{"id":"https://openalex.org/W3134354193","doi":"https://doi.org/10.1162/tacl_a_00434","title":"Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP","display_name":"Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP","publication_year":2021,"publication_date":"2021-01-01","ids":{"openalex":"https://openalex.org/W3134354193","doi":"https://doi.org/10.1162/tacl_a_00434","mag":"3134354193"},"language":"en","primary_location":{"id":"doi:10.1162/tacl_a_00434","is_oa":true,"landing_page_url":"https://doi.org/10.1162/tacl_a_00434","pdf_url":"https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00434/1979270/tacl_a_00434.pdf","source":{"id":"https://openalex.org/S2729999759","display_name":"Transactions of the Association for Computational Linguistics","issn_l":"2307-387X","issn":["2307-387X"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Transactions of the Association for Computational Linguistics","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","datacite","doaj"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00434/1979270/tacl_a_00434.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5076269589","display_name":"Timo Schick","orcid":null},"institutions":[{"id":"https://openalex.org/I8204097","display_name":"Ludwig-Maximilians-Universit\u00e4t M\u00fcnchen","ror":"https://ror.org/05591te55","country_code":"DE","type":"education","lineage":["https://openalex.org/I8204097"]}],"countries":["DE"],"is_corresponding":true,"raw_author_name":"Timo Schick","raw_affiliation_strings":["Center for Information and Language Processing (CIS), LMU Munich, Germany. schickt@cis.lmu.de"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Center for Information and Language Processing (CIS), LMU Munich, Germany. schickt@cis.lmu.de","institution_ids":["https://openalex.org/I8204097"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5089986506","display_name":"Sahana Udupa","orcid":"https://orcid.org/0000-0003-3647-9570"},"institutions":[{"id":"https://openalex.org/I8204097","display_name":"Ludwig-Maximilians-Universit\u00e4t M\u00fcnchen","ror":"https://ror.org/05591te55","country_code":"DE","type":"education","lineage":["https://openalex.org/I8204097"]}],"countries":["DE"],"is_corresponding":true,"raw_author_name":"Sahana Udupa","raw_affiliation_strings":["Institute of Social and Cultural Anthropology, LMU Munich, Germany. sahana.udupa@lmu.de"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Institute of Social and Cultural Anthropology, LMU Munich, Germany. sahana.udupa@lmu.de","institution_ids":["https://openalex.org/I8204097"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5071144367","display_name":"Hinrich Sch\u00fctze","orcid":null},"institutions":[{"id":"https://openalex.org/I8204097","display_name":"Ludwig-Maximilians-Universit\u00e4t M\u00fcnchen","ror":"https://ror.org/05591te55","country_code":"DE","type":"education","lineage":["https://openalex.org/I8204097"]}],"countries":["DE"],"is_corresponding":true,"raw_author_name":"Hinrich Sch\u00fctze","raw_affiliation_strings":["Center for Information and Language Processing (CIS), LMU Munich, Germany. inquiries@cislmu.org"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Center for Information and Language Processing (CIS), LMU Munich, Germany. inquiries@cislmu.org","institution_ids":["https://openalex.org/I8204097"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":["https://openalex.org/A5071144367","https://openalex.org/A5076269589","https://openalex.org/A5089986506"],"corresponding_institution_ids":["https://openalex.org/I8204097"],"apc_list":null,"apc_paid":null,"fwci":22.0642,"has_fulltext":false,"cited_by_count":217,"citation_normalized_percentile":{"value":0.99590611,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":99,"max":100},"biblio":{"volume":"9","issue":null,"first_page":"1408","last_page":"1424"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9977999925613403,"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/T10028","display_name":"Topic Modeling","score":0.9977999925613403,"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/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.9970999956130981,"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":0.9965000152587891,"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":0.9830232858657837},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8863811492919922},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.6790969967842102},{"id":"https://openalex.org/keywords/offensive","display_name":"Offensive","score":0.6495510339736938},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.5958698987960815},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.5891956686973572},{"id":"https://openalex.org/keywords/word","display_name":"Word (group