{"id":"https://openalex.org/W7164934358","doi":"https://doi.org/10.48550/arxiv.2606.14801","title":"QPILOTS: Efficient Test-Time Q-Steering for Flow Policies","display_name":"QPILOTS: Efficient Test-Time Q-Steering for Flow Policies","publication_year":2026,"publication_date":"2026-06-11","ids":{"openalex":"https://openalex.org/W7164934358","doi":"https://doi.org/10.48550/arxiv.2606.14801"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.14801","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.14801","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.2606.14801","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5037980400","display_name":"Yifan Ruan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ruan, Yifan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104253466","display_name":"Chenyang Cao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cao, Chenyang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102974710","display_name":"Andreas Burger","orcid":"https://orcid.org/0000-0001-7417-3527"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Burger, Andreas","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5055670823","display_name":"Ali Pesaranghader","orcid":"https://orcid.org/0000-0002-5971-6180"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pesaranghader, Ali","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004976943","display_name":"Kaveh Kamali","orcid":"https://orcid.org/0000-0002-7154-9688"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kamali, Kaveh","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138709812","display_name":"Jaehong Kim","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kim, Jaehong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5080873211","display_name":"Nandita Vijaykumar","orcid":"https://orcid.org/0000-0003-3315-9336"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Vijaykumar, Nandita","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5071495561","display_name":"Al\u00e1n Aspuru\u2010Guzik","orcid":"https://orcid.org/0000-0002-8277-4434"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Aspuru-Guzik, Alan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138756490","display_name":"Igor Gilitschenski","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gilitschenski, Igor","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5077198404","display_name":"Nicholas Rhinehart","orcid":"https://orcid.org/0000-0003-4242-1236"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rhinehart, Nicholas","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":0.6678000092506409,"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":0.6678000092506409,"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":0.08640000224113464,"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/T11206","display_name":"Model Reduction and Neural Networks","score":0.039500001817941666,"subfield":{"id":"https://openalex.org/subfields/3109","display_name":"Statistical and Nonlinear Physics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.7016000151634216},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.5885000228881836},{"id":"https://openalex.org/keywords/noise-reduction","display_name":"Noise reduction","score":0.5831999778747559},{"id":"https://openalex.org/keywords/action","display_name":"Action (physics)","score":0.5501000285148621},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.5045999884605408},{"id":"https://openalex.org/keywords/reduction","display_name":"Reduction (mathematics)","score":0.476500004529953},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4560000002384186},{"id":"https://openalex.org/keywords/differentiable-function","display_name":"Differentiable function","score":0.45100000500679016}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7529000043869019},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7016000151634216},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.5885000228881836},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.5831999778747559},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.5501000285148621},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5198000073432922},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.5045999884605408},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.476500004529953},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4560000002384186},{"id":"https://openalex.org/C202615002","wikidata":"https://www.wikidata.org/wiki/Q783507","display_name":"Differentiable function","level":2,"score":0.45100000500679016},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.4278999865055084},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.36579999327659607},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.357699990272522},{"id":"https://openalex.org/C38349280","wikidata":"https://www.wikidata.org/wiki/Q1434290","display_name":"Flow (mathematics)","level":2,"score":0.34290000796318054},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.31139999628067017},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.303600013256073},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.273499995470047},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.2720000147819519},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.26969999074935913},{"id":"https://openalex.org/C167879884","wikidata":"https://www.wikidata.org/wiki/Q727568","display_name":"Balanced flow","level":2,"score":0.2694999873638153},{"id":"https://openalex.org/C115680565","wikidata":"https://www.wikidata.org/wiki/Q5977448","display_name":"Gradient method","level":2,"score":0.2606000006198883},{"id":"https://openalex.org/C30814859","wikidata":"https://www.wikidata.org/wiki/Q4119603","display_name":"Video denoising","level":5,"score":0.2538999915122986}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.14801","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.14801","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.2606.14801","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.14801","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":{"Flow-matching":[0],"and":[1,80,121],"diffusion":[2],"policies":[3],"are":[4,105],"expressive":[5],"action":[6,25,101,120],"generators,":[7],"but":[8],"optimizing":[9],"them":[10],"with":[11],"temporal-difference":[12],"reinforcement":[13],"learning":[14],"(RL)":[15],"remains":[16],"difficult.":[17],"Effective":[18],"policy":[19,53,64,78],"extraction":[20],"requires":[21],"exploiting":[22],"the":[23,52,62,66,76,82,95,98,117,123,156],"critic's":[24],"gradient,":[26],"yet":[27],"directly":[28],"backpropagating":[29],"this":[30,45],"signal":[31],"through":[32],"a":[33,55,72,133,144,149,176],"multi-step":[34],"denoising":[35,63,83,90],"process":[36,84],"can":[37],"be":[38],"numerically":[39],"unstable.":[40],"Existing":[41],"methods":[42],"work":[43],"around":[44],"either":[46],"by":[47],"discarding":[48],"gradient":[49,125],"information,":[50],"distilling":[51],"into":[54],"simpler":[56],"one-step":[57],"actor,":[58],"or":[59,186],"repeatedly":[60],"fine-tuning":[61],"as":[65],"critic":[67,96,103,124],"improves.":[68],"We":[69,127,170],"propose":[70],"QPILOTS,":[71],"method":[73],"that":[74,110],"leaves":[75],"original":[77],"unmodified":[79],"steers":[81],"at":[85],"inference":[86],"time.":[87],"At":[88],"each":[89],"step,":[91],"instead":[92],"of":[93,116,165],"evaluating":[94],"on":[97],"noisy":[99],"intermediate":[100,111],"where":[102],"predictions":[104],"unreliable,":[106],"we":[107],"first":[108],"project":[109],"state":[112],"to":[113,174],"an":[114,161],"estimate":[115],"final":[118],"clean":[119],"compute":[122],"there.":[126],"introduce":[128],"two":[129],"variants:":[130],"QPILOTS-U":[131],"uses":[132],"fast":[134],"single-point":[135],"approximation,":[136],"while":[137],"QPILOTS-M":[138],"draws":[139],"differentiable":[140],"posterior":[141],"samples":[142],"via":[143],"learned":[145],"auxiliary":[146],"network.":[147],"On":[148],"standard":[150],"offline-to-online":[151],"RL":[152],"benchmark,":[153],"QPILOTS":[154,173],"achieves":[155],"best":[157],"aggregate":[158],"performance,":[159],"reaching":[160],"average":[162],"success":[163],"rate":[164],"90%":[166],"across":[167,191],"50":[168],"tasks.":[169],"also":[171],"apply":[172],"steer":[175],"large,":[177],"frozen,":[178],"pretrained":[179],"Vision-Language":[180],"Action":[181],"(VLA)":[182],"foundation":[183],"model,":[184],"outperforming":[185],"matching":[187],"prior":[188],"inference-time":[189],"approaches":[190],"six":[192],"manipulation":[193],"tasks":[194],"in":[195],"simulation.":[196]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-17T00:00:00"}