{"id":"https://openalex.org/W7138189871","doi":"https://doi.org/10.1609/aaai.v40i30.39739","title":"Conformal Constrained Policy Optimization for Cost-Effective LLM Agents","display_name":"Conformal Constrained Policy Optimization for Cost-Effective LLM Agents","publication_year":2026,"publication_date":"2026-03-14","ids":{"openalex":"https://openalex.org/W7138189871","doi":"https://doi.org/10.1609/aaai.v40i30.39739"},"language":null,"primary_location":{"id":"doi:10.1609/aaai.v40i30.39739","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i30.39739","pdf_url":"https://ojs.aaai.org/index.php/AAAI/article/download/39739/43700","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":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"diamond","oa_url":"https://ojs.aaai.org/index.php/AAAI/article/download/39739/43700","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5057113510","display_name":"Wenwen Si","orcid":"https://orcid.org/0000-0002-7066-9901"},"institutions":[{"id":"https://openalex.org/I36788626","display_name":"California University of Pennsylvania","ror":"https://ror.org/01spssf70","country_code":"US","type":"education","lineage":["https://openalex.org/I36788626"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Wenwen Si","raw_affiliation_strings":["University of Pennsylvania"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Pennsylvania","institution_ids":["https://openalex.org/I36788626"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129699521","display_name":"Sooyong Jang","orcid":null},"institutions":[{"id":"https://openalex.org/I36788626","display_name":"California University of Pennsylvania","ror":"https://ror.org/01spssf70","country_code":"US","type":"education","lineage":["https://openalex.org/I36788626"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Sooyong Jang","raw_affiliation_strings":["University of Pennsylvania"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Pennsylvania","institution_ids":["https://openalex.org/I36788626"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129659119","display_name":"Insup Lee","orcid":null},"institutions":[{"id":"https://openalex.org/I36788626","display_name":"California University of Pennsylvania","ror":"https://ror.org/01spssf70","country_code":"US","type":"education","lineage":["https://openalex.org/I36788626"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Insup Lee","raw_affiliation_strings":["University of Pennsylvania"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Pennsylvania","institution_ids":["https://openalex.org/I36788626"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5129747254","display_name":"Osbert Bastani","orcid":null},"institutions":[{"id":"https://openalex.org/I36788626","display_name":"California University of Pennsylvania","ror":"https://ror.org/01spssf70","country_code":"US","type":"education","lineage":["https://openalex.org/I36788626"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Osbert Bastani","raw_affiliation_strings":["University of Pennsylvania"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Pennsylvania","institution_ids":["https://openalex.org/I36788626"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I36788626"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":"40","issue":"30","first_page":"25446","last_page":"25453"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.5127000212669373,"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":0.5127000212669373,"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":0.10189999639987946,"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":0.0731000006198883,"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/orchestration","display_name":"Orchestration","score":0.5932999849319458},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.5532000064849854},{"id":"https://openalex.org/keywords/constraint","display_name":"Constraint (computer-aided design)","score":0.5160999894142151},{"id":"https://openalex.org/keywords/conformal-map","display_name":"Conformal map","score":0.44600000977516174},{"id":"https://openalex.org/keywords/sequence","display_name":"Sequence (biology)","score":0.4422000050544739},{"id":"https://openalex.org/keywords/reduction","display_name":"Reduction (mathematics)","score":0.44200000166893005}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7049000263214111},{"id":"https://openalex.org/C199168358","wikidata":"https://www.wikidata.org/wiki/Q3367000","display_name":"Orchestration","level":3,"score":0.5932999849319458},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5532000064849854},{"id":"https://openalex.org/C2776036281","wikidata":"https://www.wikidata.org/wiki/Q48769818","display_name":"Constraint (computer-aided design)","level":2,"score":0.5160999894142151},{"id":"https://openalex.org/C98214594","wikidata":"https://www.wikidata.org/wiki/Q850275","display_name":"Conformal map","level":2,"score":0.44600000977516174},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.4422000050544739},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.44200000166893005},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.43790000677108765},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.3702000081539154},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.3425999879837036},{"id":"https://openalex.org/C173404611","wikidata":"https://www.wikidata.org/wiki/Q528588","display_name":"Constraint programming","level":3,"score":0.3425000011920929},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3156000077724457},{"id":"https://openalex.org/C2778820799","wikidata":"https://www.wikidata.org/wiki/Q3454688","display_name":"Cost reduction","level":2,"score":0.3075000047683716},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.30300000309944153},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.2702000141143799}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1609/aaai.v40i30.39739","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i30.39739","pdf_url":"https://ojs.aaai.org/index.php/AAAI/article/download/39739/43700","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.v40i30.39739","is_oa":true,"landing_page_url":"https://doi.org/10.1609/aaai.v40i30.39739","pdf_url":"https://ojs.aaai.org/index.php/AAAI/article/download/39739/43700","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":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7138189871.pdf","grobid_xml":"https://content.openalex.org/works/W7138189871.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"While":[0],"large":[1],"language":[2],"models":[3,36,46],"(LLMs)":[4],"have":[5,16],"recently":[6],"made":[7],"tremendous":[8],"progress":[9],"towards":[10],"solving":[11],"challenging":[12],"AI":[13],"problems,":[14],"they":[15],"done":[17],"so":[18],"at":[19],"increasingly":[20],"steep":[21],"computational":[22],"and":[23,47,102,117,140,151],"API":[24],"costs.":[25],"We":[26],"propose":[27,84],"a":[28,64,90,112,131,149],"novel":[29],"strategy":[30],"where":[31,45],"we":[32,83],"combine":[33],"multiple":[34],"LLM":[35,156],"with":[37,98],"varying":[38],"cost/accuracy":[39],"tradeoffs":[40],"in":[41,51,105],"an":[42,56,118],"agentic":[43],"manner,":[44],"tools":[48],"are":[49,159],"run":[50],"sequence":[52],"as":[53],"determined":[54],"by":[55],"orchestration":[57],"model":[58],"to":[59,63,76,130,136],"minimize":[60],"cost":[61,133],"subject":[62],"user-specified":[65],"level":[66],"of":[67],"reliability;":[68],"this":[69,81],"constraint":[70],"is":[71],"formalized":[72],"using":[73],"conformal":[74,107],"prediction":[75],"provide":[77],"guarantees.":[78],"To":[79],"solve":[80],"problem,":[82],"Conformal":[85],"Constrained":[86],"Policy":[87],"Optimization":[88],"(CCPO),":[89],"training":[91],"paradigm":[92],"that":[93,158],"integrates":[94],"constrained":[95],"policy":[96,114],"optimization":[97],"off-policy":[99],"reinforcement":[100],"learning":[101],"recent":[103],"advances":[104],"online":[106],"prediction.":[108],"CCPO":[109,127],"jointly":[110],"optimizes":[111],"cost-aware":[113,138],"(score":[115],"function)":[116],"adaptive":[119],"threshold.":[120],"Across":[121],"two":[122],"multi-hop":[123],"question":[124],"answering":[125],"benchmarks,":[126],"achieves":[128],"up":[129],"30%":[132],"reduction":[134],"compared":[135],"other":[137],"baselines":[139],"LLM-guided":[141],"methods":[142],"without":[143],"compromising":[144],"reliability.":[145,165],"Our":[146],"approach":[147],"provides":[148],"principled":[150],"practical":[152],"framework":[153],"for":[154],"deploying":[155],"agents":[157],"significantly":[160],"more":[161],"cost-effective":[162],"while":[163],"maintaining":[164]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2026-03-18T00:00:00"}