{"id":"https://openalex.org/W7214775583","doi":"https://doi.org/10.48550/arxiv.2609.35084","title":"GraphHCA: Closed-Form Hindsight Credit Assignment for Long-Horizon LLM Agents","display_name":"GraphHCA: Closed-Form Hindsight Credit Assignment for Long-Horizon LLM Agents","publication_year":2026,"publication_date":"2026-09-28","ids":{"openalex":"https://openalex.org/W7214775583","doi":"https://doi.org/10.48550/arxiv.2609.35084"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.35084","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.35084","pdf_url":"https://arxiv.org/pdf/2609.35084","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.35084","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5153453495","display_name":"Haodong Zhu","orcid":null,"observed_orcids":[]},"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":"Zhu, Haodong","raw_affiliation_strings":["Beihang University","Zhongguancun Academy"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beihang University","institution_ids":["https://openalex.org/I82880672"]},{"raw_affiliation_string":"Zhongguancun Academy","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042575284","display_name":"Yangyang Ren","orcid":"https://orcid.org/0009-0009-3191-6365","observed_orcids":["https://orcid.org/0009-0009-3191-6365"]},"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":"Ren, Yangyang","raw_affiliation_strings":["Beihang University","Zhongguancun Academy"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beihang University","institution_ids":["https://openalex.org/I82880672"]},{"raw_affiliation_string":"Zhongguancun Academy","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5062954540","display_name":"Changbai Li","orcid":null,"observed_orcids":[]},"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":"Li, Changbai","raw_affiliation_strings":["Beihang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beihang University","institution_ids":["https://openalex.org/I82880672"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153407089","display_name":"Sheng Xu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I75689368","display_name":"Communication University of China","ror":"https://ror.org/04facbs33","country_code":"CN","type":"education","lineage":["https://openalex.org/I75689368"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xu, Sheng","raw_affiliation_strings":["Communication University of China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Communication University of China","institution_ids":["https://openalex.org/I75689368"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153533004","display_name":"Linlin Yang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I75689368","display_name":"Communication University of China","ror":"https://ror.org/04facbs33","country_code":"CN","type":"education","lineage":["https://openalex.org/I75689368"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang, Linlin","raw_affiliation_strings":["Communication University of China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Communication University of China","institution_ids":["https://openalex.org/I75689368"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128205393","display_name":"Haiguang Liu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"liu, haiguang","raw_affiliation_strings":["Zhongguancun Academy"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhongguancun Academy","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5153430658","display_name":"Baochang Zhang","orcid":null,"observed_orcids":[]},"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":"Zhang, Baochang","raw_affiliation_strings":["Beihang University","Hangzhou Innovation Institute of Beihang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Beihang University","institution_ids":["https://openalex.org/I82880672"]},{"raw_affiliation_string":"Hangzhou Innovation Institute of Beihang University","institution_ids":["https://openalex.org/I82880672"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":0.9990000128746033,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.9990000128746033,"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/T10906","display_name":"AI-based Problem Solving and Planning","score":9.999999747378752e-05,"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/T11574","display_name":"Artificial Intelligence in Games","score":9.999999747378752e-05,"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/llm-based-agents","display_name":"LLM