{"id":"https://openalex.org/W7204228970","doi":"https://doi.org/10.48550/arxiv.2608.24479","title":"WarpSAC: Towards the Pinnacle of Scalable Off-policy RL by Rethinking Exploration and Exploitation","display_name":"WarpSAC: Towards the Pinnacle of Scalable Off-policy RL by Rethinking Exploration and Exploitation","publication_year":2026,"publication_date":"2026-08-25","ids":{"openalex":"https://openalex.org/W7204228970","doi":"https://doi.org/10.48550/arxiv.2608.24479"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.24479","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24479","pdf_url":"https://arxiv.org/pdf/2608.24479","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/2608.24479","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5148274862","display_name":"Zihao Wu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I162868743","display_name":"Tianjin University","ror":"https://ror.org/012tb2g32","country_code":"CN","type":"education","lineage":["https://openalex.org/I162868743"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wu, Zihao","raw_affiliation_strings":["Tianjin University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tianjin University ,","institution_ids":["https://openalex.org/I162868743"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138465306","display_name":"Hongyao Tang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I162868743","display_name":"Tianjin University","ror":"https://ror.org/012tb2g32","country_code":"CN","type":"education","lineage":["https://openalex.org/I162868743"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Tang, Hongyao","raw_affiliation_strings":["Tianjin University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tianjin University ,","institution_ids":["https://openalex.org/I162868743"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138422067","display_name":"Yi Ma","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I181877577","display_name":"Shanxi University","ror":"https://ror.org/03y3e3s17","country_code":"CN","type":"education","lineage":["https://openalex.org/I181877577"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Ma, Yi","raw_affiliation_strings":["Shanxi University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanxi University ,","institution_ids":["https://openalex.org/I181877577"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5035638865","display_name":"Huizhong Song","orcid":"https://orcid.org/0009-0004-7520-6692","observed_orcids":["https://orcid.org/0009-0004-7520-6692","https://orcid.org/0000-0002-8829-1062"]},"institutions":[{"id":"https://openalex.org/I181877577","display_name":"Shanxi University","ror":"https://ror.org/03y3e3s17","country_code":"CN","type":"education","lineage":["https://openalex.org/I181877577"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Song, Huizhong","raw_affiliation_strings":["Shanxi University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanxi University ,","institution_ids":["https://openalex.org/I181877577"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148263376","display_name":"Pengyi Li","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I162868743","display_name":"Tianjin University","ror":"https://ror.org/012tb2g32","country_code":"CN","type":"education","lineage":["https://openalex.org/I162868743"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Li, Pengyi","raw_affiliation_strings":["Tianjin University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tianjin University ,","institution_ids":["https://openalex.org/I162868743"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126344606","display_name":"Yifu Yuan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I162868743","display_name":"Tianjin University","ror":"https://ror.org/012tb2g32","country_code":"CN","type":"education","lineage":["https://openalex.org/I162868743"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yuan, Yifu","raw_affiliation_strings":["Tianjin University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tianjin University ,","institution_ids":["https://openalex.org/I162868743"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126338504","display_name":"Fei Ni","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I47508984","display_name":"Imperial College London","ror":"https://ror.org/041kmwe10","country_code":"GB","type":"education","lineage":["https://openalex.org/I47508984"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Ni, Fei","raw_affiliation_strings":["Imperial College London"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Imperial College London","institution_ids":["https://openalex.org/I47508984"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148282787","display_name":"Jinyi