theory)","score":0.48201504349708557},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.4066106975078583},{"id":"https://openalex.org/keywords/linguistics","display_name":"Linguistics","score":0.10751137137413025}],"concepts":[{"id":"https://openalex.org/C2779458634","wikidata":"https://www.wikidata.org/wiki/Q24963715","display_name":"Debiasing","level":2,"score":0.9830232858657837},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8863811492919922},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.6790969967842102},{"id":"https://openalex.org/C176856949","wikidata":"https://www.wikidata.org/wiki/Q2001676","display_name":"Offensive","level":2,"score":0.6495510339736938},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.5958698987960815},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5891956686973572},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.48201504349708557},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4066106975078583},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.10751137137413025},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.0},{"id":"https://openalex.org/C187736073","wikidata":"https://www.wikidata.org/wiki/Q2920921","display_name":"Management","level":1,"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/C188147891","wikidata":"https://www.wikidata.org/wiki/Q147638","display_name":"Cognitive science","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":4,"locations":[{"id":"doi:10.1162/tacl_a_00434","is_oa":true,"landing_page_url":"https://doi.org/10.1162/tacl_a_00434","pdf_url":"https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00434/1979270/tacl_a_00434.pdf","source":{"id":"https://openalex.org/S2729999759","display_name":"Transactions of the Association for Computational Linguistics","issn_l":"2307-387X","issn":["2307-387X"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Transactions of the Association for Computational Linguistics","raw_type":"journal-article"},{"id":"pmh:oai:doaj.org/article:7865d581bc554481bb1d3d28fe5f98e4","is_oa":true,"landing_page_url":"https://doaj.org/article/7865d581bc554481bb1d3d28fe5f98e4","pdf_url":null,"source":{"id":"https://openalex.org/S4306401280","display_name":"DOAJ (DOAJ: Directory of Open Access Journals)","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":"cc-by-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Transactions of the Association for Computational Linguistics, Vol 9, Pp 1408-1424 (2021)","raw_type":"article"},{"id":"pmh:oai:epub.ub.uni-muenchen.de:92231","is_oa":true,"landing_page_url":"http://nbn-resolving.de/urn:nbn:de:bvb:19-epub-92231-5","pdf_url":null,"source":{"id":"https://openalex.org/S2729999759","display_name":"Transactions of the Association for Computational Linguistics","issn_l":"2307-387X","issn":["2307-387X"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"Transactions of the Association for Computational Linguistics","raw_type":"info:eu-repo/semantics/article"},{"id":"doi:10.5282/ubm/epub.92231","is_oa":true,"landing_page_url":"https://doi.org/10.5282/ubm/epub.92231","pdf_url":null,"source":{"id":"https://openalex.org/S7407052039","display_name":"Universit\u00e4tsbibliothek der LMU","issn_l":null,"issn":[],"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article-journal"}],"best_oa_location":{"id":"doi:10.1162/tacl_a_00434","is_oa":true,"landing_page_url":"https://doi.org/10.1162/tacl_a_00434","pdf_url":"https://direct.mit.edu/tacl/article-pdf/doi/10.1162/tacl_a_00434/1979270/tacl_a_00434.pdf","source":{"id":"https://openalex.org/S2729999759","display_name":"Transactions of the Association for Computational Linguistics","issn_l":"2307-387X","issn":["2307-387X"],"is_oa":true,"is_in_doaj":true,"is_core":true,"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Transactions of the Association for Computational Linguistics","raw_type":"journal-article"},"sustainable_development_goals":[{"display_name":"Gender equality","id":"https://metadata.un.org/sdg/5","score":0.4000000059604645}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":79,"referenced_works":["https://openalex.org/W1566289585","https://openalex.org/W1614298861","https://openalex.org/W2483215953","https://openalex.org/W2493916176","https://openalex.org/W2511277317","https://openalex.org/W2559655401","https://openalex.org/W2789910672","https://openalex.org/W2889624842","https://openalex.org/W2893425640","https://openalex.org/W2949433733","https://openalex.org/W2962739339","https://openalex.org/W2962787423","https://openalex.org/W2963341956","https://openalex.org/W2963457723","https://openalex.org/W2963494889","https://openalex.org/W2963809228","https://openalex.org/W2965373594","https://openalex.org/W2971307358","https://openalex.org/W2972668795","https://openalex.org/W2973049837","https://openalex.org/W2979826702","https://openalex.org/W2995335514","https://openalex.org/W2997195635","https://openalex.org/W2997588435","