{"id":"https://openalex.org/W7127491001","doi":"https://doi.org/10.48550/arxiv.2602.00686","title":"Learning to Accelerate Vision-Language-Action Models through Adaptive Visual Token Caching","display_name":"Learning to Accelerate Vision-Language-Action Models through Adaptive Visual Token Caching","publication_year":2026,"publication_date":"2026-01-31","ids":{"openalex":"https://openalex.org/W7127491001","doi":"https://doi.org/10.48550/arxiv.2602.00686"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2602.00686","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5124936318","display_name":"Yujie Wei","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Wei, Yujie","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124888262","display_name":"Jiahan Fan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fan, Jiahan","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124964613","display_name":"Jiyu Guo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Jiyu","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5080207705","display_name":"Ruichen Zhen","orcid":"https://orcid.org/0000-0002-1212-6538"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhen, Ruichen","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124880909","display_name":"Rui Shao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shao, Rui","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102814914","display_name":"Xiu Su","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Su, Xiu","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124842349","display_name":"Zeke Xie","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xie, Zeke","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5124930361","display_name":"Shuo Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Shuo","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5124936318"],"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.8335000276565552,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.8335000276565552,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10036","display_name":"Advanced Neural Network Applications","score":0.05530000105500221,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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":0.015799999237060547,"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/security-token","display_name":"Security token","score":0.6478999853134155},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6230999827384949},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.5874999761581421},{"id":"https://openalex.org/keywords/cache","display_name":"Cache","score":0.5644000172615051},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.527999997138977},{"id":"https://openalex.org/keywords/overhead","display_name":"Overhead (engineering)","score":0.5023999810218811},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.486299991607666},{"id":"https://openalex.org/keywords/heuristic","display_name":"Heuristic","score":0.448199987411499},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.4302999973297119},{"id":"https://openalex.org/keywords/acceleration","display_name":"Acceleration","score":0.38190001249313354}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8428999781608582},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.6478999853134155},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6230999827384949},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.5874999761581421},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.5644000172615051},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.5547999739646912},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.527999997138977},{"id":"https://openalex.org/C2779960059","wikidata":"https://www.wikidata.org/wiki/Q7113681","display_name":"Overhead (engineering)","level":2,"score":0.5023999810218811},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.486299991607666},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4657999873161316},{"id":"https://openalex.org/C173801870","wikidata":"https://www.wikidata.org/wiki/Q201413","display_name":"Heuristic","level":2,"score":0.448199987411499},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.4302999973297119},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.38190001249313354},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3774000108242035},{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.37540000677108765},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.37459999322891235},{"id":"https://openalex.org/C12713177","wikidata":"https://www.wikidata.org/wiki/Q1900281","display_name":"Perspective (graphical)","level":2,"score":0.37279999256134033},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.3675000071525574},{"id":"https://openalex.org/C2776650193","wikidata":"https://www.wikidata.org/wiki/Q264661","display_name":"Obstacle","level":2,"score":0.3547999858856201},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.32749998569488525},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.32280001044273376},{"id":"https://openalex.org/C2777472644","wikidata":"https://www.wikidata.org/wiki/Q16968992","display_name":"Approximate inference","level":3,"score":0.32019999623298645},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.32010000944137573},{"id":"https://openalex.org/C115067241","wikidata":"https://www.wikidata.org/wiki/Q1639854","display_name":"Token