{"id":"https://openalex.org/W7223697015","doi":"https://doi.org/10.48550/arxiv.2610.32125","title":"Conditional Flow Models for Offline Maximum Entropy Reinforcement Learning","display_name":"Conditional Flow Models for Offline Maximum Entropy Reinforcement Learning","publication_year":2026,"publication_date":"2026-10-23","ids":{"openalex":"https://openalex.org/W7223697015","doi":"https://doi.org/10.48550/arxiv.2610.32125"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.32125","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.32125","pdf_url":"https://arxiv.org/pdf/2610.32125","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/2610.32125","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5159581286","display_name":"Hao Cheng","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Cheng, Hao","raw_affiliation_strings":["School of Mathematics and Statistics , Wuhan University , Wuhan , 430072 , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Mathematics and Statistics , Wuhan University , Wuhan , 430072 , China","institution_ids":["https://openalex.org/I37461747"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5080185313","display_name":"Yuling Jiao","orcid":"https://orcid.org/0000-0003-2762-0421","observed_orcids":["https://orcid.org/0000-0003-2762-0421"]},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiao, Yuling","raw_affiliation_strings":["School of Artificial Intelligence , and Hubei Key Laboratory of Computational Science , Wuhan University , Wuhan , 430072 , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Artificial Intelligence , and Hubei Key Laboratory of Computational Science , Wuhan University , Wuhan , 430072 , China","institution_ids":["https://openalex.org/I37461747"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086788173","display_name":"Lican Kang","orcid":"https://orcid.org/0000-0003-3136-9225","observed_orcids":["https://orcid.org/0000-0003-3136-9225"]},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Kang, Lican","raw_affiliation_strings":["Institute for Math and AI , Hubei Key Laboratory of Computational Science , and School of Artificial Intelligence , Wuhan University , Wuhan , 430072 , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Institute for Math and AI , Hubei Key Laboratory of Computational Science , and School of Artificial Intelligence , Wuhan University , Wuhan , 430072 , China","institution_ids":["https://openalex.org/I37461747"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057896052","display_name":"Jerry Zhijian Yang","orcid":"https://orcid.org/0000-0002-0402-4056","observed_orcids":["https://orcid.org/0000-0002-0402-4056"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Jerry Zhijian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5159730103","display_name":"Chen Zhong","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I37461747","display_name":"Wuhan University","ror":"https://ror.org/033vjfk17","country_code":"CN","type":"education","lineage":["https://openalex.org/I37461747"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhong, Chen","raw_affiliation_strings":["School of Mathematics and Statistics , Wuhan University , Wuhan , 430072 , China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Mathematics and Statistics , Wuhan University , Wuhan , 430072 , China","institution_ids":["https://openalex.org/I37461747"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"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":0.9994999766349792,"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":0.9994999766349792,"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/T12794","display_name":"Adaptive Dynamic Programming Control","score":0.00019999999494757503,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12101","display_name":"Advanced Bandit Algorithms Research","score":0.0,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/offline-reinforcement-learning","display_name":"offline reinforcement learning","score":0.8610000014305115},{"id":"https://openalex.org/keywords/maximum-entropy-reinforcement-learning","display_name":"maximum