{"id":"https://openalex.org/W4411446012","doi":"https://doi.org/10.1109/tai.2025.3581507","title":"Latent Variable Modeling for Controllable and Diverse Generation From Large Language Models","display_name":"Latent Variable Modeling for Controllable and Diverse Generation From Large Language Models","publication_year":2025,"publication_date":"2025-06-19","ids":{"openalex":"https://openalex.org/W4411446012","doi":"https://doi.org/10.1109/tai.2025.3581507"},"language":"en","primary_location":{"id":"doi:10.1109/tai.2025.3581507","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tai.2025.3581507","pdf_url":null,"source":{"id":"https://openalex.org/S4210169448","display_name":"IEEE Transactions on Artificial Intelligence","issn_l":"2691-4581","issn":["2691-4581"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Artificial Intelligence","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Jianfei Zhang","orcid":"https://orcid.org/0009-0005-3212-8871"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Jianfei Zhang","raw_affiliation_strings":["School of Computer Science and Engineering, Beihang University, Beijing, China"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Beihang University, Beijing, China","institution_ids":["https://openalex.org/I82880672"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Bei Li","orcid":"https://orcid.org/0000-0001-7617-9041"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bei Li","raw_affiliation_strings":["Meituan Inc., Beijing, China"],"affiliations":[{"raw_affiliation_string":"Meituan Inc., Beijing, China","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Zhuofan Chen","orcid":"https://orcid.org/0009-0000-5649-2751"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhuofan Chen","raw_affiliation_strings":["School of Computer Science and Engineering, Beihang University, Beijing, China"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Beihang University, Beijing, China","institution_ids":["https://openalex.org/I82880672"]}]},{"author_position":"middle","author":{"id":null,"display_name":"Chang Liu","orcid":"https://orcid.org/0009-0000-8324-7153"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chang Liu","raw_affiliation_strings":["School of Computer Science and Engineering, Beihang University, Beijing, China"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Beihang University, Beijing, China","institution_ids":["https://openalex.org/I82880672"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5092339687","display_name":"Chen Li","orcid":null},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Chen Li","raw_affiliation_strings":["School of Computer Science and Engineering, Beihang University, Beijing, China"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Beihang University, Beijing, China","institution_ids":["https://openalex.org/I82880672"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024599321","display_name":"Chenghua Lin","orcid":"https://orcid.org/0000-0003-3454-2468"},"institutions":[{"id":"https://openalex.org/I28407311","display_name":"University of Manchester","ror":"https://ror.org/027m9bs27","country_code":"GB","type":"education","lineage":["https://openalex.org/I28407311"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Chenghua Lin","raw_affiliation_strings":["Department of Computer Science, University of Manchester, Manchester, U.K","Department of Computer Science, University of Manchester, UK"],"affiliations":[{"raw_affiliation_string":"Department of Computer Science, University of Manchester, Manchester, U.K","institution_ids":["https://openalex.org/I28407311"]},{"raw_affiliation_string":"Department of Computer Science, University of Manchester, UK","institution_ids":["https://openalex.org/I28407311"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5055420596","display_name":"Wenge Rong","orcid":"https://orcid.org/0000-0002-4229-7215"},"institutions":[{"id":"https://openalex.org/I82880672","display_name":"Beihang University","ror":"https://ror.org/00wk2mp56","country_code":"CN","type":"education","lineage":["https://openalex.org/I82880672"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wenge Rong","raw_affiliation_strings":["School of Computer Science and Engineering, Beihang University, Beijing, China"],"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering, Beihang University, Beijing, China","institution_ids":["https://openalex.org/I82880672"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":7,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I82880672"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.0671786,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"7","issue":"2","first_page":"791","last_page":"805"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9718999862670898,"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":0.9718999862670898,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.9657999873161316,"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/latent-variable","display_name":"Latent variable","score":0.7223753929138184},{"id":"https://openalex.org/keywords/variable","display_name":"Variable (mathematics)","score":0.5915532112121582},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.4906965494155884},{"id":"https://openalex.org/keywords/latent-variable-model","display_name":"Latent variable