{"id":"https://openalex.org/W7166714858","doi":"https://doi.org/10.48550/arxiv.2606.28707","title":"BV-Blend: Uncertainty-Weighted Historical Baselines for Stable Critic-Free RL with Verifiable Rewards","display_name":"BV-Blend: Uncertainty-Weighted Historical Baselines for Stable Critic-Free RL with Verifiable Rewards","publication_year":2026,"publication_date":"2026-06-27","ids":{"openalex":"https://openalex.org/W7166714858","doi":"https://doi.org/10.48550/arxiv.2606.28707"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.28707","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.28707","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.28707","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5102930702","display_name":"Yupeng Chang","orcid":"https://orcid.org/0000-0001-7178-6088"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chang, Yupeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139661427","display_name":"Yuan Wu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Yuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139644250","display_name":"Yi Chang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chang, Yi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"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":0.7275999784469604,"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":0.7275999784469604,"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":0.05350000038743019,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.021299999207258224,"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/verifiable-secret-sharing","display_name":"Verifiable secret sharing","score":0.7785000205039978},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.5472999811172485},{"id":"https://openalex.org/keywords/normalization","display_name":"Normalization (sociology)","score":0.47769999504089355},{"id":"https://openalex.org/keywords/variance","display_name":"Variance (accounting)","score":0.46540001034736633},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.46230000257492065},{"id":"https://openalex.org/keywords/function","display_name":"Function (biology)","score":0.4383000135421753},{"id":"https://openalex.org/keywords/bellman-equation","display_name":"Bellman equation","score":0.41449999809265137},{"id":"https://openalex.org/keywords/binary-number","display_name":"Binary number","score":0.3959999978542328}],"concepts":[{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.7785000205039978},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6141999959945679},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5472999811172485},{"id":"https://openalex.org/C136886441","wikidata":"https://www.wikidata.org/wiki/Q926129","display_name":"Normalization (sociology)","level":2,"score":0.47769999504089355},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.46540001034736633},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.46230000257492065},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.4383000135421753},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.43389999866485596},{"id":"https://openalex.org/C14646407","wikidata":"https://www.wikidata.org/wiki/Q1430750","display_name":"Bellman equation","level":2,"score":0.41449999809265137},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.3959999978542328},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3921999931335449},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.3612000048160553},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3571999967098236},{"id":"https://openalex.org/C2776291640","wikidata":"https://www.wikidata.org/wiki/Q2912517","display_name":"Value (mathematics)","level":2,"score":0.3296999931335449},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.32589998841285706},{"id":"https://openalex.org/C2781311116","wikidata":"https://www.wikidata.org/wiki/Q83306","display_name":"Group (periodic table)","level":2,"score":0.313400000333786},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.31139999628067017},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.30790001153945923},{"id":"https://openalex.org/C201236551","wikidata":"https://www.wikidata.org/wiki/Q7310809","display_name":"Relative value","level":2,"score":0.2989000082015991},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.2808000147342682},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.2581000030040741},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.2531999945640564},{"id":"https://openalex.org/C97399411","wikidata":"https://www.wikidata.org/wiki/Q825367","display_name":"Coin flipping","level":2,"score":0.25119999051094055},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.25}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.28707","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.28707","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.28707","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.28707","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Critic-free":[0],"reinforcement":[1],"learning":[2,78],"with":[3,82,100],"verifiable":[4,148],"rewards":[5],"(RLVR),":[6],"exemplified":[7],"by":[8,95],"Group":[9],"Relative":[10],"Policy":[11],"Optimization":[12],"(GRPO),":[13],"avoids":[14],"training":[15,155],"a":[16,56,88,113,117,139],"value":[17],"function":[18],"(critic)":[19],"and":[20,23,46,68,125,132,135,157,159],"reduces":[21],"memory":[22],"compute":[24],"overhead":[25],"relative":[26],"to":[27,129],"critic-based":[28],"PPO":[29],"pipelines":[30],"for":[31,74,109,142],"aligning":[32],"large":[33],"language":[34],"models.":[35],"However,":[36],"GRPO-style":[37],"advantage":[38,93,141],"estimation":[39,94],"depends":[40],"on":[41,147],"prompt-local":[42,97,133],"(within-prompt-group)":[43],"reward":[44,64,107],"statistics":[45,99,137],"can":[47],"be":[48],"unstable.":[49],"In":[50],"particular,":[51],"when":[52],"all":[53],"rollouts":[54],"in":[55,79,162],"prompt":[57],"group":[58,69],"receive":[59],"identical":[60],"rewards,":[61],"the":[62,121],"within-group":[63],"variance":[65,136],"becomes":[66],"zero,":[67],"normalization":[70],"yields":[71],"zero":[72],"advantages":[73],"that":[75,91,152],"group,":[76],"impeding":[77],"cold-start":[80],"regimes":[81,163],"binary":[83],"verifiers.":[84],"We":[85],"introduce":[86],"BV-Blend,":[87],"critic-free":[89],"framework":[90],"stabilizes":[92],"combining":[96],"on-policy":[98],"semantic-cluster-conditioned":[101],"historical":[102,131],"moments.":[103],"BV-Blend":[104,153],"maintains":[105],"EMA-tracked":[106],"moments":[108],"each":[110],"cluster,":[111],"derives":[112],"confidence":[114],"weight":[115,128],"from":[116],"standard":[118],"error":[119],"of":[120],"mean":[122],"(SEM)":[123],"proxy,":[124],"uses":[126],"this":[127],"blend":[130],"baseline":[134],"into":[138],"standardized":[140],"PPO-style":[143],"clipped":[144],"updates.":[145],"Experiments":[146],"reasoning":[149],"benchmarks":[150],"show":[151],"improves":[154],"stability":[156],"performance,":[158],"remains":[160],"robust":[161],"where":[164],"group-normalized":[165],"methods":[166],"may":[167],"stall.":[168]},"counts_by_year":[],"updated_date":"2026-07-01T06:29:00.853634","created_date":"2026-07-01T00:00:00"}