{"id":"https://openalex.org/W7213675700","doi":null,"title":"Dual-Axis Policy Optimization for LLM Agents: Bayesian Feedback Attribution and Trajectory Mass Normalization","display_name":"Dual-Axis Policy Optimization for LLM Agents: Bayesian Feedback Attribution and Trajectory Mass Normalization","publication_year":2026,"publication_date":"2026-09-17","ids":{"openalex":"https://openalex.org/W7213675700"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.19830","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.19830","pdf_url":"https://arxiv.org/pdf/2609.19830","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","indexed_in":["arxiv"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.19830","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5138298428","display_name":"Yingxuan Zhuang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhuang, Yingxuan","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5152224863","display_name":"Binhe Yu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yu, Binhe","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5152186822","display_name":"Jingxiao Yang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yang, Jingxiao","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5012442431","display_name":"Ruopei Sun","orcid":"https://orcid.org/0009-0000-4196-1738","observed_orcids":["https://orcid.org/0009-0000-4196-1738"]},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Sun, Ruopei","raw_affiliation_strings":["University of Science and Technology of China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Science and Technology of China","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057193902","display_name":"Ziting Li","orcid":"https://orcid.org/0009-0000-2903-4595","observed_orcids":["https://orcid.org/0009-0000-2903-4595"]},"institutions":[{"id":"https://openalex.org/I31746571","display_name":"UNSW Sydney","ror":"https://ror.org/03r8z3t63","country_code":"AU","type":"education","lineage":["https://openalex.org/I31746571"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Li, Ziting","raw_affiliation_strings":["University of New South Wales"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of New South Wales","institution_ids":["https://openalex.org/I31746571"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5152246594","display_name":"Cheng Tan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Tan, Cheng","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5152254094","display_name":"Xuhong Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhang, Xuhong","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133935664","display_name":"Jianwei Yin","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yin, Jianwei","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5125708739","display_name":"Jintao Chen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Jintao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"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":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":0.9980000257492065,"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":0.9980000257492065,"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/T12101","display_name":"Advanced Bandit Algorithms Research","score":0.00019999999494757503,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T10906","display_name":"AI-based Problem Solving and Planning","score":0.00019999999494757503,"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":0.9089999794960022},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"reinforcement learning","score":0.6039999723434448},{"id":"https://openalex.org/keywords/baton","display_name":"baton","score":0.6740000247955322},{"id":"https://openalex.org/keywords/group-relative-policy-optimization","display_name":"Group Relative Policy Optimization","score":0.5799999833106995},{"id":"https://openalex.org/keywords/alfworld","display_name":"ALFWorld","score":0.7620000243186951},{"id":"https://openalex.org/keywords/webshop","display_name":"WebShop","score":0.8920000195503235}],"concepts":[{"id":"https://openalex.org/C136886441","wikidata":"https://www.wikidata.org/wiki/Q926129","display_name":"Normalization (sociology)","level":2,"score":0.7842000126838684},{"id":"https://openalex.org/C2778049539","wikidata":"https://www.wikidata.org/wiki/Q17002908","display_name":"Bayesian optimization","level":2,"score":0.6195999979972839},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6172000169754028},{"id":"https://openalex.org/C107673813","wikidata":"https://www.wikidata.org/wiki/Q812534","display_name":"Bayesian probability","level":2,"score":0.5949000120162964},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.5752000212669373},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4413999915122986},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.4277999997138977},{"id":"https://openalex.org/C143299363","wikidata":"https://www.wikidata.org/wiki/Q900584","display_name":"Attribution","level":2,"score":0.4009999930858612},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.3564000129699707},{"id":"https://openalex.org/C173246807","wikidata":"https://www.wikidata.org/wiki/Q7833062","display_name":"Trajectory optimization","level":3,"score":0.33160001039505005},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.33000001311302185},{"id":"https://openalex.org/C47446073","wikidata":"https://www.wikidata.org/wiki/Q5165890","display_name":"Control theory (sociology)","level":3,"score":0.2872999906539917},{"id":"https://openalex.org/C160234255","wikidata":"https://www.wikidata.org/wiki/Q812535","display_name":"Bayesian inference","level":3,"score":0.2867000102996826}],"mesh":[],"locations_count":1,"locations":[{"id":"pmh:oai:arXiv.org:2609.19830","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.19830","pdf_url":"https://arxiv.org/pdf/2609.19830","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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.19830","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.19830","pdf_url":"https://arxiv.org/pdf/2609.19830","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/W7213675700.pdf","grobid_xml":"https://content.openalex.org/works/W7213675700.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7160921403","https://openalex.org/W4417093501","https://openalex.org/W7167748479","https://openalex.org/W4415954378","https://openalex.org/W7171929010","https://openalex.org/W4414982524","https://openalex.org/W7216351443","https://openalex.org/W7160600533","https://openalex.org/W7155061005","https://openalex.org/W7168291329"],"abstract_inverted_index":{"Reinforcement":[0],"learning":[1],"for":[2],"LLM":[3],"agents":[4],"involves":[5],"two":[6],"distinct":[7],"optimization":[8,54,85],"di-":[9],"mensions:":[10],"how":[11,20],"environment":[12],"feedback":[13],"is":[14],"exploited":[15],"within":[16],"a":[17,26,51,67],"trajectory,":[18],"and":[19,37,42,47,73,94,99,108],"complete":[21],"trajectories":[22],"are":[23],"aggregated":[24],"across":[25,119],"batch.":[27],"We":[28],"formulate":[29],"these":[30],"dimen-":[31],"sions":[32],"as":[33],"Intra-Trajectory":[34],"Feedback":[35,63],"Attribution":[36,46],"Inter-Trajectory":[38],"Objec-":[39],"tive":[40],"Aggregation,":[41],"introduce":[43],"BATON":[44,56],"(Bayesian":[45],"Trajectory":[48],"Objective":[49],"Normalization),":[50],"dual-axis":[52],"policy":[53],"framework.":[55],"instantiates":[57],"the":[58,74,115],"first":[59],"axis":[60],"with":[61,76,92],"Bayesian":[62],"Attribution,":[64],"which":[65,82],"constructs":[66],"feedback-conditioned":[68],"posterior":[69],"over":[70],"sampled":[71],"actions,":[72],"second":[75],"Trajec-":[77],"tory":[78],"Mass":[79],"Normalization":[80],"(TMN),":[81],"assigns":[83],"equal":[84],"mass":[86],"to":[87],"com-":[88],"plete":[89],"trajectories.":[90],"Experiments":[91],"GRPO":[93],"GiGPO":[95],"on":[96],"ALFWorld,":[97],"WebShop,":[98],"SearchQA":[100],"show":[101],"that":[102,109],"both":[103],"axes":[104],"provide":[105],"independent":[106],"gains":[107],"their":[110],"combi-":[111],"nation":[112],"consistently":[113],"achieves":[114],"strongest":[116],"overall":[117],"performance":[118],"model":[120],"scales.":[121]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-09-20T00:00:00"}