{"id":"https://openalex.org/W7166867861","doi":"https://doi.org/10.18653/v1/2026.acl-long.2006","title":"Simple Agents, Biased Judges: Efficient Multi-Party Dialogue Generation & The Evaluation Gap","display_name":"Simple Agents, Biased Judges: Efficient Multi-Party Dialogue Generation & The Evaluation Gap","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166867861","doi":"https://doi.org/10.18653/v1/2026.acl-long.2006"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.acl-long.2006","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.2006","pdf_url":"https://aclanthology.org/2026.acl-long.2006.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 64th 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/2026.acl-long.2006.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139790577","display_name":"Kunal Samanta","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kunal Samanta","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5064003601","display_name":"Faisal Tareque Shohan","orcid":"https://orcid.org/0000-0002-3872-5758"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Faisal Tareque Shohan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5038712780","display_name":"Amine Trabelsi","orcid":"https://orcid.org/0000-0002-1852-5265"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Amine Trabelsi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5065888080","display_name":"Richard Khoury","orcid":"https://orcid.org/0000-0001-8237-4548"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Richard Khoury","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":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.85676668,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"43326","last_page":"43345"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12031","display_name":"Speech and dialogue systems","score":0.7954000234603882,"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/T12031","display_name":"Speech and dialogue systems","score":0.7954000234603882,"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":0.03799999877810478,"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"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.03189999982714653,"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/simple","display_name":"Simple (philosophy)","score":0.5361999869346619},{"id":"https://openalex.org/keywords/measure","display_name":"Measure (data warehouse)","score":0.28870001435279846},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.26980000734329224},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.2685999870300293},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.25619998574256897}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5511999726295471},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.5361999869346619},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3400000035762787},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3278999924659729},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.290800005197525},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.28870001435279846},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.26980000734329224},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.2685999870300293},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.25619998574256897},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.24719999730587006}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.acl-long.2006","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.2006","pdf_url":"https://aclanthology.org/2026.acl-long.2006.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 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.acl-long.2006","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.2006","pdf_url":"https://aclanthology.org/2026.acl-long.2006.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 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G2614460161","display_name":null,"funder_award_id":"-2022","funder_id":"https://openalex.org/F4320334593","funder_display_name":"Natural Sciences and Engineering Research Council of Canada"}],"funders":[{"id":"https://openalex.org/F4320334593","display_name":"Natural Sciences and Engineering Research Council of Canada","ror":"https://ror.org/01h531d29"},{"id":"https://openalex.org/F4320334841","display_name":"Fonds de recherche du Qu\u00e9bec \u2013 Nature et technologies","ror":"https://ror.org/00b9f9778"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166867861.pdf","grobid_xml":"https://content.openalex.org/works/W7166867861.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Multi-party":[0],"social":[1,44,69,203],"dialogue":[2,24,55,80],"remains":[3],"underexplored":[4],"in":[5,8,43,68,201],"the":[6,12,134],"literature,":[7],"part":[9],"due":[10],"to":[11,118,175],"difficulty":[13],"and":[14,31,52,66,82,100,115,178,198],"cost":[15],"of":[16,77],"evaluation.As":[17],"a":[18,50,141,153],"result,":[19],"recent":[20],"work":[21],"on":[22,28],"synthetic":[23,213],"generation":[25,56,65],"often":[26],"relies":[27],"automated":[29,101],"metrics":[30],"LLM-as-a-Judge":[32],"frameworks,":[33],"despite":[34],"limited":[35],"evidence":[36,183],"that":[37,146,161,184],"such":[38],"judges":[39,102,148,162,188],"reflect":[40,192],"human":[41,75,85,193],"preferences":[42,194],"settings.In":[45],"this":[46,71,137],"work,":[47],"we":[48,73,94,139],"introduce":[49,140],"lightweight":[51],"controllable":[53],"multi-party":[54,79,202],"framework":[57],"(MPOD)":[58],"as":[59],"an":[60],"experimental":[61],"instrument":[62],"for":[63,195,211],"studying":[64],"evaluation":[67],"interaction.Using":[70],"framework,":[72],"conduct":[74],"evaluations":[76],"opendomain":[78],"simulation":[81],"directly":[83],"compare":[84],"judgments":[86],"against":[87],"stateof-the-art":[88],"LLM":[89,147,187],"judges.Across":[90],"319":[91],"pairwise":[92],"comparisons,":[93],"observe":[95],"near-random":[96],"agreement":[97,123],"between":[98],"humans":[99],"(Cohen's":[103],"\u03ba":[104],"\u2248":[105],"0.11),":[106],"driven":[107],"by":[108],"systematic":[109],"behaviors":[110],"including":[111],"extreme":[112],"tie":[113],"aversion":[114],"strong":[116],"sensitivity":[117],"assistant-style":[119],"verbosity.Crucially,":[120],"human-human":[121],"inter-annotator":[122],"(\u03ba":[124],"=":[125],"0.29)":[126],"is":[127,170],"substantially":[128],"higher":[129],"than":[130],"human-LLM":[131],"agreement.To":[132],"isolate":[133],"mechanism":[135],"underlying":[136],"misalignment,":[138],"controlled":[142,182],"Transplant":[143],"Ablation,":[144],"showing":[145],"consistently":[149],"prefer":[150,163],"conversations":[151,165],"containing":[152],"single":[154],"proprietary,":[155],"assistantstyle":[156],"agent.Additional":[157],"stress":[158],"tests":[159],"show":[160],"GPT-style":[164],"even":[166],"when":[167],"utterance":[168],"order":[169],"randomly":[171],"shuffled,":[172],"indicating":[173],"insensitivity":[174],"conversational":[176,214],"structure":[177],"coherence.Our":[179],"findings":[180],"provide":[181],"current":[185],"instruction-tuned":[186],"do":[189],"not":[190],"reliably":[191],"naturalness,":[196],"engagingness,":[197],"overall":[199],"quality":[200],"dialogue,":[204],"calling":[205],"into":[206],"question":[207],"their":[208],"widespread":[209],"use":[210],"validating":[212],"data.":[215]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}