{"id":"https://openalex.org/W7168362381","doi":"https://doi.org/10.1145/3805712.3808508","title":"CSMAD: Hallucination Detection via Multi-Agent Debate with NLI-Verified Contradictory Statements","display_name":"CSMAD: Hallucination Detection via Multi-Agent Debate with NLI-Verified Contradictory Statements","publication_year":2026,"publication_date":"2026-07-15","ids":{"openalex":"https://openalex.org/W7168362381","doi":"https://doi.org/10.1145/3805712.3808508"},"language":null,"primary_location":{"id":"doi:10.1145/3805712.3808508","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3805712.3808508","pdf_url":null,"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 49th International ACM SIGIR Conference on Research and Development in Information Retrieval","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3805712.3808508","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5140853045","display_name":"Swapnil Gupta","orcid":"https://orcid.org/0000-0002-6095-2957"},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Swapnil Gupta","raw_affiliation_strings":["Amazon, Gurugram, India"],"raw_orcid":"https://orcid.org/0000-0002-6095-2957","affiliations":[{"raw_affiliation_string":"Amazon, Gurugram, India","institution_ids":["https://openalex.org/I1311688040"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140813743","display_name":"Akshay Verma","orcid":"https://orcid.org/0009-0006-7718-2705"},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Akshay Verma","raw_affiliation_strings":["Amazon, Gurugram, India"],"raw_orcid":"https://orcid.org/0009-0006-7718-2705","affiliations":[{"raw_affiliation_string":"Amazon, Gurugram, India","institution_ids":["https://openalex.org/I1311688040"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140783526","display_name":"Khushi Gupta","orcid":"https://orcid.org/0009-0000-9252-4407"},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Khushi Gupta","raw_affiliation_strings":["Amazon, Gurugram, India"],"raw_orcid":"https://orcid.org/0009-0000-9252-4407","affiliations":[{"raw_affiliation_string":"Amazon, Gurugram, India","institution_ids":["https://openalex.org/I1311688040"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140760674","display_name":"Prateek Sircar","orcid":"https://orcid.org/0000-0002-9152-9238"},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Prateek Sircar","raw_affiliation_strings":["Amazon, Gurugram, India"],"raw_orcid":"https://orcid.org/0000-0002-9152-9238","affiliations":[{"raw_affiliation_string":"Amazon, Gurugram, India","institution_ids":["https://openalex.org/I1311688040"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5140815186","display_name":"Deepak Gupta","orcid":"https://orcid.org/0000-0003-0972-1317"},"institutions":[{"id":"https://openalex.org/I1311688040","display_name":"Amazon (United States)","ror":"https://ror.org/04mv4n011","country_code":"US","type":"company","lineage":["https://openalex.org/I1311688040"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Deepak Gupta","raw_affiliation_strings":["Amazon, Gurugram, India"],"raw_orcid":"https://orcid.org/0000-0003-0972-1317","affiliations":[{"raw_affiliation_string":"Amazon, Gurugram, India","institution_ids":["https://openalex.org/I1311688040"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I1311688040"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"4627","last_page":"4632"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":null,"topics":[],"keywords":[{"id":"https://openalex.org/keywords/meaning","display_name":"Meaning (existential)","score":0.2881999909877777},{"id":"https://openalex.org/keywords/identification","display_name":"Identification (biology)","score":0.27950000762939453},{"id":"https://openalex.org/keywords/subject","display_name":"Subject (documents)","score":0.26759999990463257},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.25440001487731934}],"concepts":[{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.4424000084400177},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.3718999922275543},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.3118000030517578},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.3095000088214874},{"id":"https://openalex.org/C2780876879","wikidata":"https://www.wikidata.org/wiki/Q3054749","display_name":"Meaning (existential)","level":2,"score":0.2881999909877777},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.28630000352859497},{"id":"https://openalex.org/C77805123","wikidata":"https://www.wikidata.org/wiki/Q161272","display_name":"Social psychology","level":1,"score":0.2840000092983246},{"id":"https://openalex.org/C116834253","wikidata":"https://www.wikidata.org/wiki/Q2039217","display_name":"Identification (biology)","level":2,"score":0.27950000762939453},{"id":"https://openalex.org/C2777855551","wikidata":"https://www.wikidata.org/wiki/Q12310021","display_name":"Subject (documents)","level":2,"score":0.26759999990463257},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.25440001487731934},{"id":"https://openalex.org/C95457728","wikidata":"https://www.wikidata.org/wiki/Q309","display_name":"History","level":0,"score":0.25189998745918274}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3805712.3808508","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3805712.3808508","pdf_url":null,"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 49th International ACM SIGIR Conference on Research and Development in Information Retrieval","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3805712.3808508","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3805712.3808508","pdf_url":null,"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 49th International ACM SIGIR Conference on Research and Development in Information Retrieval","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":13,"referenced_works":["https://openalex.org/W2946659172","https://openalex.org/W3099977667","https://openalex.org/W4287887788","https://openalex.org/W4375949262","https://openalex.org/W4389518962","https://openalex.org/W4389519928","https://openalex.org/W4402672029","https://openalex.org/W4402672090","https://openalex.org/W4402683414","https://openalex.org/W4404782209","https://openalex.org/W4411120368","https://openalex.org/W7140093991","https://openalex.org/W7140101146"],"related_works":[],"abstract_inverted_index":{"Large":[0],"Language":[1,149],"Models":[2],"(LLMs)":[3],"are":[4,106],"prone":[5],"to":[6,79,173],"hallucinations,":[7],"producing":[8],"fluent":[9],"but":[10],"factually":[11],"incorrect":[12,44],"statements.":[13],"Recent":[14],"multi-agent":[15,60],"debate":[16],"methods":[17],"improve":[18],"hallucination":[19],"detection":[20],"by":[21,66,217,228],"jointly":[22],"improving":[23,215],"reasoning":[24,92],"and":[25,83,115,184,210,219],"decision-making.":[26],"However,":[27],"existing":[28],"approaches":[29],"either":[30,107],"collaborate":[31],"which":[32,87],"amplifies":[33],"shared":[34],"overconfidence,":[35],"or":[36,109],"adopt":[37],"adversarial":[38],"preset":[39,95],"stances,":[40],"that":[41,62,154,198],"can":[42],"inject":[43],"information":[45],"complicating":[46],"decision":[47],"making.":[48],"To":[49,139],"address":[50],"this,":[51],"we":[52,145,196],"propose":[53],"Contradictory":[54],"Statement":[55],"Multi-Agent":[56],"Debate":[57],"(CSMAD),":[58],"a":[59,68,147,191],"framework":[61],"creates":[63],"structured":[64],"disagreement":[65],"generating":[67],"contradictory":[69,85,104,141],"claim":[70,82,186],"for":[71,181,205],"each":[72],"input":[73],"claim.":[74],"CSMAD":[75,199],"asks":[76],"independent":[77],"agents":[78,112],"evaluate":[80],"the":[81,84,98,103,111,128,131,157,162,169,202],"claim,":[86,133],"encourages":[88],"different":[89],"lines":[90],"of":[91,130],"without":[93],"assigning":[94],"stances.":[96],"When":[97],"outcome":[99],"is":[100],"non-discriminative;":[101],"both":[102,135,206],"statements":[105],"accepted":[108],"rejected;":[110],"exchange":[113],"rationales":[114],"update":[116],"their":[117],"judgments":[118],"after":[119],"considering":[120],"opposing":[121],"evidence.":[122],"A":[123],"final":[124],"judge":[125],"then":[126],"decides":[127],"truth":[129],"original":[132,163],"using":[134],"arguments":[136],"as":[137,188,190],"context.":[138],"make":[140],"statement":[142,159],"generation":[143],"reliable,":[144],"add":[146],"Natural":[148],"Inference":[150],"(NLI)":[151],"based":[152],"verifier":[153],"checks":[155],"whether":[156],"generated":[158],"actually":[160],"contradicts":[161],"claim;":[164],"if":[165],"it":[166],"does":[167],"not,":[168],"system":[170],"falls":[171],"back":[172],"an":[174],"explicit":[175],"negation-based":[176],"contradiction.":[177],"Across":[178],"public":[179],"benchmarks":[180],"question":[182],"answering":[183],"scientific":[185],"verification,":[187],"well":[189],"proprietary":[192],"e-commerce":[193],"claims":[194],"dataset,":[195],"show":[197],"consistently":[200],"outperforms":[201],"strongest":[203],"baseline":[204],"large":[207],"(Claude-3.5":[208],"Sonnet)":[209],"medium-sized":[211],"(Qwen3-8B)":[212],"language":[213],"models,":[214],"F1":[216],"+2.3":[218],"+4.1":[220],"points,":[221],"respectively,":[222],"while":[223],"reducing":[224],"LLM":[225],"token":[226],"cost":[227],"28%.":[229]},"counts_by_year":[],"updated_date":"2026-07-17T05:52:16.776730","created_date":"2026-07-16T00:00:00"}