Liu","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I162868743","display_name":"Tianjin University","ror":"https://ror.org/012tb2g32","country_code":"CN","type":"education","lineage":["https://openalex.org/I162868743"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Liu, Jinyi","raw_affiliation_strings":["Tianjin University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tianjin University ,","institution_ids":["https://openalex.org/I162868743"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137626880","display_name":"Wei Wei","orcid":"https://orcid.org/0000-0002-0655-056X","observed_orcids":["https://orcid.org/0000-0002-0655-056X","https://orcid.org/0000-0003-0875-5603","https://orcid.org/0000-0002-3173-6905"]},"institutions":[{"id":"https://openalex.org/I181877577","display_name":"Shanxi University","ror":"https://ror.org/03y3e3s17","country_code":"CN","type":"education","lineage":["https://openalex.org/I181877577"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wei, Wei","raw_affiliation_strings":["Shanxi University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Shanxi University ,","institution_ids":["https://openalex.org/I181877577"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148335543","display_name":"Jianrong Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I162868743","display_name":"Tianjin University","ror":"https://ror.org/012tb2g32","country_code":"CN","type":"education","lineage":["https://openalex.org/I162868743"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wang, Jianrong","raw_affiliation_strings":["Tianjin University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tianjin University ,","institution_ids":["https://openalex.org/I162868743"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138697700","display_name":"Yan Zheng","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I162868743","display_name":"Tianjin University","ror":"https://ror.org/012tb2g32","country_code":"CN","type":"education","lineage":["https://openalex.org/I162868743"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zheng, Yan","raw_affiliation_strings":["Tianjin University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tianjin University ,","institution_ids":["https://openalex.org/I162868743"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5047509839","display_name":"Jianye Hao","orcid":"https://orcid.org/0000-0002-0422-8235","observed_orcids":["https://orcid.org/0000-0002-0422-8235"]},"institutions":[{"id":"https://openalex.org/I162868743","display_name":"Tianjin University","ror":"https://ror.org/012tb2g32","country_code":"CN","type":"education","lineage":["https://openalex.org/I162868743"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hao, Jianye","raw_affiliation_strings":["Tianjin University ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Tianjin University ,","institution_ids":["https://openalex.org/I162868743"]}]}],"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":1.0,"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":1.0,"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/T11195","display_name":"Simulation Techniques and Applications","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"}},{"id":"https://openalex.org/T12794","display_name":"Adaptive Dynamic Programming Control","score":0.0,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/off-policy-learning","display_name":"off-policy learning","score":0.9929999709129333},{"id":"https://openalex.org/keywords/exploration-exploitation-trade-off","display_name":"exploration-exploitation balance","score":0.8289999961853027},{"id":"https://openalex.org/keywords/parameter-normalization","display_name":"parameter normalization","score":0.6690000295639038},{"id":"https://openalex.org/keywords/massively-parallel-simulation","display_name":"massively parallel simulation","score":0.7350000143051147},{"id":"https://openalex.org/keywords/unitree-g1","display_name":"Unitree G1","score":0.5899999737739563}],"concepts":[{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.746399998664856},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7211999893188477},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.6877999901771545},{"id":"https://openalex.org/C183115368","wikidata":"https://www.wikidata.org/wiki/Q856577","display_name":"Weighting","level":2,"score":0.6122000217437744},{"id":"https://openalex.org/C136886441","wikidata":"https://www.wikidata.org/wiki/Q926129","display_name":"Normalization (sociology)","level":2,"score":0.6057000160217285},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5425999760627747},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.4749999940395355},{"id":"https://openalex.org/C75917345","wikidata":"https://www.wikidata.org/wiki/Q2725298","display_name":"Sampling