{"id":"https://openalex.org/W4412945073","doi":"https://doi.org/10.18653/v1/2025.acl-long.867","title":"Faithful and Robust LLM-Driven Theorem Proving for NLI Explanations","display_name":"Faithful and Robust LLM-Driven Theorem Proving for NLI Explanations","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412945073","doi":"https://doi.org/10.18653/v1/2025.acl-long.867"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.acl-long.867","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.867","pdf_url":"https://aclanthology.org/2025.acl-long.867.pdf","source":null,"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 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.acl-long.867.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Xin Quan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xin Quan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5032107191","display_name":"Marco Valentino","orcid":"https://orcid.org/0000-0002-9959-8385"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Marco Valentino","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5040721287","display_name":"Louise A. Dennis","orcid":"https://orcid.org/0000-0003-1426-1896"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Louise A. Dennis","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":null,"display_name":"Andre Freitas","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Andre Freitas","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":null,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"17734","last_page":"17755"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.98089998960495,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.98089998960495,"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/T11986","display_name":"Scientific Computing and Data Management","score":0.9345999956130981,"subfield":{"id":"https://openalex.org/subfields/1802","display_name":"Information Systems and Management"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision 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":0.9332000017166138,"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/computer-science","display_name":"Computer science","score":0.5706089735031128},{"id":"https://openalex.org/keywords/calculus","display_name":"Calculus (dental)","score":0.4435029923915863},{"id":"https://openalex.org/keywords/automated-theorem-proving","display_name":"Automated theorem proving","score":0.4328486919403076},{"id":"https://openalex.org/keywords/theoretical-computer-science","display_name":"Theoretical computer science","score":0.33514684438705444},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.320201575756073},{"id":"https://openalex.org/keywords/medicine","display_name":"Medicine","score":0.08913359045982361}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5706089735031128},{"id":"https://openalex.org/C2777686260","wikidata":"https://www.wikidata.org/wiki/Q144037","display_name":"Calculus (dental)","level":2,"score":0.4435029923915863},{"id":"https://openalex.org/C206880738","wikidata":"https://www.wikidata.org/wiki/Q431667","display_name":"Automated theorem proving","level":2,"score":0.4328486919403076},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.33514684438705444},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.320201575756073},{"id":"https://openalex.org/C71924100","wikidata":"https://www.wikidata.org/wiki/Q11190","display_name":"Medicine","level":0,"score":0.08913359045982361},{"id":"https://openalex.org/C199343813","wikidata":"https://www.wikidata.org/wiki/Q12128","display_name":"Dentistry","level":1,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.18653/v1/2025.acl-long.867","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.867","pdf_url":"https://aclanthology.org/2025.acl-long.867.pdf","source":null,"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 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},{"id":"pmh:oai:eprints.whiterose.ac.uk:231199","is_oa":false,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4306400854","display_name":"White Rose Research Online (University of Leeds, The University of Sheffield, University of York)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I2800616092","host_organization_name":"White Rose University Consortium","host_organization_lineage":["https://openalex.org/I2800616092"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"acceptedVersion","is_accepted":true,"is_published":false,"raw_source_name":null,"raw_type":"Proceedings