{"id":"https://openalex.org/W7162393363","doi":"https://doi.org/10.48550/arxiv.2605.25477","title":"EXPO-FT: Sample-Efficient Reinforcement Learning Finetuning for Vision-Language-Action Models","display_name":"EXPO-FT: Sample-Efficient Reinforcement Learning Finetuning for Vision-Language-Action Models","publication_year":2026,"publication_date":"2026-05-25","ids":{"openalex":"https://openalex.org/W7162393363","doi":"https://doi.org/10.48550/arxiv.2605.25477"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2605.25477","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2605.25477","pdf_url":"https://arxiv.org/pdf/2605.25477","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":"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/2605.25477","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5113246490","display_name":"Perry Dong","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Dong, Perry","raw_affiliation_strings":["Stanford University https :// pd-perry "],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Stanford University https :// pd-perry ","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5079506491","display_name":"Kuo-Han Hung","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hung, Kuo-Han","raw_affiliation_strings":["Stanford University https :// pd-perry "],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Stanford University https :// pd-perry ","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137027820","display_name":"Tian Gao","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Gao, Tian","raw_affiliation_strings":["Stanford University https :// pd-perry "],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Stanford University https :// pd-perry ","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5080725225","display_name":"Dorsa Sadigh","orcid":"https://orcid.org/0000-0002-7802-9183","observed_orcids":["https://orcid.org/0000-0002-7802-9183"]},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Sadigh, Dorsa","raw_affiliation_strings":["Stanford University https :// pd-perry "],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Stanford University https :// pd-perry ","institution_ids":["https://openalex.org/I97018004"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5005431772","display_name":"Chelsea Finn","orcid":"https://orcid.org/0000-0001-6298-0874","observed_orcids":["https://orcid.org/0000-0001-6298-0874"]},"institutions":[{"id":"https://openalex.org/I97018004","display_name":"Stanford University","ror":"https://ror.org/00f54p054","country_code":"US","type":"education","lineage":["https://openalex.org/I97018004"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Finn, Chelsea","raw_affiliation_strings":["Stanford University https :// pd-perry "],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Stanford University https :// pd-perry ","institution_ids":["https://openalex.org/I97018004"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I97018004"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":18,"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/T10653","display_name":"Robot Manipulation and Learning","score":0.5044000148773193,"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"}},"topics":[{"id":"https://openalex.org/T10653","display_name":"Robot Manipulation and Learning","score":0.5044000148773193,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.4936999976634979,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.0005000000237487257,"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/vision-language-action-models","display_name":"vision-language-action models","score":0.9470000267028809},{"id":"https://openalex.org/keywords/reinforcement-learning-fine-tuning","display_name":"reinforcement learning fine-tuning","score":0.9380000233650208},{"id":"https://openalex.org/keywords/sample-efficiency","display_name":"sample efficiency","score":0.9559999704360962},{"id":"https://openalex.org/keywords/robotic-manipulation","display_name":"robotic manipulation","score":0.6819999814033508},{"id":"https://openalex.org/keywords/manipulation-tasks","display_name":"manipulation tasks","score":0.35100001096725464}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8097000122070312},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7882999777793884},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.621999979019165},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.5252000093460083},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.5040000081062317},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5031999945640564},{"id":"https://openalex.org/C79581498","wikidata":"https://www.wikidata.org/wiki/Q1367530","display_name":"Suite","level":2,"score":0.44020000100135803},{"id":"https://openalex.org/C51929080","wikidata":"https://www.wikidata.org/wiki/Q2425187","display_name":"Codebase","level":3,"score":0.3917999863624573},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.38850000500679016},{"id":"https://openalex.org/C90509273","wikidata":"https://www.wikidata.org/wiki/Q11012","display_name":"Robot","level":2,"score":0.3725999891757965},{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.36410000920295715},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.362199991941452},{"id":"https://openalex.org/C63000827","wikidata":"https://www.wikidata.org/wiki/Q3080428","display_name":"Software