bias","level":3,"score":0.3862999975681305},{"id":"https://openalex.org/C190475519","wikidata":"https://www.wikidata.org/wiki/Q544384","display_name":"Massively parallel","level":2,"score":0.3783999979496002},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3555999994277954},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3522000014781952},{"id":"https://openalex.org/C198531522","wikidata":"https://www.wikidata.org/wiki/Q485146","display_name":"Sample (material)","level":2,"score":0.28769999742507935},{"id":"https://openalex.org/C2776291640","wikidata":"https://www.wikidata.org/wiki/Q2912517","display_name":"Value (mathematics)","level":2,"score":0.27059999108314514},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.2635999917984009},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.26010000705718994}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.24479","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24479","pdf_url":"https://arxiv.org/pdf/2608.24479","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.2608.24479","doi":"https://doi.org/10.48550/arxiv.2608.24479","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.24479","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:2608.24479","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24479","pdf_url":"https://arxiv.org/pdf/2608.24479","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/W7204228970.pdf","grobid_xml":"https://content.openalex.org/works/W7204228970.grobid-xml"},"referenced_works_count":11,"referenced_works":["https://openalex.org/W2781585732","https://openalex.org/W2904246096","https://openalex.org/W3037932933","https://openalex.org/W4387964172","https://openalex.org/W4402354138","https://openalex.org/W4403819516","https://openalex.org/W4414079055","https://openalex.org/W4415337908","https://openalex.org/W7126263443","https://openalex.org/W7133231913","https://openalex.org/W7151342152"],"related_works":["https://openalex.org/W4385262479","https://openalex.org/W7151342152","https://openalex.org/W2978455699","https://openalex.org/W4281258657","https://openalex.org/W4389157295","https://openalex.org/W3204075372","https://openalex.org/W3005089802","https://openalex.org/W4415795631","https://openalex.org/W3023239315","https://openalex.org/W7084148507"],"abstract_inverted_index":{"Massively":[0],"parallel":[1],"simulation":[2],"changes":[3],"the":[4,177],"data":[5,47,179],"regime":[6],"in":[7,56],"which":[8],"off-policy":[9,83,170],"reinforcement":[10],"learning":[11,63],"(RL)":[12],"is":[13],"trained,":[14],"challenging":[15],"stabilizers":[16,32,175],"designed":[17],"for":[18,91,103,112],"data-limited":[19,104],"replay.":[20],"Through":[21],"controlled":[22],"experiments":[23],"across":[24,65,125,131],"eight":[25],"benchmark":[26],"families,":[27],"we":[28,76],"show":[29,167],"that":[30,168],"these":[31,74],"are":[33,48],"data-regime-dependent:":[34],"parameter":[35],"normalization":[36],"helps":[37],"with":[38,68],"narrow":[39],"replay":[40,60],"coverage":[41],"but":[42],"restricts":[43],"value":[44],"fitting":[45],"when":[46],"abundant,":[49],"while":[50],"clipped":[51,101],"double-Q":[52],"can":[53],"be":[54],"relaxed":[55],"high-throughput":[57],"manipulation.":[58],"Age-biased":[59],"weighting":[61],"improves":[62,117,144],"efficiency":[64],"regimes,":[66],"especially":[67],"limited":[69],"network":[70],"capacity.":[71],"Based":[72],"on":[73,149,160],"findings,":[75],"propose":[77],"WarpSAC,":[78],"a":[79],"regime-aware":[80],"family":[81],"of":[82],"RL":[84,171],"algorithms.":[85],"WarpSAC":[86,116],"uses":[87],"Sample":[88],"Weight":[89],"Decay":[90],"efficient":[92],"exploitation":[93],"and":[94,107,129,154],"provides":[95],"two":[96],"variants:":[97],"WarpSAC-L":[98],"(Norm":[99,109],"ON,":[100],"double-Q)":[102],"CPU-scale":[105,127],"training,":[106],"WarpSAC-A":[108],"OFF,":[110],"single-Q)":[111],"data-abundant":[113],"GPU-parallel":[114,133],"training.":[115],"normalized":[118,146],"score--step":[119],"AUC":[120,148],"over":[121],"FlashSAC":[122],"by":[123,152],"4.5%":[124],"nine":[126],"environments":[128],"23.1%":[130],"fourteen":[132],"environments.":[134],"It":[135],"increases":[136],"UnitreeG1TransportBox-v1":[137],"success":[138],"rate":[139],"from":[140],"19.8%":[141],"to":[142,176],"96.4%,":[143],"mean":[145],"wall-time":[147],"MuJoCo":[150],"Playground":[151],"19.1%,":[153],"achieves":[155],"36.4%":[156],"faster":[157],"sim-to-real":[158],"deployment":[159],"Unitree":[161],"G1":[162],"than":[163],"FlashSAC.":[164],"These":[165],"results":[166],"scalable":[169],"should":[172],"adapt":[173],"its":[174],"available":[178],"regime.":[180]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-08-27T00:00:00"}