entropy reinforcement learning","score":0.9940000176429749},{"id":"https://openalex.org/keywords/conditional-diffusion-models","display_name":"conditional diffusion models","score":0.8059999942779541},{"id":"https://openalex.org/keywords/probability-flow-ode","display_name":"probability flow ODE","score":0.9419999718666077},{"id":"https://openalex.org/keywords/holder-regularity","display_name":"H\u00f6lder regularity","score":0.8980000019073486}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.6358000040054321},{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.5885999798774719},{"id":"https://openalex.org/C14646407","wikidata":"https://www.wikidata.org/wiki/Q1430750","display_name":"Bellman equation","level":2,"score":0.5618000030517578},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.4932999908924103},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.4693000018596649},{"id":"https://openalex.org/C106301342","wikidata":"https://www.wikidata.org/wiki/Q4117933","display_name":"Entropy (arrow of time)","level":2,"score":0.4345000088214874},{"id":"https://openalex.org/C17020691","wikidata":"https://www.wikidata.org/wiki/Q139677","display_name":"Operator (biology)","level":5,"score":0.40619999170303345},{"id":"https://openalex.org/C43555835","wikidata":"https://www.wikidata.org/wiki/Q2300258","display_name":"Conditional probability distribution","level":2,"score":0.38839998841285706},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.37790000438690186},{"id":"https://openalex.org/C91575142","wikidata":"https://www.wikidata.org/wiki/Q1971426","display_name":"Optimal control","level":2,"score":0.37070000171661377},{"id":"https://openalex.org/C33676613","wikidata":"https://www.wikidata.org/wiki/Q13415176","display_name":"Dimension (graph theory)","level":2,"score":0.36820000410079956},{"id":"https://openalex.org/C186215838","wikidata":"https://www.wikidata.org/wiki/Q772232","display_name":"Conditional expectation","level":2,"score":0.3587999939918518},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.34380000829696655},{"id":"https://openalex.org/C149441793","wikidata":"https://www.wikidata.org/wiki/Q200726","display_name":"Probability distribution","level":2,"score":0.3264000117778778},{"id":"https://openalex.org/C37404715","wikidata":"https://www.wikidata.org/wiki/Q380679","display_name":"Dynamic programming","level":2,"score":0.3190999925136566},{"id":"https://openalex.org/C155512373","wikidata":"https://www.wikidata.org/wiki/Q287450","display_name":"Residual","level":2,"score":0.31029999256134033},{"id":"https://openalex.org/C9679016","wikidata":"https://www.wikidata.org/wiki/Q1417473","display_name":"Principle of maximum entropy","level":2,"score":0.3034000098705292},{"id":"https://openalex.org/C122383733","wikidata":"https://www.wikidata.org/wiki/Q865920","display_name":"Approximation error","level":2,"score":0.29319998621940613},{"id":"https://openalex.org/C77553402","wikidata":"https://www.wikidata.org/wiki/Q13222579","display_name":"Upper and lower bounds","level":2,"score":0.29159998893737793},{"id":"https://openalex.org/C101721835","wikidata":"https://www.wikidata.org/wiki/Q813908","display_name":"Conditional entropy","level":3,"score":0.29100000858306885},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.27889999747276306},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.2727999985218048},{"id":"https://openalex.org/C119322782","wikidata":"https://www.wikidata.org/wiki/Q2662236","display_name":"VC dimension","level":2,"score":0.259799987077713},{"id":"https://openalex.org/C91873725","wikidata":"https://www.wikidata.org/wiki/Q3445816","display_name":"Function approximation","level":3,"score":0.257099986076355},{"id":"https://openalex.org/C103784038","wikidata":"https://www.wikidata.org/wiki/Q386228","display_name":"Cumulative distribution function","level":3,"score":0.25459998846054077},{"id":"https://openalex.org/C171752962","wikidata":"https://www.wikidata.org/wiki/Q255166","display_name":"Kullback\u2013Leibler