Paper"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.acl-long.867","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.867","pdf_url":"https://aclanthology.org/2025.acl-long.867.pdf","source":null,"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 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G6661291015","display_name":"EnnCore: End-to-End Conceptual Guarding of Neural Architectures","funder_award_id":"EP/T026995/1","funder_id":"https://openalex.org/F4320334627","funder_display_name":"Engineering and Physical Sciences Research Council"},{"id":"https://openalex.org/G8579816056","display_name":null,"funder_award_id":"200021_204617","funder_id":"https://openalex.org/F4320320924","funder_display_name":"Schweizerischer Nationalfonds zur F\u00f6rderung der Wissenschaftlichen Forschung"}],"funders":[{"id":"https://openalex.org/F4320319985","display_name":"Cancer Research UK","ror":"https://ror.org/054225q67"},{"id":"https://openalex.org/F4320319990","display_name":"National Institute for Health and Care Research","ror":"https://ror.org/0187kwz08"},{"id":"https://openalex.org/F4320320924","display_name":"Schweizerischer Nationalfonds zur F\u00f6rderung der Wissenschaftlichen Forschung","ror":"https://ror.org/00yjd3n13"},{"id":"https://openalex.org/F4320333687","display_name":"Manchester Biomedical Research Centre","ror":null},{"id":"https://openalex.org/F4320334627","display_name":"Engineering and Physical Sciences Research Council","ror":"https://ror.org/0439y7842"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412945073.pdf","grobid_xml":"https://content.openalex.org/works/W4412945073.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W4391375266","https://openalex.org/W1979597421","https://openalex.org/W2007980826","https://openalex.org/W2061531152","https://openalex.org/W3002753104","https://openalex.org/W2077600819","https://openalex.org/W2142036596","https://openalex.org/W2072657027"],"abstract_inverted_index":{"Natural":[0,8],"language":[1,27,48],"explanations":[2],"play":[3],"a":[4,53],"fundamental":[5],"role":[6],"in":[7,72,84,122,133,168],"Language":[9],"Inference":[10],"(NLI)":[11],"by":[12,69],"revealing":[13],"how":[14],"premises":[15],"logically":[16],"entail":[17],"hypotheses.Recent":[18],"work":[19],"has":[20],"shown":[21],"that":[22,55,161,183],"the":[23,39,57,162,178,187,197],"interaction":[24],"of":[25,41,59,143,199],"large":[26],"models":[28],"(LLMs)":[29],"with":[30,77],"theorem":[31],"provers":[32],"(TPs)":[33],"can":[34,191],"help":[35],"verify":[36],"and":[37,63,89,102,118,138,155,172],"improve":[38,193],"validity":[40],"NLI":[42],"explanations.However,":[43],"TPs":[44],"require":[45],"translating":[46],"natural":[47],"into":[49],"machine-verifiable":[50],"formal":[51,94],"representations,":[52,124],"process":[54],"introduces":[56],"risk":[58],"semantic":[60,111],"information":[61],"loss":[62,112],"unfaithful":[64],"interpretation,":[65],"an":[66],"issue":[67],"compounded":[68],"LLMs'":[70,141],"challenges":[71],"capturing":[73],"critical":[74],"logical":[75,123,128],"structures":[76],"sufficient":[78],"precision.Moreover,":[79],"LLMs":[80,132,159],"are":[81],"still":[82],"limited":[83],"their":[85],"capacity":[86,142],"for":[87,147,202],"rigorous":[88],"robust":[90],"proof":[91,136],"construction":[92],"within":[93],"verification":[95],"frameworks.To":[96],"mitigate":[97],"issues":[98],"related":[99],"to":[100,108,130],"faithfulness":[101],"robustness,":[103],"this":[104],"paper":[105],"investigates":[106],"strategies":[107,164],"(1)":[109],"alleviate":[110],"during":[113],"autoformalisation,":[114],"(2)":[115],"efficiently":[116],"identify":[117],"correct":[119],"syntactic":[120],"errors":[121],"(3)":[125],"explicitly":[126],"use":[127],"expressions":[129],"guide":[131],"generating":[134],"structured":[135],"sketches,":[137],"(4)":[139],"increase":[140],"interpreting":[144],"TP's":[145],"feedback":[146],"iterative":[148],"refinement.Our":[149],"empirical":[150],"results":[151],"on":[152,186],"e-SNLI,":[153],"QASC":[154],"WorldTree":[156],"using":[157],"different":[158],"demonstrate":[160],"proposed":[163],"yield":[165],"significant":[166],"improvements":[167],"autoformalisation":[169],"(+18.46%,+34.2%,":[170],"+39.77%)":[171],"explanation":[173],"refinement":[174],"(+29.5%,+51.5%,":[175],"+41.25%)":[176],"over":[177],"state-of-the-art":[179],"model.Moreover,":[180],"we":[181],"show":[182],"specific":[184],"interventions":[185],"hybrid":[188],"LLM-TP":[189],"architecture":[190],"substantially":[192],"efficiency,":[194],"drastically":[195],"reducing":[196],"number":[198],"iterations":[200],"required":[201],"successful":[203],"verification.":[204],"1":[205]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}