agents","score":0.8100000023841858},{"id":"https://openalex.org/keywords/state-transition-diagrams","display_name":"state diagrams","score":0.6159999966621399},{"id":"https://openalex.org/keywords/group-relative-policy-optimization","display_name":"Group Relative Policy Optimization","score":0.625},{"id":"https://openalex.org/keywords/alfworld","display_name":"ALFWorld","score":0.7599999904632568},{"id":"https://openalex.org/keywords/webshop","display_name":"WebShop","score":0.8029999732971191},{"id":"https://openalex.org/keywords/sokoban","display_name":"Sokoban","score":0.8899999856948853},{"id":"https://openalex.org/keywords/vision-language-agents","display_name":"vision-language agents","score":0.8130000233650208}],"concepts":[{"id":"https://openalex.org/C10347200","wikidata":"https://www.wikidata.org/wiki/Q1960297","display_name":"Hindsight bias","level":2,"score":0.9922999739646912},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5439000129699707},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.51910001039505},{"id":"https://openalex.org/C168773036","wikidata":"https://www.wikidata.org/wiki/Q264164","display_name":"Recursion (computer science)","level":2,"score":0.5034999847412109},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.4902999997138977},{"id":"https://openalex.org/C25343380","wikidata":"https://www.wikidata.org/wiki/Q277521","display_name":"Relation (database)","level":2,"score":0.42989999055862427},{"id":"https://openalex.org/C2781089630","wikidata":"https://www.wikidata.org/wiki/Q21856745","display_name":"Realization (probability)","level":2,"score":0.3910999894142151},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.3702999949455261},{"id":"https://openalex.org/C148220186","wikidata":"https://www.wikidata.org/wiki/Q7111912","display_name":"Outcome (game theory)","level":2,"score":0.3675999939441681},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3659000098705292},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.36550000309944153},{"id":"https://openalex.org/C2777027219","wikidata":"https://www.wikidata.org/wiki/Q1284190","display_name":"Constant (computer programming)","level":2,"score":0.328900009393692},{"id":"https://openalex.org/C162324750","wikidata":"https://www.wikidata.org/wiki/Q8134","display_name":"Economics","level":0,"score":0.3276999890804291},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.32710000872612},{"id":"https://openalex.org/C162118730","wikidata":"https://www.wikidata.org/wiki/Q1128453","display_name":"Actuarial science","level":1,"score":0.32659998536109924},{"id":"https://openalex.org/C2777894999","wikidata":"https://www.wikidata.org/wiki/Q4781758","display_name":"Approx","level":2,"score":0.3208000063896179},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.31850001215934753},{"id":"https://openalex.org/C44648626","wikidata":"https://www.wikidata.org/wiki/Q1049848","display_name":"Percentage point","level":2,"score":0.3151000142097473},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.30239999294281006},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.289900004863739},{"id":"https://openalex.org/C96250715","wikidata":"https://www.wikidata.org/wiki/Q965330","display_name":"Estimation","level":2,"score":0.2870999872684479},{"id":"https://openalex.org/C134121241","wikidata":"https://www.wikidata.org/wiki/Q899301","display_name":"Yield (engineering)","level":2,"score":0.2815999984741211},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.27900001406669617},{"id":"https://openalex.org/C144237770","wikidata":"https://www.wikidata.org/wiki/Q747534","display_name":"Mathematical economics","level":1,"score":0.2734000086784363},{"id":"https://openalex.org/C88626702","wikidata":"https://www.wikidata.org/wiki/Q1128903","display_name":"Continuation","level":2,"score":0.27250000834465027},{"id":"https://openalex.org/C41426520","wikidata":"https://www.wikidata.org/wiki/Q1192065","display_name":"Point estimation","level":2,"score":0.26019999384880066},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.25780001282691956},{"id":"https://openalex.org/C159023740","wikidata":"https://www.wikidata.org/wiki/Q623276","display_name":"Deadlock","level":2,"score":0.25360000133514404},{"id":"https://openalex.org/C6422946","wikidata":"https://www.wikidata.org/wiki/Q857354","display_name":"Panel