divergence","level":2,"score":0.2535000145435333}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.32125","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.32125","pdf_url":"https://arxiv.org/pdf/2610.32125","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.2610.32125","doi":"https://doi.org/10.48550/arxiv.2610.32125","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.32125","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2610.32125","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.32125","pdf_url":"https://arxiv.org/pdf/2610.32125","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":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7223697015.pdf","grobid_xml":"https://content.openalex.org/works/W7223697015.grobid-xml"},"referenced_works_count":2,"referenced_works":["https://openalex.org/W2965140792","https://openalex.org/W4409974863"],"related_works":["https://openalex.org/W7203650104","https://openalex.org/W190762192","https://openalex.org/W4416398156","https://openalex.org/W3133661462","https://openalex.org/W4313914531","https://openalex.org/W7102329132","https://openalex.org/W4411945999","https://openalex.org/W7148729830","https://openalex.org/W4415797558","https://openalex.org/W7131130268"],"abstract_inverted_index":{"Offline":[0],"maximum":[1],"entropy":[2],"reinforcement":[3,222],"learning":[4,188],"seeks":[5],"to":[6,120,167,215],"estimate":[7,84,121],"the":[8,22,29,34,44,48,57,85,102,128,141,163,175,186,204,208,234,239,242],"optimal":[9,31,104],"soft":[10,30,103,157,164],"action-value":[11],"function":[12,65],"\\(Q^*\\)":[13,25,55],"from":[14,174],"a":[15,69,97,114],"fixed":[16],"dataset":[17],"without":[18],"further":[19],"interaction":[20],"with":[21],"environment.":[23],"Although":[24],"is":[26,111,203],"characterized":[27],"by":[28],"Bellman":[32,36,105,109,165,193,229],"equation,":[33],"corresponding":[35],"operator":[37,132],"cannot":[38],"be":[39],"evaluated":[40],"directly":[41],"because":[42],"both":[43],"reward":[45,87,177],"distribution":[46,88],"and":[47,63,89,153,178,210,237],"transition":[49,90,179],"kernel":[50],"are":[51],"unknown.":[52],"Consequently,":[53],"estimating":[54],"requires":[56],"joint":[58],"treatment":[59],"of":[60,101,116,197,207,220,241],"environment":[61],"modeling":[62],"value":[64,158,187],"learning.":[66],"We":[67,123],"propose":[68],"framework":[70],"that":[71,145,169],"addresses":[72],"these":[73],"tasks":[74],"separately.":[75],"Conditional":[76],"diffusion":[77],"models":[78],"based":[79],"on":[80],"probability":[81],"flow":[82,143],"ODEs":[83],"conditional":[86,142,176],"kernel.":[91],"The":[92],"resulting":[93],"distributional":[94],"estimators":[95,144],"define":[96],"data":[98],"driven":[99],"approximation":[100],"operator,":[106],"whose":[107],"empirical":[108],"residual":[110,194],"minimized":[112],"over":[113],"class":[115],"deep":[117,221],"neural":[118],"networks":[119],"\\(Q^*\\).":[122],"establish":[124],"nonasymptotic":[125],"guarantees":[126],"for":[127,140,148],"proposed":[129,243],"method.":[130,244],"For":[131,156,181],"learning,":[133,159,223],"we":[134,160],"derive":[135],"total":[136,205],"variation":[137],"error":[138],"bounds":[139],"account":[146],"explicitly":[147],"score":[149],"estimation,":[150],"numerical":[151],"discretization,":[152],"terminal":[154],"smoothing.":[155],"first":[161],"use":[162],"equation":[166],"show":[168],"$Q^*$":[170],"inherits":[171],"H\u00f6lder":[172,183],"regularity":[173],"densities.":[180],"any":[182],"index":[184],"$\u03b2$,":[185],"stage":[189],"achieves":[190],"an":[191],"excess":[192],"risk":[195],"bound":[196],"order":[198],"$\\widetilde{\\mathcal":[199],"O}\\big(n^{-\\frac{2\u03b2}{d+2\u03b2}}\\big),$":[200],"where":[201],"$d$":[202],"dimension":[206],"state":[209],"action":[211],"spaces.":[212],"In":[213],"contrast":[214],"many":[216],"existing":[217],"theoretical":[218,235],"analyses":[219],"our":[224],"results":[225],"do":[226],"not":[227],"require":[228],"completeness.":[230],"Numerical":[231],"experiments":[232],"support":[233],"findings":[236],"demonstrate":[238],"effectiveness":[240]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-10-27T00:00:00"}