{"id":"https://openalex.org/W7162069534","doi":"https://doi.org/10.48550/arxiv.2605.21854","title":"CrossVLA: Cross-Paradigm Post-Training and Inference Optimization for Vision-Language-Action Models","display_name":"CrossVLA: Cross-Paradigm Post-Training and Inference Optimization for Vision-Language-Action Models","publication_year":2026,"publication_date":"2026-05-21","ids":{"openalex":"https://openalex.org/W7162069534","doi":"https://doi.org/10.48550/arxiv.2605.21854"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.21854","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.21854","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.2605.21854","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136766521","display_name":"Zhi Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Liu, Zhi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5136766521"],"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":0.9542999863624573,"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":0.9542999863624573,"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":0.0142000000923872,"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":0.008899999782443047,"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/estimator","display_name":"Estimator","score":0.5931000113487244},{"id":"https://openalex.org/keywords/autoregressive-model","display_name":"Autoregressive model","score":0.5918999910354614},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4749000072479248},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.4438000023365021},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.41940000653266907},{"id":"https://openalex.org/keywords/ode","display_name":"Ode","score":0.4138999879360199},{"id":"https://openalex.org/keywords/cache","display_name":"Cache","score":0.39239999651908875},{"id":"https://openalex.org/keywords/projection","display_name":"Projection (relational algebra)","score":0.3874000012874603},{"id":"https://openalex.org/keywords/object","display_name":"Object (grammar)","score":0.32989999651908875}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6980999708175659},{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.5931000113487244},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.5918999910354614},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4749000072479248},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4490000009536743},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.4438000023365021},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.41940000653266907},{"id":"https://openalex.org/C34862557","wikidata":"https://www.wikidata.org/wiki/Q178985","display_name":"Ode","level":2,"score":0.4138999879360199},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.39239999651908875},{"id":"https://openalex.org/C57493831","wikidata":"https://www.wikidata.org/wiki/Q3134666","display_name":"Projection (relational algebra)","level":2,"score":0.3874000012874603},{"id":"https://openalex.org/C2781238097","wikidata":"https://www.wikidata.org/wiki/Q175026","display_name":"Object (grammar)","level":2,"score":0.32989999651908875},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.30660000443458557},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.3057999908924103},{"id":"https://openalex.org/C42812","wikidata":"https://www.wikidata.org/wiki/Q1082910","display_name":"Partition (number theory)","level":2,"score":0.2989000082015991},{"id":"https://openalex.org/C61423126","wikidata":"https://www.wikidata.org/wiki/Q187432","display_name":"Scripting language","level":2,"score":0.29600000381469727},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.29409998655319214},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.28769999742507935},{"id":"https://openalex.org/C3261483","wikidata":"https://www.wikidata.org/wiki/Q119565","display_name":"Frame rate","level":2,"score":0.2815000116825104},{"id":"https://openalex.org/C52146309","wikidata":"https://www.wikidata.org/wiki/Q7431116","display_name":"Schema (genetic algorithms)","level":2,"score":0.25949999690055847},{"id":"https://openalex.org/C61224824","wikidata":"https://www.wikidata.org/wiki/Q2260434","display_name":"Mixture model","level":2,"score":0.2581999897956848},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.25619998574256897},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2558000087738037},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.2538999915122986},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.25220000743865967},{"id":"https://openalex.org/C89128539","wikidata":"https://www.wikidata.org/wiki/Q1949963","display_name":"Statistic","level":2,"score":0.2515999972820282},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.25130000710487366}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.21854","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.21854","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.2605.21854","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.21854","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":{"Vision-Language-Action":[0],"(VLA)":[1],"models":[2,37],"have":[3],"rapidly":[4],"converged":[5],"on":[6,44,69,124,127,181],"a":[7,60,77,98,148,153,175,197],"small":[8],"set":[9],"of":[10,53,80,129,142],"architectural":[11],"patterns:":[12],"discrete-token":[13],"autoregression":[14],"(e.g.":[15,20],"OpenVLA)":[16],"and":[17,82,145,160,205],"continuous-action":[18,70],"flow-matching":[19,62],"pi-0.5).":[21],"Yet":[22],"preference":[23],"alignment":[24],"via":[25],"Direct":[26],"Preference":[27],"Optimisation":[28],"(DPO)":[29],"--":[30,38,109,119,157],"the":[31,85,137],"de-facto":[32],"post-training":[33],"step":[34],"in":[35,169],"language":[36],"has":[39],"been":[40],"studied":[41],"almost":[42],"exclusively":[43],"autoregressive":[45],"VLAs.":[46],"We":[47,172],"present":[48],"CrossVLA,":[49],"an":[50,133],"empirical":[51],"study":[52],"cross-paradigm":[54],"VLA":[55,89],"post-training.":[56],"Three":[57],"contributions:":[58],"(i)":[59],"surrogate":[61],"log-probability":[63],"estimator":[64],"that":[65],"lets":[66],"DPO":[67],"operate":[68],"backbones":[71],"without":[72],"probability-flow":[73],"ODE":[74],"integration;":[75],"(ii)":[76],"head-to-head":[78],"comparison":[79],"LoRA":[81],"DoRA":[83,92],"as":[84,196],"parameter-efficient":[86],"layer":[87],"for":[88,189],"DPO,":[90],"finding":[91],"improves":[93],"over":[94,193],"OpenVLA":[95],"SFT":[96],"by":[97],"mean":[99],"+10.4":[100],"pp":[101],"across":[102],"LIBERO":[103,183],"4-suite":[104],"(600":[105],"trials,":[106],"3":[107,130],"seeds)":[108],"per-suite":[110],"+20.0":[111],"Object,":[112],"+11.0":[113],"Long-horizon,":[114],"+8.0":[115],"Goal,":[116],"+2.7":[117],"Spatial":[118],"with":[120],"zero":[121],"seed":[122],"variance":[123],"Object":[125],"(38/50":[126],"each":[128],"seeds);":[131],"(iii)":[132],"inference-time":[134],"anatomy":[135],"showing":[136],"denoise":[138],"loop":[139],"dominates":[140],"78.6%":[141],"sample_actions":[143],"latency":[144],"prefix-K/V":[146],"caching":[147],"la":[149],"VLA-Cache":[150],"caps":[151],"at":[152,210],"21%":[154],"acceleration":[155],"ceiling":[156],"both":[158],"chunk-level":[159],"token-level":[161],"cache":[162],"strategies":[163],"degrade":[164],"success":[165],"rate":[166],"to":[167],"0-80%":[168],"our":[170],"benchmarks.":[171],"further":[173],"pretrain":[174],"multi-view":[176],"+":[177],"temporal":[178],"projection":[179],"head":[180],"6000":[182],"frames,":[184],"achieving":[185],"99.5%":[186],"k-NN":[187],"recall@1":[188],"same-task":[190],"retrieval":[191],"(36x":[192],"random),":[194],"available":[195],"downstream":[198],"initialisation.":[199],"All":[200],"code,":[201],"ckpts,":[202],"training":[203],"logs,":[204],"reproduction":[206],"scripts":[207],"are":[208],"open":[209],"https://github.com/lz-googlefycy/vla-lab.":[211]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-23T00:00:00"}