portability","level":2,"score":0.3343000113964081},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.32690000534057617},{"id":"https://openalex.org/C199190896","wikidata":"https://www.wikidata.org/wiki/Q3509276","display_name":"Learning classifier system","level":3,"score":0.2784000039100647},{"id":"https://openalex.org/C4924752","wikidata":"https://www.wikidata.org/wiki/Q184148","display_name":"Plug-in","level":2,"score":0.2727999985218048},{"id":"https://openalex.org/C110332635","wikidata":"https://www.wikidata.org/wiki/Q629498","display_name":"Genetic programming","level":2,"score":0.26109999418258667},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.26030001044273376},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.26010000705718994}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2605.25477","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2605.25477","pdf_url":"https://arxiv.org/pdf/2605.25477","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2605.25477","doi":"https://doi.org/10.48550/arxiv.2605.25477","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.25477","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:2605.25477","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2605.25477","pdf_url":"https://arxiv.org/pdf/2605.25477","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G6022427645","display_name":"CAREER: Safe and Influencing Interactions for Human-Robot Systems","funder_award_id":"1941722","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320337345","display_name":"Office of Naval Research","ror":"https://ror.org/00rk2pe57"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7162393363.pdf","grobid_xml":"https://content.openalex.org/works/W7162393363.grobid-xml"},"referenced_works_count":25,"referenced_works":["https://openalex.org/W1982803779","https://openalex.org/W2100235553","https://openalex.org/W2125612430","https://openalex.org/W2554830522","https://openalex.org/W2895144567","https://openalex.org/W2904246096","https://openalex.org/W2949650786","https://openalex.org/W4377865266","https://openalex.org/W4401416214","https://openalex.org/W4405254492","https://openalex.org/W4413362511","https://openalex.org/W4414079285","https://openalex.org/W4414581628","https://openalex.org/W4414674749","https://openalex.org/W4415164172","https://openalex.org/W4415251629","https://openalex.org/W4415795631","https://openalex.org/W4416398156","https://openalex.org/W4416453792","https://openalex.org/W4416967268","https://openalex.org/W7103139230","https://openalex.org/W7155175472","https://openalex.org/W7170084358","https://openalex.org/W7171815610","https://openalex.org/W7204808299"],"related_works":["https://openalex.org/W7213529142","https://openalex.org/W4416157228","https://openalex.org/W7203881002","https://openalex.org/W4414079285","https://openalex.org/W4414581628","https://openalex.org/W7131476132","https://openalex.org/W7157012058","https://openalex.org/W4414050921","https://openalex.org/W4416255006","https://openalex.org/W4399695759"],"abstract_inverted_index":{"The":[0],"ability":[1],"to":[2,48,114,143],"efficiently":[3],"and":[4,72,110,125,141,173],"reliably":[5],"learn":[6],"new":[7],"tasks":[8,158],"has":[9],"been":[10],"a":[11,45,82,100,119,123,127,130],"foundational":[12],"challenge":[13],"in":[14,195],"robotics.":[15,196],"Vision-Language-Action":[16],"(VLA)":[17],"models":[18,194],"have":[19],"demonstrated":[20],"strong":[21],"generalization":[22],"across":[23,155],"diverse":[24],"manipulation":[25,104],"tasks,":[26,105],"yet":[27],"pretrained":[28,62,90],"policies":[29,92],"consistently":[30],"fall":[31],"short":[32],"of":[33,89,102,136,162,165,185,189,192],"the":[34,69,112,183],"reliability":[35],"required":[36],"for":[37,84],"real-world":[38],"deployment.":[39],"Reinforcement":[40],"learning":[41],"(RL)":[42],"fine-tuning":[43],"offers":[44],"promising":[46],"path":[47],"bridge":[49],"this":[50,95],"gap,":[51],"but":[52],"existing":[53],"approaches":[54],"either":[55],"train":[56],"from":[57],"scratch":[58],"without":[59,67],"fully":[60],"leveraging":[61],"priors,":[63],"or":[64],"fine-tune":[65],"VLAs":[66],"achieving":[68],"sample":[70],"efficiency":[71],"success":[73],"rates":[74],"that":[75,93],"practical":[76],"deployment":[77],"demands.":[78],"We":[79,177],"present":[80],"EXPO-FT,":[81],"system":[83,98,148],"stable,":[85],"sample-efficient":[86],"RL":[87,190],"finetuning":[88,175,191],"VLA":[91,174,193],"closes":[94],"gap.":[96],"Our":[97,147],"solves":[99],"suite":[101],"challenging":[103],"including":[106],"routing":[107],"string":[108],"lights":[109],"inserting":[111,126],"plug":[113],"light":[115],"it":[116],"up,":[117],"striking":[118],"pool":[120],"ball":[121],"into":[122,129],"pocket,":[124],"flower":[128],"wine":[131],"bottle,":[132],"each":[133],"requiring":[134],"combinations":[135],"high":[137],"precision,":[138],"dynamic":[139],"actions,":[140],"robustness":[142],"varied":[144],"initial":[145],"states.":[146],"achieves":[149],"perfect":[150],"task":[151],"performance":[152],"(30/30":[153],"successes)":[154],"all":[156],"evaluated":[157],"within":[159],"an":[160,179],"average":[161],"19.1":[163],"minutes":[164],"online":[166],"robot":[167],"data,":[168],"outperforming":[169],"both":[170],"prior":[171],"RL-from-scratch":[172],"approaches.":[176],"release":[178],"open-source":[180],"codebase":[181],"with":[182],"aim":[184],"facilitating":[186],"broader":[187],"adoption":[188]},"counts_by_year":[{"year":2026,"cited_by_count":18}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-05-27T00:00:00"}