04095","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.2109.04095","doi":"https://doi.org/10.48550/arxiv.2109.04095","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2109.04095","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},{"id":"mag:3211881375","doi":null,"is_oa":false,"landing_page_url":"https://aclanthology.org/2021.emnlp-main.116/","pdf_url":null,"source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":"Empirical Methods in Natural Language Processing","raw_type":null}],"best_oa_location":{"id":"doi:10.18653/v1/2021.emnlp-main.116","doi":"https://doi.org/10.18653/v1/2021.emnlp-main.116","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2021.emnlp-main.116","pdf_url":"https://aclanthology.org/2021.emnlp-main.116.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320323139","display_name":"Azrieli Foundation","ror":"https://ror.org/009svts64"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W3198192082.pdf","grobid_xml":"https://content.openalex.org/works/W3198192082.grobid-xml"},"referenced_works_count":34,"referenced_works":["https://openalex.org/W1821462560","https://openalex.org/W1840435438","https://openalex.org/W2116064496","https://openalex.org/W2515741950","https://openalex.org/W2607892599","https://openalex.org/W2789566302","https://openalex.org/W2790415926","https://openalex.org/W2798358706","https://openalex.org/W2893425640","https://openalex.org/W2900065283","https://openalex.org/W2906152891","https://openalex.org/W2951286828","https://openalex.org/W2962776659","https://openalex.org/W2963341956","https://openalex.org/W2963351448","https://openalex.org/W2963381846","https://openalex.org/W2964204621","https://openalex.org/W2969009758","https://openalex.org/W2970019270","https://openalex.org/W2970862333","https://openalex.org/W2975429091","https://openalex.org/W2979826702","https://openalex.org/W2984256198","https://openalex.org/W3013891696","https://openalex.org/W3016970897","https://openalex.org/W3023038345","https://openalex.org/W3035139434","https://openalex.org/W3037697022","https://openalex.org/W3088349782","https://openalex.org/W3099624838","https://openalex.org/W3110511794","https://openalex.org/W3126493605","https://openalex.org/W3193068792","https://openalex.org/W6893675597"],"related_works":["https://openalex.org/W3088349782","https://openalex.org/W4382202569","https://openalex.org/W3175352680","https://openalex.org/W4385570386","https://openalex.org/W3035139434","https://openalex.org/W4283796213","https://openalex.org/W4385571840","https://openalex.org/W4389520452","https://openalex.org/W3023038345","https://openalex.org/W6964939956"],"abstract_inverted_index":{"Model":[0],"robustness":[1],"to":[2,44,76],"bias":[3,112],"is":[4,39,104,113],"often":[5],"determined":[6],"by":[7,27],"the":[8,45,52,78,84,99,110],"generalization":[9],"on":[10,24],"carefully":[11],"designed":[12],"out-of-distribution":[13],"datasets.Recent":[14],"debiasing":[15],"methods":[16,38],"in":[17,51,68,116],"natural":[18],"language":[19,69,102],"understanding":[20],"(NLU)":[21],"improve":[22],"performance":[23],"such":[25,37],"datasets":[26,91],"pressuring":[28],"models":[29],"into":[30],"making":[31],"unbiased":[32],"predictions.An":[33],"underlying":[34],"assumption":[35],"behind":[36],"that":[40,61],"this":[41],"also":[42],"leads":[43],"discovery":[46],"of":[47,66,80],"more":[48,100,111],"robust":[49],"features":[50],"model's":[53,85],"inner":[54,118],"representations.We":[55,86],"propose":[56],"a":[57,101,107],"general":[58],"probing-based":[59],"framework":[60],"allows":[62],"for":[63],"posthoc":[64],"interpretation":[65],"biases":[67,82],"models,":[70],"and":[71,92,95],"use":[72],"an":[73],"information-theoretic":[74],"approach":[75],"measure":[77],"extractability":[79],"certain":[81],"from":[83],"experiment":[87],"with":[88],"several":[89],"NLU":[90],"known":[93],"biases,":[94],"show":[96],"that,":[97],"counter-intuitively,":[98],"model":[103],"pushed":[105],"towards":[106],"debiased":[108],"regime,":[109],"actually":[114],"encoded":[115],"its":[117],"representations.":[119],"1":[120]},"counts_by_year":[{"year":2025,"cited_by_count":3},{"year":2024,"cited_by_count":1},{"year":2023,"cited_by_count":6},{"year":2022,"cited_by_count":4}],"updated_date":"2026-10-31T12:09:11.096373","created_date":"2021-09-13T00:00:00"}