{"id":"https://openalex.org/W4403486668","doi":"https://doi.org/10.3233/faia240970","title":"REFINE-LM: Mitigating Language Model Stereotypes via Reinforcement Learning","display_name":"REFINE-LM: Mitigating Language Model Stereotypes via Reinforcement Learning","publication_year":2024,"publication_date":"2024-10-16","ids":{"openalex":"https://openalex.org/W4403486668","doi":"https://doi.org/10.3233/faia240970"},"language":"en","primary_location":{"id":"doi:10.3233/faia240970","is_oa":true,"landing_page_url":"https://doi.org/10.3233/faia240970","pdf_url":"https://ebooks.iospress.nl/pdf/doi/10.3233/FAIA240970","source":{"id":"https://openalex.org/S4210201731","display_name":"Frontiers in artificial intelligence and applications","issn_l":"0922-6389","issn":["0922-6389","1879-8314"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Frontiers in Artificial Intelligence and Applications","raw_type":"book-chapter"},"type":"book-chapter","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://ebooks.iospress.nl/pdf/doi/10.3233/FAIA240970","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5107179436","display_name":"Rameez Qureshi","orcid":null},"institutions":[{"id":"https://openalex.org/I205274468","display_name":"Trinity College Dublin","ror":"https://ror.org/02tyrky19","country_code":"IE","type":"education","lineage":["https://openalex.org/I205274468"]},{"id":"https://openalex.org/I42934936","display_name":"Dublin City University","ror":"https://ror.org/04a1a1e81","country_code":"IE","type":"education","lineage":["https://openalex.org/I42934936"]}],"countries":["IE"],"is_corresponding":false,"raw_author_name":"Rameez Qureshi","raw_affiliation_strings":["ADAPT Centre, Trinity College Dublin, Ireland","ADAPT Centre (School of Computing, Dublin City University, Glasnevin, Dublin 9 - Ireland)"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ADAPT Centre, Trinity College Dublin, Ireland","institution_ids":["https://openalex.org/I205274468"]},{"raw_affiliation_string":"ADAPT Centre (School of Computing, Dublin City University, Glasnevin, Dublin 9 - Ireland)","institution_ids":["https://openalex.org/I42934936"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5107179437","display_name":"Na\u00efm Es-Sebbani","orcid":null},"institutions":[{"id":"https://openalex.org/I1294671590","display_name":"Centre National de la Recherche Scientifique","ror":"https://ror.org/02feahw73","country_code":"FR","type":"government","lineage":["https://openalex.org/I1294671590"]},{"id":"https://openalex.org/I44563897","display_name":"Universit\u00e9 d'Artois","ror":"https://ror.org/053x9s498","country_code":"FR","type":"education","lineage":["https://openalex.org/I44563897"]}],"countries":["FR"],"is_corresponding":false,"raw_author_name":"Na\u00efm Es-Sebbani","raw_affiliation_strings":["CRIL CNRS, Univ Artois, France"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"CRIL CNRS, Univ Artois, France","institution_ids":["https://openalex.org/I44563897","https://openalex.org/I1294671590"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5081390951","display_name":"Luis Gal\u00e1rraga","orcid":"https://orcid.org/0000-0002-0241-5379"},"institutions":[{"id":"https://openalex.org/I1326498283","display_name":"Institut national de recherche en sciences et technologies du num\u00e9rique","ror":"https://ror.org/02kvxyf05","country_code":"FR","type":"government","lineage":["https://openalex.org/I1326498283"]},{"id":"https://openalex.org/I2802519937","display_name":"Institut de Recherche en Informatique et Syst\u00e8mes Al\u00e9atoires","ror":"https://ror.org/00myn0z94","country_code":"FR","type":"facility","lineage":["https://openalex.org/I1294671590","https://openalex.org/I1294671590","https://openalex.org/I1326498283","https://openalex.org/I205703379","https://openalex.org/I2802204017","https://openalex.org/I2802519937","https://openalex.org/I28221208","https://openalex.org/I4210127572","https://openalex.org/I4210159245","https://openalex.org/I56067802"]}],"countries":["FR"],"is_corresponding":false,"raw_author_name":"Luis Gal\u00e1rraga","raw_affiliation_strings":["INRIA/IRISA Rennes, France","IRISA - LA ou URA 227 - INRIA Rennes (France)"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"INRIA/IRISA Rennes, France","institution_ids":["https://openalex.org/I2802519937","https://openalex.org/I1326498283"]},{"raw_affiliation_string":"IRISA - LA ou URA 227 - INRIA Rennes (France)","institution_ids":["https://openalex.org/I2802519937","https://openalex.org/I1326498283"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5107179438","display_name":"Yvette Graham","orcid":null},"institutions":[{"id":"https://openalex.org/I205274468","display_name":"Trinity College Dublin","ror":"https://ror.org/02tyrky19","country_code":"IE","type":"education","lineage":["https://openalex.org/I205274468"]},{"id":"https://openalex.org/I42934936","display_name":"Dublin City University","ror":"https://ror.org/04a1a1e81","country_code":"IE","type":"education","lineage":["https://openalex.org/I42934936"]}],"countries":["IE"],"is_corresponding":false,"raw_author_name":"Yvette Graham","raw_affiliation_strings":["ADAPT Centre, Trinity College Dublin, Ireland","ADAPT Centre (School of Computing, Dublin City University, Glasnevin, Dublin 9 - Ireland)"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"ADAPT Centre, Trinity College Dublin, Ireland","institution_ids":["https://openalex.org/I205274468"]},{"raw_affiliation_string":"ADAPT Centre (School of Computing, Dublin City University, Glasnevin, Dublin 9 - Ireland)","institution_ids":["https://openalex.org/I42934936"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5066732273","display_name":"Miguel Couceiro","orcid":"https://orcid.org/0000-0003-2316-7623"},"institutions":[{"id":"https://openalex.org/I121345201","display_name":"Instituto de Engenharia de Sistemas e Computadores Investiga\u00e7\u00e3o e Desenvolvimento","ror":"https://ror.org/04mqy3p58","country_code":"PT","type":"nonprofit","lineage":["https://openalex.org/I121345201","https://openalex.org/I4210125590"]},{"id":"https://openalex.org/I1294671590","display_name":"Centre National de la Recherche Scientifique","ror":"https://ror.org/02feahw73","country_code":"FR","type":"government","lineage":["https://openalex.org/I1294671590"]},{"id":"https://openalex.org/I141596103","display_name":"University of Lisbon","ror":"https://ror.org/01c27hj86","country_code":"PT","type":"education","lineage":["https://openalex.org/I141596103"]},{"id":"https://openalex.org/I4210121838","display_name":"Laboratoire Lorrain de Recherche en Informatique et ses Applications","ror":"https://ror.org/02vnf0c38","country_code":"FR","type":"facility","lineage":["https://openalex.org/I1294671590","https://openalex.org/I1294671590","https://openalex.org/I1326498283","https://openalex.org/I277688954","https://openalex.org/I4210107720","https://openalex.org/I4210121838","https://openalex.org/I4210159245","https://openalex.org/I90183372"]},{"id":"https://openalex.org/I90183372","display_name":"Universit\u00e9 de Lorraine","ror":"https://ror.org/04vfs2w97","country_code":"FR","type":"education","lineage":["https://openalex.org/I90183372"]}],"countries":["FR","PT"],"is_corresponding":false,"raw_author_name":"Miguel Couceiro","raw_affiliation_strings":["INESC-ID, IST, Universidade de Lisboa, Portugal","Universit\u00e9 de Lorraine, CNRS, LORIA, France","ORPAILLEUR - Knowledge representation, reasonning (France)"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"INESC-ID, IST, Universidade de Lisboa, Portugal","institution_ids":["https://openalex.org/I121345201","https://openalex.org/I141596103"]},{"raw_affiliation_string":"Universit\u00e9 de Lorraine, CNRS, LORIA, France","institution_ids":["https://openalex.org/I90183372","https://openalex.org/I4210121838","https://openalex.org/I1294671590"]},{"raw_affiliation_string":"ORPAILLEUR - Knowledge representation, reasonning (France)","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5107179439","display_name":"Zied Bouraoui","orcid":null},"institutions":[{"id":"https://openalex.org/I1294671590","display_name":"Centre National de la Recherche Scientifique","ror":"https://ror.org/02feahw73","country_code":"FR","type":"government","lineage":["https://openalex.org/I1294671590"]},{"id":"https://openalex.org/I44563897","display_name":"Universit\u00e9 d'Artois","ror":"https://ror.org/053x9s498","country_code":"FR","type":"education","lineage":["https://openalex.org/I44563897"]}],"countries":["FR"],"is_corresponding":false,"raw_author_name":"Zied Bouraoui","raw_affiliation_strings":["CRIL CNRS, Univ Artois, France"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"CRIL