passing","level":3,"score":0.3158000111579895},{"id":"https://openalex.org/C174998907","wikidata":"https://www.wikidata.org/wiki/Q357662","display_name":"Work in process","level":2,"score":0.3091999888420105},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.30820000171661377},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.288100004196167},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.28780001401901245},{"id":"https://openalex.org/C29202148","wikidata":"https://www.wikidata.org/wiki/Q287260","display_name":"Resource allocation","level":2,"score":0.28529998660087585},{"id":"https://openalex.org/C138236772","wikidata":"https://www.wikidata.org/wiki/Q25098575","display_name":"Edge device","level":3,"score":0.2815000116825104},{"id":"https://openalex.org/C2776459999","wikidata":"https://www.wikidata.org/wiki/Q2119376","display_name":"Fidelity","level":2,"score":0.27950000762939453},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.2782000005245209},{"id":"https://openalex.org/C159023740","wikidata":"https://www.wikidata.org/wiki/Q623276","display_name":"Deadlock","level":2,"score":0.27730000019073486},{"id":"https://openalex.org/C106189395","wikidata":"https://www.wikidata.org/wiki/Q176789","display_name":"Markov decision process","level":3,"score":0.2720000147819519},{"id":"https://openalex.org/C117251300","wikidata":"https://www.wikidata.org/wiki/Q1849855","display_name":"Parametric statistics","level":2,"score":0.26420000195503235},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.25870001316070557},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.25690001249313354},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.25040000677108765}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2602.00686","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"doi:10.48550/arxiv.2602.00686","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.00686","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":"article"}],"best_oa_location":{"id":"pmh:doi:10.48550/arxiv.2602.00686","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"sustainable_development_goals":[{"score":0.6526364684104919,"id":"https://metadata.un.org/sdg/16","display_name":"Peace, Justice and strong institutions"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Vision-Language-Action":[0],"(VLA)":[1],"models":[2,215],"have":[3],"demonstrated":[4],"remarkable":[5],"generalization":[6],"capabilities":[7],"in":[8],"robotic":[9,32],"manipulation":[10],"tasks,":[11,194],"yet":[12],"their":[13,131],"substantial":[14],"computational":[15,207],"overhead":[16],"remains":[17],"a":[18,71,78,83,101,113,139,165],"critical":[19],"obstacle":[20],"to":[21,57,59,122,130,183],"real-world":[22,193],"deployment.":[23],"Improving":[24],"inference":[25,68,168],"efficiency":[26],"is":[27,127],"therefore":[28],"essential":[29],"for":[30,213],"practical":[31],"applications.":[33],"Existing":[34],"acceleration":[35,69],"methods":[36],"often":[37],"rely":[38],"on":[39,149,185,192],"heuristic":[40],"or":[41,48],"static":[42],"strategies--such":[43],"as":[44,70,155,157],"rule-based":[45],"token":[46],"caching":[47],"pruning--that":[49],"are":[50,96,217],"decoupled":[51],"from":[52],"task":[53],"objectives":[54],"and":[55,76,112,152,187,220],"fail":[56],"adapt":[58],"dynamic":[60],"scene":[61],"changes.":[62],"In":[63],"this":[64,136],"work,":[65],"we":[66],"reformulate":[67],"learnable":[72],"policy":[73],"optimization":[74],"problem":[75],"propose":[77],"novel":[79],"framework":[80],"that":[81,142,161,216],"integrates":[82],"dynamic,":[84],"task-aware":[85,206],"decision-making":[86],"process":[87],"directly":[88],"into":[89],"the":[90,150,173,202,211],"VLA":[91,214],"model.":[92],"At":[93],"its":[94],"core":[95],"two":[97],"lightweight,":[98],"cooperative":[99],"modules:":[100],"Cached":[102],"Token":[103],"Selector,":[104],"which":[105,107,117],"determines":[106],"tokens":[108,121],"should":[109],"be":[110],"reused,":[111],"Cache":[114],"Ratio":[115],"Predictor,":[116],"controls":[118],"how":[119],"many":[120],"reuse.":[123],"Training":[124],"these":[125],"modules":[126],"non-trivial":[128],"due":[129],"discrete":[132],"decisions.":[133],"We":[134],"address":[135],"by":[137,177,188],"adopting":[138],"differentiable":[140],"relaxation":[141],"allows":[143],"gradient-based":[144],"end-to-end":[145],"optimization.":[146],"Extensive":[147],"experiments":[148],"LIBERO":[151,186],"SIMPLER":[153],"benchmarks,":[154],"well":[156],"real-robot":[158],"evaluations,":[159],"show":[160],"our":[162],"method":[163],"achieves":[164],"1.76x":[166],"wall-clock":[167],"speedup":[169],"while":[170],"simultaneously":[171],"improving":[172],"average":[174],"success":[175],"rate":[176],"1.9":[178],"percentage":[179,190],"points":[180,191],"(from":[181],"75.0%":[182],"76.9%)":[184],"5.0":[189],"significantly":[195],"outperforming":[196],"existing":[197],"baselines.":[198],"This":[199],"work":[200],"highlights":[201],"potential":[203],"of":[204],"learning":[205],"allocation":[208],"policies,":[209],"paving":[210],"way":[212],"both":[218],"powerful":[219],"efficient.":[221]},"counts_by_year":[],"updated_date":"2026-04-04T16:13:02.066488","created_date":"2026-02-04T00:00:00"}