{"id":"https://openalex.org/W4413926098","doi":"https://doi.org/10.1109/icra55743.2025.11127931","title":"Inference-Time Policy Steering Through Human Interactions","display_name":"Inference-Time Policy Steering Through Human Interactions","publication_year":2025,"publication_date":"2025-05-19","ids":{"openalex":"https://openalex.org/W4413926098","doi":"https://doi.org/10.1109/icra55743.2025.11127931"},"language":"en","primary_location":{"id":"doi:10.1109/icra55743.2025.11127931","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icra55743.2025.11127931","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE International Conference on Robotics and Automation (ICRA)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100642503","display_name":"Yanwei Wang","orcid":"https://orcid.org/0000-0002-8067-1909"},"institutions":[{"id":"https://openalex.org/I4405258931","display_name":"MIT Computer Science and Artificial Intelligence Laboratory","ror":"https://ror.org/01e85r662","country_code":"US","type":"facility","lineage":["https://openalex.org/I4405258931","https://openalex.org/I63966007"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yanwei Wang","raw_affiliation_strings":["MIT CSAIL"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"MIT CSAIL","institution_ids":["https://openalex.org/I4405258931"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5007415599","display_name":"Lirui Wang","orcid":"https://orcid.org/0000-0001-9285-7084"},"institutions":[{"id":"https://openalex.org/I4405258931","display_name":"MIT Computer Science and Artificial Intelligence Laboratory","ror":"https://ror.org/01e85r662","country_code":"US","type":"facility","lineage":["https://openalex.org/I4405258931","https://openalex.org/I63966007"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Lirui Wang","raw_affiliation_strings":["MIT CSAIL"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"MIT CSAIL","institution_ids":["https://openalex.org/I4405258931"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041777043","display_name":"Yilun Du","orcid":null},"institutions":[{"id":"https://openalex.org/I4405258931","display_name":"MIT Computer Science and Artificial Intelligence Laboratory","ror":"https://ror.org/01e85r662","country_code":"US","type":"facility","lineage":["https://openalex.org/I4405258931","https://openalex.org/I63966007"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yilun Du","raw_affiliation_strings":["MIT CSAIL"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"MIT CSAIL","institution_ids":["https://openalex.org/I4405258931"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5053446613","display_name":"Balakumar Sundaralingam","orcid":"https://orcid.org/0000-0003-3106-4755"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Balakumar Sundaralingam","raw_affiliation_strings":["NVIDIA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Xuning Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xuning Yang","raw_affiliation_strings":["NVIDIA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5067607006","display_name":"Yu-Wei Chao","orcid":"https://orcid.org/0000-0002-5476-4343"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yu-Wei Chao","raw_affiliation_strings":["NVIDIA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5005763357","display_name":"Claudia P\u00e9rez-D\u2019Arpino","orcid":"https://orcid.org/0000-0002-2949-4214"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Claudia P\u00e9rez-D\u2019Arpino","raw_affiliation_strings":["NVIDIA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5108257764","display_name":"Dieter Fox","orcid":"https://orcid.org/0009-0009-4694-9127"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dieter Fox","raw_affiliation_strings":["NVIDIA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"NVIDIA","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5044369720","display_name":"Julie Shah","orcid":"https://orcid.org/0000-0003-1338-8107"},"institutions":[{"id":"https://openalex.org/I4405258931","display_name":"MIT Computer Science and Artificial Intelligence Laboratory","ror":"https://ror.org/01e85r662","country_code":"US","type":"facility","lineage":["https://openalex.org/I4405258931","https://openalex.org/I63966007"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Julie Shah","raw_affiliation_strings":["MIT CSAIL"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"MIT CSAIL","institution_ids":["https://openalex.org/I4405258931"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"15626","last_page":"15633"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11810","display_name":"Complex Systems and Decision Making","score":0.09510000050067902,"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"}},"topics":[{"id":"https://openalex.org/T11810","display_name":"Complex Systems and Decision Making","score":0.09510000050067902,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6230832934379578},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5858350992202759},{"id":"https://openalex.org/keywords/human\u2013computer-interaction","display_name":"Human\u2013computer interaction","score":0.33075064420700073},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.2637579143047333}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6230832934379578},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5858350992202759},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.33075064420700073},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2637579143047333}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icra55743.2025.11127931","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icra55743.2025.11127931","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE International Conference on Robotics and Automation (ICRA)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":23,"referenced_works":["https://openalex.org/W2076332785","https://openalex.org/W2103450459","https://openalex.org/W2784881541","https://openalex.org/W2794908222","https://openalex.org/W3016541531","https://openalex.org/W3150635270","https://openalex.org/W3208223622","https://openalex.org/W4226091687","https://openalex.org/W4240592325","https://openalex.org/W4293498911","https://openalex.org/W4312824283","https://openalex.org/W4367016688","https://openalex.org/W4383109488","https://openalex.org/W4383172031","https://openalex.org/W4385430497","https://openalex.org/W4385430539","https://openalex.org/W4385430588","https://openalex.org/W4385430674","https://openalex.org/W4388430323","https://openalex.org/W4402354098","https://openalex.org/W4402354171","https://openalex.org/W4403337227","https://openalex.org/W4405779052"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"Generative":[0],"policies":[1],"trained":[2],"with":[3,62,123],"human":[4,43,63,79,108],"demonstrations":[5],"can":[6],"autonomously":[7],"accomplish":[8],"multimodal,":[9],"longhorizon":[10],"tasks.":[11],"However,":[12],"during":[13],"inference,":[14],"humans":[15],"are":[16,136],"often":[17],"removed":[18],"from":[19],"the":[20,25,83,90,127],"policy":[21,31,60,91,125],"execution":[22,55],"loop,":[23],"limiting":[24],"ability":[26],"to":[27,51,81],"guide":[28],"a":[29,33],"pre-trained":[30],"towards":[32],"specific":[34],"sub-goal":[35],"or":[36,54],"trajectory":[37],"shape":[38],"among":[39],"multiple":[40],"predictions.":[41],"Naive":[42],"intervention":[44],"may":[45],"inadvertently":[46],"exacerbate":[47],"distribution":[48,133],"shift,":[49],"leading":[50],"constraint":[52],"violations":[53],"failures.":[56],"To":[57],"better":[58],"align":[59],"output":[61],"intent":[64],"without":[65],"inducing":[66],"out-of-distribution":[67],"errors,":[68],"we":[69],"propose":[70],"an":[71],"Inference-Time":[72],"Policy":[73],"Steering":[74],"(ITPS)":[75],"framework":[76],"that":[77],"leverages":[78],"interactions":[80],"bias":[82],"generative":[84],"sampling":[85,117,122],"process,":[86],"rather":[87],"than":[88],"finetuning":[89],"on":[92],"interaction":[93,109],"data.":[94],"We":[95],"evaluate":[96],"ITPS":[97],"across":[98],"three":[99,105],"simulated":[100],"and":[101,110,132],"real-world":[102],"benchmarks,":[103],"testing":[104],"forms":[106],"of":[107],"associated":[111],"alignment":[112,131],"distance":[113],"metrics.":[114],"Among":[115],"six":[116],"strategies,":[118],"our":[119],"proposed":[120],"stochastic":[121],"diffusion":[124],"achieves":[126],"best":[128],"trade-off":[129],"between":[130],"shift.":[134],"Videos":[135],"available":[137],"at":[138],"https://yanweiw.github.io/itps/.":[139]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}