data","level":2,"score":0.25290000438690186}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.35084","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.35084","pdf_url":"https://arxiv.org/pdf/2609.35084","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.35084","doi":"https://doi.org/10.48550/arxiv.2609.35084","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.35084","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.35084","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.35084","pdf_url":"https://arxiv.org/pdf/2609.35084","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7214775583.pdf","grobid_xml":"https://content.openalex.org/works/W7214775583.grobid-xml"},"referenced_works_count":14,"referenced_works":["https://openalex.org/W2736601468","https://openalex.org/W2978455699","https://openalex.org/W4223908421","https://openalex.org/W4327810158","https://openalex.org/W4390041933","https://openalex.org/W4391631327","https://openalex.org/W4402427562","https://openalex.org/W4405655184","https://openalex.org/W4406774573","https://openalex.org/W4407794264","https://openalex.org/W4414281281","https://openalex.org/W4417422257","https://openalex.org/W7134953399","https://openalex.org/W7162494674"],"related_works":["https://openalex.org/W7162494674","https://openalex.org/W7212730484","https://openalex.org/W7134953399","https://openalex.org/W7214358361","https://openalex.org/W7203896364","https://openalex.org/W7103150324","https://openalex.org/W7221537839","https://openalex.org/W7221199531","https://openalex.org/W7202145059","https://openalex.org/W7222111281"],"abstract_inverted_index":{"Group-based":[0],"reinforcement":[1],"learning":[2],"(RL)":[3],"has":[4],"advanced":[5],"large":[6],"language":[7],"models":[8],"(LLMs)":[9],"and":[10,180,199,205,228],"is":[11,163,187],"increasingly":[12],"extending":[13],"to":[14,46,62,108,223,231],"agentic":[15],"tasks,":[16],"where":[17],"sparse":[18],"terminal":[19],"rewards":[20],"make":[21],"step-level":[22,131,161,185,237],"credit":[23,29,51,56],"assignment":[24,52],"essential.":[25],"Existing":[26],"methods":[27],"assign":[28],"from":[30,137],"what":[31],"follows":[32],"an":[33,71,75,176],"action":[34],"in":[35],"sampled":[36],"rollouts,":[37],"but":[38,65],"do":[39],"not":[40],"explicitly":[41],"capture":[42],"its":[43],"retrospective":[44],"relation":[45],"the":[47,58,67,105,145,166,184,235],"realized":[48],"outcome.":[49],"Hindsight":[50],"(HCA)":[53],"instead":[54],"attributes":[55],"through":[57,140],"ratio":[59,107,110],"of":[60,89,111],"hindsight":[61,68,106,173],"behavior-policy":[63,112],"probabilities,":[64],"estimating":[66],"distribution":[69],"requires":[70],"auxiliary":[72],"model":[73,174],"or":[74],"extra":[76,177],"pass.":[77],"To":[78],"address":[79],"this":[80,135],"estimation":[81],"bottleneck,":[82],"we":[83],"propose":[84],"GraphHCA,":[85],"a":[86,109,121,128,141,151,171,209],"model-free":[87],"realization":[88],"HCA":[90],"that":[91],"eliminates":[92],"explicit":[93],"hindsight-distribution":[94],"estimation.":[95],"For":[96,212],"terminal-goal":[97],"tasks":[98],"with":[99,165,208],"deterministic":[100],"transitions,":[101],"Bayes'":[102],"rule":[103],"reduces":[104],"success":[113,123,219],"probabilities":[114],"at":[115,201],"consecutive":[116],"states.":[117],"Taking":[118],"logs":[119],"yields":[120],"state-wise":[122],"potential,":[124],"whose":[125],"increment":[126],"across":[127],"transition":[129,147],"provides":[130],"credit.":[132],"GraphHCA":[133,193],"estimates":[134],"potential":[136],"pooled":[138],"rollouts":[139],"discounted":[142],"recursion":[143],"on":[144,155,197,206,214],"induced":[146],"graph,":[148],"which":[149],"admits":[150],"unique":[152],"fixed":[153],"point":[154],"any":[156],"directed":[157],"graph.":[158],"The":[159],"resulting":[160],"signal":[162],"combined":[164],"trajectory-level":[167],"advantage,":[168],"requiring":[169],"neither":[170],"learned":[172],"nor":[175],"forward":[178],"pass":[179],"recovering":[181],"GRPO":[182,227],"when":[183],"weight":[186],"zero.":[188],"Among":[189],"all":[190],"compared":[191],"baselines,":[192],"achieves":[194],"state-of-the-art":[195],"results":[196],"ALFWorld":[198,215],"WebShop":[200],"both":[202],"LLM":[203],"scales,":[204],"Sokoban":[207],"vision-language":[210],"agent.":[211],"example,":[213],"it":[216],"improves":[217],"overall":[218],"rate":[220],"by":[221,229],"up":[222,230],"24.6":[224],"points":[225,233],"over":[226,234],"4.7":[232],"strongest":[236],"baseline.":[238]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-09-30T00:00:00"}