https://openalex.org/W3019416653","https://openalex.org/W3024308166","https://openalex.org/W3034238904","https://openalex.org/W3034775979","https://openalex.org/W3034951181","https://openalex.org/W3034995113","https://openalex.org/W3035241006","https://openalex.org/W3037697022","https://openalex.org/W3044438666","https://openalex.org/W3082274269","https://openalex.org/W3085190015","https://openalex.org/W3088059392","https://openalex.org/W3097738726","https://openalex.org/W3098824823","https://openalex.org/W3098903812","https://openalex.org/W3100355250","https://openalex.org/W3105882417","https://openalex.org/W3111372071","https://openalex.org/W3111792198","https://openalex.org/W3116459227","https://openalex.org/W3119866685","https://openalex.org/W3125649664","https://openalex.org/W3133702157","https://openalex.org/W3153427360","https://openalex.org/W3155655882","https://openalex.org/W3172642864","https://openalex.org/W3176477796","https://openalex.org/W3184144760","https://openalex.org/W4212774754","https://openalex.org/W4233907442","https://openalex.org/W4241439391","https://openalex.org/W4248576335","https://openalex.org/W4287391717","https://openalex.org/W4288089799","https://openalex.org/W4385970303","https://openalex.org/W6636510571","https://openalex.org/W6721933647","https://openalex.org/W6727099177","https://openalex.org/W6748736812","https://openalex.org/W6755207826","https://openalex.org/W6760413209","https://openalex.org/W6761994077","https://openalex.org/W6766673545","https://openalex.org/W6768028577","https://openalex.org/W6769627184","https://openalex.org/W6771395532","https://openalex.org/W6771522293","https://openalex.org/W6778883912","https://openalex.org/W6782866170","https://openalex.org/W6787309020","https://openalex.org/W6787315605","https://openalex.org/W6788404175","https://openalex.org/W6788811087","https://openalex.org/W6788910479","https://openalex.org/W6790057370"],"related_works":["https://openalex.org/W4386875279","https://openalex.org/W4362554880","https://openalex.org/W4281684980","https://openalex.org/W2171721708","https://openalex.org/W3214527415","https://openalex.org/W4287887864","https://openalex.org/W1495104519","https://openalex.org/W4390963114","https://openalex.org/W4225584739","https://openalex.org/W2944691285"],"abstract_inverted_index":{"Abstract":[0],"\u26a0":[1],"This":[2],"paper":[3],"contains":[4],"prompts":[5],"and":[6,27,96],"model":[7,137],"outputs":[8],"that":[9,34],"are":[10],"offensive":[11],"in":[12,38,193],"nature.":[13],"When":[14],"trained":[15],"on":[16,112,152],"large,":[17],"unfiltered":[18],"crawls":[19],"from":[20,70],"the":[21,39,97,100,128,132,167,176],"Internet,":[22],"language":[23,86,136,179],"models":[24,53,87,180],"pick":[25],"up":[26],"reproduce":[28],"all":[29],"kinds":[30],"of":[31,56,99,127,134,178],"undesirable":[32,94],"biases":[33,95],"can":[35],"be":[36,189],"found":[37],"data:":[40],"They":[41],"often":[42],"generate":[43],"racist,":[44],"sexist,":[45],"violent,":[46],"or":[47,164],"otherwise":[48],"toxic":[49],"language.":[50],"As":[51],"large":[52],"require":[54,160],"millions":[55],"training":[57,162],"examples":[58],"to":[59,66,73,89,106,143,166,188],"achieve":[60],"good":[61],"performance,":[62],"it":[63,159],"is":[64],"difficult":[65],"completely":[67],"prevent":[68],"them":[69],"being":[71],"exposed":[72],"such":[74],"content.":[75],"In":[76],"this":[77,107,113,144,194],"paper,":[78],"we":[79,115,171,184],"first":[80],"demonstrate":[81],"a":[82,90,118,124,135],"surprising":[83],"finding:":[84],"Pretrained":[85],"recognize,":[88],"considerable":[91],"degree,":[92],"their":[93],"toxicity":[98],"content":[101],"they":[102],"produce.":[103],"We":[104,141],"refer":[105,142],"capability":[108],"as":[109,146],"self-diagnosis.":[110],"Based":[111],"finding,":[114],"then":[116],"propose":[117],"decoding":[119],"algorithm":[120],"that,":[121],"given":[122],"only":[123],"textual":[125],"description":[126],"undesired":[129],"behavior,":[130],"reduces":[131],"probability":[133],"producing":[138],"problematic":[139],"text.":[140],"approach":[145,187],"self-debiasing.":[147],"Self-debiasing":[148],"does":[149,158],"not":[150],"rely":[151],"manually":[153],"curated":[154],"word":[155],"lists,":[156],"nor":[157],"any":[161],"data":[163],"changes":[165],"model\u2019s":[168],"parameters.":[169],"While":[170],"by":[172],"no":[173],"means":[174],"eliminate":[175],"issue":[177],"generating":[181],"biased":[182],"text,":[183],"believe":[185],"our":[186],"an":[190],"important":[191],"step":[192],"direction.1":[195]},"counts_by_year":[{"year":2026,"cited_by_count":14},{"year":2025,"cited_by_count":41},{"year":2024,"cited_by_count":37},{"year":2023,"cited_by_count":72},{"year":2022,"cited_by_count":36},{"year":2021,"cited_by_count":17}],"updated_date":"2026-07-02T09:51:11.867554","created_date":"2025-10-10T00:00:00"}