{"id":"https://openalex.org/W4412944516","doi":"https://doi.org/10.18653/v1/2025.findings-acl.1302","title":"Sparse Rewards Can Self-Train Dialogue Agents","display_name":"Sparse Rewards Can Self-Train Dialogue Agents","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412944516","doi":"https://doi.org/10.18653/v1/2025.findings-acl.1302"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.findings-acl.1302","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.1302","pdf_url":"https://aclanthology.org/2025.findings-acl.1302.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":"Findings of the Association for Computational Linguistics: ACL 2025","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-acl.1302.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5092427250","display_name":"Barrett Martin Lattimer","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Barrett Martin Lattimer","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022752740","display_name":"Varun Gangal","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Varun Prashant Gangal","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022489655","display_name":"Ryan McDonald","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ryan McDonald","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5019929567","display_name":"Yi Yang","orcid":"https://orcid.org/0000-0001-5570-8610"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yi Yang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":4,"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":0.08836811,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"25395","last_page":"25413"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12031","display_name":"Speech and dialogue systems","score":0.995199978351593,"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":0.995199978351593,"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/T10456","display_name":"Multi-Agent Systems and Negotiation","score":0.9882000088691711,"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/T10028","display_name":"Topic Modeling","score":0.9735999703407288,"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":0.6795214414596558},{"id":"https://openalex.org/keywords/human\u2013computer-interaction","display_name":"Human\u2013computer interaction","score":0.41152167320251465},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.3766602873802185}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6795214414596558},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.41152167320251465},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3766602873802185}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-acl.1302","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.1302","pdf_url":"https://aclanthology.org/2025.findings-acl.1302.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":"Findings of the Association for Computational Linguistics: ACL 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-acl.1302","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.1302","pdf_url":"https://aclanthology.org/2025.findings-acl.1302.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":"Findings of the Association for Computational Linguistics: ACL 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412944516.pdf","grobid_xml":"https://content.openalex.org/works/W4412944516.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"Recent":[0],"advancements":[1],"in":[2,11],"state-of-the-art":[3],"(SOTA)":[4],"Large":[5],"Language":[6],"Model":[7],"(LLM)":[8],"agents,":[9],"especially":[10],"multi-turn":[12],"dialogue":[13],"tasks,":[14],"have":[15],"been":[16],"primarily":[17],"driven":[18],"by":[19],"supervised":[20],"fine-tuning":[21],"and":[22,41,101,129,144],"high-quality":[23],"human":[24,35,51,77],"feedback.However,":[25],"as":[26],"base":[27,45],"LLM":[28,46,68,105],"models":[29,123],"continue":[30],"to":[31,70,97],"improve,":[32],"acquiring":[33],"meaningful":[34],"feedback":[36],"has":[37],"become":[38],"increasingly":[39],"challenging":[40],"costly.In":[42],"certain":[43],"domains,":[44],"agents":[47,69],"may":[48],"eventually":[49],"exceed":[50],"capabilities,":[52],"making":[53],"traditional":[54],"feedback-driven":[55],"methods":[56],"impractical.In":[57],"this":[58],"paper,":[59],"we":[60],"introduce":[61],"a":[62,87,92,112],"novel":[63],"self-improvement":[64],"paradigm":[65],"that":[66,90,122],"empowers":[67],"autonomously":[71],"enhance":[72],"their":[73],"performance":[74],"without":[75],"external":[76],"feedback.Our":[78],"method,":[79],"Juxtaposed":[80],"Outcomes":[81],"for":[82],"Simulation":[83],"Harvesting":[84],"(JOSH),":[85],"is":[86],"self-alignment":[88],"algorithm":[89],"leverages":[91],"sparse":[93,113],"reward":[94,114],"simulation":[95,116],"environment":[96,117],"extract":[98],"ideal":[99],"behaviors":[100],"further":[102],"train":[103],"the":[104],"on":[106,149],"its":[107],"own":[108],"outputs.We":[109],"present":[110],"ToolWOZ,":[111],"tool-calling":[115],"derived":[118],"from":[119],"MultiWOZ.We":[120],"demonstrate":[121],"trained":[124],"with":[125],"JOSH,":[126],"both":[127],"small":[128],"frontier,":[130],"significantly":[131],"improve":[132],"tool-based":[133],"interactions":[134],"while":[135],"preserving":[136],"general":[137],"model":[138],"capabilities":[139],"across":[140],"diverse":[141],"benchmarks.Our":[142],"code":[143],"data":[145],"are":[146],"publicly":[147],"available":[148],"GitHub":[150],"at":[151],"https://github.com/asappresearch/":[152],"josh-llm-simulation-training.git":[153]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}