CNRS, Univ Artois, France","institution_ids":["https://openalex.org/I44563897","https://openalex.org/I1294671590"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":10,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.35732039,"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/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.9613000154495239,"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/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.9613000154495239,"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":0.9498999714851379,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.9312000274658203,"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/debiasing","display_name":"Debiasing","score":0.9689981937408447},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7439891695976257},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.6434183716773987},{"id":"https://openalex.org/keywords/preprocessor","display_name":"Preprocessor","score":0.5272048115730286},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5135574340820312},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.45680755376815796},{"id":"https://openalex.org/keywords/simple","display_name":"Simple (philosophy)","score":0.45268407464027405},{"id":"https://openalex.org/keywords/range","display_name":"Range (aeronautics)","score":0.44351261854171753},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.3394731879234314},{"id":"https://openalex.org/keywords/psychology","display_name":"Psychology","score":0.1677524447441101},{"id":"https://openalex.org/keywords/social-psychology","display_name":"Social psychology","score":0.15784865617752075}],"concepts":[{"id":"https://openalex.org/C2779458634","wikidata":"https://www.wikidata.org/wiki/Q24963715","display_name":"Debiasing","level":2,"score":0.9689981937408447},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7439891695976257},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.6434183716773987},{"id":"https://openalex.org/C34736171","wikidata":"https://www.wikidata.org/wiki/Q918333","display_name":"Preprocessor","level":2,"score":0.5272048115730286},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5135574340820312},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.45680755376815796},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.45268407464027405},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.44351261854171753},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3394731879234314},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.1677524447441101},{"id":"https://openalex.org/C77805123","wikidata":"https://www.wikidata.org/wiki/Q161272","display_name":"Social psychology","level":1,"score":0.15784865617752075},{"id":"https://openalex.org/C159985019","wikidata":"https://www.wikidata.org/wiki/Q181790","display_name":"Composite material","level":1,"score":0.0},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.0},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.3233/faia240970","is_oa":true,"landing_page_url":"https://doi.org/10.3233/faia240970","pdf_url":"https://ebooks.iospress.nl/pdf/doi/10.3233/FAIA240970","source":{"id":"https://openalex.org/S4210201731","display_name":"Frontiers in artificial intelligence and applications","issn_l":"0922-6389","issn":["0922-6389","1879-8314"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Frontiers in Artificial Intelligence and Applications","raw_type":"book-chapter"},{"id":"pmh:oai:HAL:hal-04872144v1","is_oa":true,"landing_page_url":"https://hal.science/hal-04872144v1/document","pdf_url":"https://hal.science/hal-04872144/document","source":{"id":"https://openalex.org/S4306402512","display_name":"HAL (Le Centre pour la Communication Scientifique Directe)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1294671590","host_organization_name":"Centre National de la Recherche Scientifique","host_organization_lineage":["https://openalex.org/I1294671590"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"ECAI 2024 - 27th European Conference on Artificial Intelligence, Oct 2024, Santiago de Compostela, Spain. pp.4027-4034, &#x27E8;10.3233/FAIA240970&#x27E9;","raw_type":"info:eu-repo/semantics/conferenceObject"}],"best_oa_location":{"id":"doi:10.3233/faia240970","is_oa":true,"landing_page_url":"https://doi.org/10.3233/faia240970","pdf_url":"https://ebooks.iospress.nl/pdf/doi/10.3233/FAIA240970","source":{"id":"https://openalex.org/S4210201731","display_name":"Frontiers