{"id":"https://openalex.org/W4395064977","doi":"https://doi.org/10.48550/arxiv.2404.13390","title":"Explanation based Bias Decoupling Regularization for Natural Language Inference","display_name":"Explanation based Bias Decoupling Regularization for Natural Language Inference","publication_year":2024,"publication_date":"2024-04-20","ids":{"openalex":"https://openalex.org/W4395064977","doi":"https://doi.org/10.48550/arxiv.2404.13390"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2404.13390","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2404.13390","pdf_url":"https://arxiv.org/pdf/2404.13390","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"text"},"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2404.13390","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5108583753","display_name":"Jianxiang Zang","orcid":"https://orcid.org/0000-0003-0236-8215"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zang, Jianxiang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5055455306","display_name":"Hui Liu","orcid":"https://orcid.org/0000-0002-8166-0216"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Hui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":1,"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/T10028","display_name":"Topic Modeling","score":0.6722000241279602,"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":0.6722000241279602,"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/regularization","display_name":"Regularization (linguistics)","score":0.6500924825668335},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6130589842796326},{"id":"https://openalex.org/keywords/decoupling","display_name":"Decoupling (probability)","score":0.6110180020332336},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.4894239008426666},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.41589099168777466},{"id":"https://openalex.org/keywords/econometrics","display_name":"Econometrics","score":0.3882589340209961},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.3668772578239441},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.2944685220718384},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.07053294777870178},{"id":"https://openalex.org/keywords/control-engineering","display_name":"Control engineering","score":0.054811716079711914}],"concepts":[{"id":"https://openalex.org/C2776135515","wikidata":"https://www.wikidata.org/wiki/Q17143721","display_name":"Regularization (linguistics)","level":2,"score":0.6500924825668335},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6130589842796326},{"id":"https://openalex.org/C205606062","wikidata":"https://www.wikidata.org/wiki/Q5249645","display_name":"Decoupling (probability)","level":2,"score":0.6110180020332336},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4894239008426666},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.41589099168777466},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.3882589340209961},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3668772578239441},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2944685220718384},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.07053294777870178},{"id":"https://openalex.org/C133731056","wikidata":"https://www.wikidata.org/wiki/Q4917288","display_name":"Control engineering","level":1,"score":0.054811716079711914}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2404.13390","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2404.13390","pdf_url":"https://arxiv.org/pdf/2404.13390","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"text"},{"id":"doi:10.48550/arxiv.2404.13390","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2404.13390","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2404.13390","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2404.13390","pdf_url":"https://arxiv.org/pdf/2404.13390","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"text"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4395064977.pdf","grobid_xml":"https://content.openalex.org/works/W4395064977.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4313561376","https://openalex.org/W3103825105","https://openalex.org/W3027880158","https://openalex.org/W1617565119","https://openalex.org/W160381218","https://openalex.org/W2512958550","https://openalex.org/W2004102934","https://openalex.org/W4366455350","https://openalex.org/W2329266651","https://openalex.org/W4321844193"],"abstract_inverted_index":{"The":[0],"robustness":[1],"of":[2,37,121,174],"Transformer-based":[3,159],"Natural":[4],"Language":[5],"Inference":[6],"encoders":[7,129,160],"is":[8],"frequently":[9],"compromised":[10],"as":[11,110,149],"they":[12],"tend":[13],"to":[14,30,80,83,115,130,161],"rely":[15],"more":[16],"on":[17,21,50,134],"dataset":[18],"biases":[19,163],"than":[20],"the":[22,35,41,59,85,113,138,145],"intended":[23],"task-relevant":[24,139],"features.":[25],"Recent":[26],"studies":[27],"have":[28],"attempted":[29],"mitigate":[31],"this":[32,76],"by":[33],"reducing":[34],"weight":[36],"biased":[38,55,60],"samples":[39,53],"during":[40,141],"training":[42],"process.":[43],"However,":[44],"these":[45],"debiasing":[46],"methods":[47,171],"primarily":[48],"focus":[49,133],"identifying":[51],"which":[52],"are":[54],"without":[56],"explicitly":[57],"determining":[58],"components":[61],"within":[62],"each":[63],"case.":[64],"This":[65,126],"limitation":[66],"restricts":[67],"those":[68],"methods'":[69],"capability":[70],"in":[71,89,172],"out-of-distribution":[72,175],"inference.":[73],"To":[74],"address":[75],"issue,":[77],"we":[78],"aim":[79],"train":[81],"models":[82],"adopt":[84],"logic":[86],"humans":[87],"use":[88],"explaining":[90],"causality.":[91],"We":[92],"propose":[93],"a":[94,117,165],"simple,":[95],"comprehensive,":[96],"and":[97,124,132],"interpretable":[98],"method:":[99],"Explanation":[100],"based":[101],"Bias":[102],"Decoupling":[103,123],"Regularization":[104],"(EBD-Reg).":[105],"EBD-Reg":[106,155],"employs":[107],"human":[108],"explanations":[109],"criteria,":[111],"guiding":[112],"encoder":[114],"establish":[116],"tripartite":[118],"parallel":[119],"supervision":[120],"Distinguishing,":[122],"Aligning.":[125],"method":[127],"enables":[128],"identify":[131],"keywords":[135],"that":[136,154],"represent":[137],"features":[140],"inference,":[142],"while":[143],"discarding":[144],"residual":[146],"elements":[147],"acting":[148],"biases.":[150],"Empirical":[151],"evidence":[152],"underscores":[153],"effectively":[156],"guides":[157],"various":[158],"decouple":[162],"through":[164],"human-centric":[166],"lens,":[167],"significantly":[168],"surpassing":[169],"other":[170],"terms":[173],"inference":[176],"capabilities.":[177]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2024-04-24T00:00:00"}