model","score":0.4433170557022095},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.2440873682498932},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.16080042719841003}],"concepts":[{"id":"https://openalex.org/C51167844","wikidata":"https://www.wikidata.org/wiki/Q4422623","display_name":"Latent variable","level":2,"score":0.7223753929138184},{"id":"https://openalex.org/C182365436","wikidata":"https://www.wikidata.org/wiki/Q50701","display_name":"Variable (mathematics)","level":2,"score":0.5915532112121582},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4906965494155884},{"id":"https://openalex.org/C65965080","wikidata":"https://www.wikidata.org/wiki/Q1806885","display_name":"Latent variable model","level":3,"score":0.4433170557022095},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2440873682498932},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.16080042719841003},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/tai.2025.3581507","is_oa":false,"landing_page_url":"https://doi.org/10.1109/tai.2025.3581507","pdf_url":null,"source":{"id":"https://openalex.org/S4210169448","display_name":"IEEE Transactions on Artificial Intelligence","issn_l":"2691-4581","issn":["2691-4581"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Artificial Intelligence","raw_type":"journal-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G4558726929","display_name":null,"funder_award_id":"62477001","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":35,"referenced_works":["https://openalex.org/W2101105183","https://openalex.org/W2194775991","https://openalex.org/W2250966211","https://openalex.org/W2251939518","https://openalex.org/W2606974598","https://openalex.org/W2896457183","https://openalex.org/W2924334974","https://openalex.org/W2962717182","https://openalex.org/W2962834855","https://openalex.org/W2963206148","https://openalex.org/W2963223306","https://openalex.org/W2963456134","https://openalex.org/W2963567641","https://openalex.org/W2964120993","https://openalex.org/W2964352131","https://openalex.org/W2970682219","https://openalex.org/W3035100081","https://openalex.org/W3035451444","https://openalex.org/W3036394672","https://openalex.org/W3098708719","https://openalex.org/W3116243435","https://openalex.org/W3129831491","https://openalex.org/W3152950858","https://openalex.org/W3191529106","https://openalex.org/W3194782062","https://openalex.org/W4221155857","https://openalex.org/W4225107720","https://openalex.org/W4225824663","https://openalex.org/W4253028564","https://openalex.org/W4287888031","https://openalex.org/W4385245566","https://openalex.org/W4385572077","https://openalex.org/W4385572481","https://openalex.org/W4389520135","https://openalex.org/W4404783034"],"related_works":["https://openalex.org/W2461917396","https://openalex.org/W2037497866","https://openalex.org/W4243467573","https://openalex.org/W62001224","https://openalex.org/W1502435251","https://openalex.org/W3032390039","https://openalex.org/W1584341211","https://openalex.org/W3122667150","https://openalex.org/W3145681568","https://openalex.org/W4393387622"],"abstract_inverted_index":{"Conditional":[0],"Variational":[1],"Auto-Encoders":[2],"(CVAEs)":[3],"represent":[4],"a":[5,93,104,159],"powerful":[6],"deep":[7],"generative":[8],"framework,":[9],"utilizing":[10],"latent":[11,59,86,99,134],"variables":[12],"(explicitly":[13],"modelled":[14],"hidden":[15],"states)":[16],"to":[17,69],"capture":[18],"underlying":[19],"factors":[20],"and":[21,48,102,123,136,163],"govern":[22],"the":[23,34,66,127,141,155],"generation":[24,112,116,139,153],"process":[25],"accordingly.":[26],"However,":[27],"this":[28,62],"idea":[29],"is":[30],"less":[31],"explored":[32],"in":[33,43,55,132,151],"era":[35],"of":[36,129,143,161,167],"Large":[37],"Language":[38],"Models":[39],"(LLMs),":[40],"facing":[41],"challenges":[42,54],"structural":[44],"differences":[45],"between":[46],"LLMs":[47,72,80],"traditional":[49],"CVAEs":[50],"as":[51,53],"well":[52],"posterior":[56],"collapse":[57],"(homogeneous":[58],"variables).":[60],"In":[61],"work,":[63],"we":[64,91,125,146],"present":[65],"first":[67],"attempt":[68],"extend":[70],"decoder-only":[71],"into":[73],"encoder-decoder":[74],"CVAEs,":[75],"aimming":[76],"at":[77],"enhancing":[78],"existing":[79],"with":[81],"flexible":[82],"control":[83],"via":[84],"low-dimensional":[85],"vectors.":[87],"To":[88],"achieve":[89],"this,":[90],"introduce":[92],"novel":[94],"optimization":[95],"objective":[96],"for":[97],"effective":[98],"variable":[100],"modeling":[101],"propose":[103],"Gradient-only":[105],"Skip":[106],"(G-Skip)":[107],"Connection,":[108],"which":[109],"jointly":[110],"enhances":[111],"controllability":[113],"while":[114],"preserving":[115],"quality.":[117],"Through":[118],"experiments":[119],"on":[120,140,154],"AGNews,":[121],"Yelp":[122],"DailyDialog,":[124],"validate":[126],"effectiveness":[128],"our":[130],"method":[131],"achieving":[133,158],"modelling":[135],"latent-guided":[137],"language":[138],"basis":[142],"Llama3-8B.":[144],"Specifically,":[145],"establish":[147],"new":[148],"state-of-the-art":[149],"performance":[150],"dialogue":[152],"DailyDialog":[156],"dataset,":[157],"BERTScore":[160],"88.30":[162],"an":[164],"FED":[165],"score":[166],"5.49.":[168]},"counts_by_year":[],"updated_date":"2026-04-09T08:11:56.329763","created_date":"2025-10-10T00:00:00"}