in artificial intelligence and applications","issn_l":"0922-6389","issn":["0922-6389","1879-8314"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Frontiers in Artificial Intelligence and Applications","raw_type":"book-chapter"},"sustainable_development_goals":[{"display_name":"Quality Education","score":0.5899999737739563,"id":"https://metadata.un.org/sdg/4"}],"awards":[{"id":"https://openalex.org/G3857068732","display_name":"Intrinsic and Extrinsic evaluation of biases in large language models","funder_award_id":"ANR-23-IAS1-0004","funder_id":"https://openalex.org/F4320320883","funder_display_name":"Agence Nationale de la Recherche"},{"id":"https://openalex.org/G589531907","display_name":"Event-Centric Reasoning for Interpreting Everyday Narratives","funder_award_id":"ANR-22-CE23-0002","funder_id":"https://openalex.org/F4320320883","funder_display_name":"Agence Nationale de la Recherche"},{"id":"https://openalex.org/G6334448832","display_name":"ADAPT_Phase 2","funder_award_id":"13/RC/2106_P2","funder_id":"https://openalex.org/F4320320847","funder_display_name":"Science Foundation Ireland"}],"funders":[{"id":"https://openalex.org/F4320320847","display_name":"Science Foundation Ireland","ror":"https://ror.org/0271asj38"},{"id":"https://openalex.org/F4320320864","display_name":"Trinity College Dublin","ror":"https://ror.org/02tyrky19"},{"id":"https://openalex.org/F4320320883","display_name":"Agence Nationale de la Recherche","ror":"https://ror.org/00rbzpz17"}],"has_content":{"grobid_xml":false,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4403486668.pdf"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4362554880","https://openalex.org/W4281684980","https://openalex.org/W4386875279","https://openalex.org/W2171721708","https://openalex.org/W4390963114","https://openalex.org/W4287887864","https://openalex.org/W3214527415","https://openalex.org/W1495104519","https://openalex.org/W4225584739","https://openalex.org/W2199432031"],"abstract_inverted_index":{"With":[0],"the":[1,13,58,109],"introduction":[2],"of":[3,26,64,76,96,108,113,138,162],"(large)":[4],"language":[5],"models,":[6,139],"there":[7],"has":[8],"been":[9],"significant":[10,129],"concern":[11],"about":[12],"unintended":[14],"bias":[15,117],"such":[16,31,167],"models":[17,32],"may":[18],"inherit":[19],"from":[20],"their":[21],"training":[22,102],"data.":[23],"A":[24],"number":[25],"studies":[27],"have":[28],"shown":[29],"that":[30,88,144],"propagate":[33],"gender":[34],"stereotypes,":[35],"as":[36,38,168],"well":[37],"geographical":[39],"and":[40,55,67,172,175],"racial":[41],"bias,":[42],"among":[43],"other":[44],"biases.":[45,77],"While":[46],"existing":[47],"works":[48],"tackle":[49],"this":[50],"issue":[51],"by":[52],"preprocessing":[53],"data":[54],"debiasing":[56,86,124],"embeddings,":[57],"proposed":[59],"methods":[60],"require":[61],"a":[62,85,103,114,135],"lot":[63],"computational":[65,130],"resources":[66],"annotation":[68],"effort":[69],"while":[70,152],"being":[71],"limited":[72],"to":[73,92,159,181],"certain":[74],"types":[75,95,161],"To":[78],"address":[79],"these":[80],"issues,":[81],"we":[82],"introduce":[83],"REFINE-LM,":[84],"method":[87,121,146],"uses":[89],"reinforcement":[90,119],"learning":[91,120],"handle":[93],"different":[94,160],"biases":[97,151],"without":[98,125],"any":[99],"fine-tuning.":[100],"By":[101],"simple":[104],"model":[105,123],"on":[106,134],"top":[107],"word":[110],"probability":[111],"distribution":[112],"LM,":[115],"our":[116,145],"agnostic":[118],"enables":[122],"human":[126],"annotations":[127],"or":[128],"resources.":[131],"Experiments":[132],"conducted":[133],"wide":[136],"range":[137],"including":[140],"several":[141],"LMs,":[142],"show":[143],"(i)":[147],"significantly":[148],"reduces":[149],"stereotypical":[150],"preserving":[153],"LMs":[154],"performance;":[155],"(ii)":[156],"is":[157,178],"applicable":[158],"biases,":[163],"generalizing":[164],"across":[165],"contexts":[166],"gender,":[169],"ethnicity,":[170],"religion,":[171],"nationality-based":[173],"biases;":[174],"(iii)":[176],"it":[177],"not":[179],"expensive":[180],"train.":[182]},"counts_by_year":[],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2024-10-18T00:00:00"}
