{"id":"https://openalex.org/W7138044216","doi":"https://doi.org/10.1609/aaai.v40i26.39343","title":"Towards Acyclic Preference Evaluation of Language Models via Multiple Evaluators","display_name":"Towards Acyclic Preference Evaluation of Language Models via Multiple Evaluators","publication_year":2026,"publication_date":"2026-03-14","ids":{"openalex":"https://openalex.org/W7138044216","doi":"https://doi.org/10.1609/aaai.v40i26.39343"},"language":null,"primary_location":{"id":"doi:10.1609/aaai.v40i26.39343","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i26.39343","pdf_url":null,"source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://doi.org/10.1609/aaai.v40i26.39343","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5129691695","display_name":"Zhengyu Hu","orcid":null},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":true,"raw_author_name":"Zhengyu Hu","raw_affiliation_strings":["The Hong Kong University of Science and Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129663836","display_name":"Jieyu Zhang","orcid":null},"institutions":[{"id":"https://openalex.org/I201448701","display_name":"University of Washington","ror":"https://ror.org/00cvxb145","country_code":"US","type":"education","lineage":["https://openalex.org/I201448701"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Jieyu Zhang","raw_affiliation_strings":["University of Washington"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Washington","institution_ids":["https://openalex.org/I201448701"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129705772","display_name":"Zhihan Xiong","orcid":null},"institutions":[{"id":"https://openalex.org/I201448701","display_name":"University of Washington","ror":"https://ror.org/00cvxb145","country_code":"US","type":"education","lineage":["https://openalex.org/I201448701"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhihan Xiong","raw_affiliation_strings":["University of Washington"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Washington","institution_ids":["https://openalex.org/I201448701"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5076215229","display_name":"Alexander Ratner","orcid":null},"institutions":[{"id":"https://openalex.org/I201448701","display_name":"University of Washington","ror":"https://ror.org/00cvxb145","country_code":"US","type":"education","lineage":["https://openalex.org/I201448701"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Alexander Ratner","raw_affiliation_strings":["Department of Computer Science, University of Washington"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science, University of Washington","institution_ids":["https://openalex.org/I201448701"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129728581","display_name":"Kaize Ding","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kaize Ding","raw_affiliation_strings":["Northwestern University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Northwestern University","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5129643365","display_name":"Ranjay Krishna","orcid":null},"institutions":[{"id":"https://openalex.org/I201448701","display_name":"University of Washington","ror":"https://ror.org/00cvxb145","country_code":"US","type":"education","lineage":["https://openalex.org/I201448701"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ranjay Krishna","raw_affiliation_strings":["University of Washington"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Washington","institution_ids":["https://openalex.org/I201448701"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5129691695"],"corresponding_institution_ids":["https://openalex.org/I200769079"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.24139579,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"40","issue":"26","first_page":"21903","last_page":"21911"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.3490999937057495,"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":0.3490999937057495,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.12110000103712082,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.06920000165700912,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/leverage","display_name":"Leverage (statistics)","score":0.7099000215530396},{"id":"https://openalex.org/keywords/preference","display_name":"Preference","score":0.6051999926567078},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.48840001225471497},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.4683000147342682},{"id":"https://openalex.org/keywords/ranking","display_name":"Ranking (information retrieval)","score":0.4392000138759613},{"id":"https://openalex.org/keywords/construct","display_name":"Construct (python library)","score":0.391400009393692},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.3894999921321869}],"concepts":[{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.7099000215530396},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6639000177383423},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.6051999926567078},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5713000297546387},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5335000157356262},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.48840001225471497},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4683000147342682},{"id":"https://openalex.org/C189430467","wikidata":"https://www.wikidata.org/wiki/Q7293293","display_name":"Ranking (information retrieval)","level":2,"score":0.4392000138759613},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.391400009393692},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.3894999921321869},{"id":"https://openalex.org/C146849305","wikidata":"https://www.wikidata.org/wiki/Q370766","display_name":"Ground truth","level":2,"score":0.3873000144958496},{"id":"https://openalex.org/C132525143","wikidata":"https://www.wikidata.org/wiki/Q141488","display_name":"Graph","level":2,"score":0.373199999332428},{"id":"https://openalex.org/C93959086","wikidata":"https://www.wikidata.org/wiki/Q6888345","display_name":"Model selection","level":2,"score":0.37059998512268066},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.36090001463890076},{"id":"https://openalex.org/C181204326","wikidata":"https://www.wikidata.org/wiki/Q7239820","display_name":"Preference learning","level":3,"score":0.28999999165534973},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2897999882698059},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.28929999470710754},{"id":"https://openalex.org/C3018395757","wikidata":"https://www.wikidata.org/wiki/Q1379672","display_name":"Evaluation methods","level":2,"score":0.2750000059604645},{"id":"https://openalex.org/C24756922","wikidata":"https://www.wikidata.org/wiki/Q1757694","display_name":"Data quality","level":3,"score":0.250900000333786}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1609/aaai.v40i26.39343","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i26.39343","pdf_url":null,"source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"}],"best_oa_location":{"id":"doi:10.1609/aaai.v40i26.39343","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i26.39343","pdf_url":null,"source":{"id":"https://openalex.org/S4210191458","display_name":"Proceedings of the AAAI Conference on Artificial Intelligence","issn_l":"2159-5399","issn":["2159-5399","2374-3468"],"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310320058","host_organization_name":"Association for the Advancement of Artificial Intelligence","host_organization_lineage":["https://openalex.org/P4310320058"],"host_organization_lineage_names":["Association for the Advancement of Artificial Intelligence"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the AAAI Conference on Artificial Intelligence","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Despite":[0],"the":[1,28,111],"remarkable":[2],"success":[3],"of":[4],"Large":[5],"Language":[6],"Models":[7],"(LLMs),":[8],"evaluating":[9],"their":[10],"outputs'":[11],"quality":[12],"regarding":[13],"preference":[14,85,114],"remains":[15],"a":[16,24,36,75],"critical":[17],"challenge.":[18],"While":[19],"existing":[20],"works":[21],"usually":[22],"leverage":[23],"strong":[25,158],"LLM":[26,150],"as":[27],"judge":[29],"for":[30,94,103,131,136,143],"comparing":[31],"LLMs'":[32],"response":[33,134],"pairwisely,":[34],"such":[35],"single-evaluator":[37],"approach":[38,77],"is":[39,47,56],"vulnerable":[40],"to":[41,83,156],"cyclic":[42],"preference,":[43],"i.e.,":[44],"output":[45],"A":[46],"better":[48,57],"than":[49,52,58],"B,":[50],"B":[51],"C,":[53],"but":[54],"C":[55],"A,":[59],"causing":[60],"contradictory":[61],"evaluation":[62,97,167],"results.":[63,98],"To":[64],"address":[65],"this,":[66],"we":[67],"introduce":[68],"PGED":[69,122,147],"(Preference":[70],"Graph":[71],"Ensemble":[72],"and":[73,87,90,139,169],"Denoise),":[74],"novel":[76],"that":[78],"leverages":[79],"multiple":[80],"model-based":[81],"evaluators":[82,151],"construct":[84],"graphs,":[86],"then":[88],"ensembles":[89],"denoises":[91],"these":[92],"graphs":[93],"acyclic,":[95],"non-contradictory":[96],"We":[99],"provide":[100],"theoretical":[101],"guarantees":[102],"our":[104],"framework,":[105],"demonstrating":[106],"its":[107,163],"efficacy":[108],"in":[109,125,165],"recovering":[110],"ground":[112],"truth":[113],"structure.":[115],"Extensive":[116],"experiments":[117],"on":[118],"ten":[119],"benchmarks":[120],"demonstrate":[121],"'s":[123],"superiority":[124],"three":[126],"applications:":[127],"1)":[128],"model":[129,144,171],"ranking":[130],"evaluation,":[132],"2)":[133],"selection":[135,142],"test-time":[137],"scaling,":[138],"3)":[140],"data":[141],"fine-tuning.":[145],"Notably,":[146],"combines":[148],"small":[149],"(e.g.,":[152,160],"Llama3-8B,":[153],"Mistral-7B,":[154],"Qwen2-7B)":[155],"outperform":[157],"ones":[159],"Qwen2-72B),":[161],"showcasing":[162],"effectiveness":[164],"enhancing":[166],"reliability":[168],"improving":[170],"performance.":[172]},"counts_by_year":[],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2026-03-18T00:00:00"}