{"id":"https://openalex.org/W7212079237","doi":"https://doi.org/10.48550/arxiv.2609.06009","title":"How to Learn from What a Human Would Avoid? Intervention-Aware World Models with Real-World RL for Dexterous Manipulation","display_name":"How to Learn from What a Human Would Avoid? Intervention-Aware World Models with Real-World RL for Dexterous Manipulation","publication_year":2026,"publication_date":"2026-09-05","ids":{"openalex":"https://openalex.org/W7212079237","doi":"https://doi.org/10.48550/arxiv.2609.06009"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.06009","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.06009","pdf_url":"https://arxiv.org/pdf/2609.06009","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.06009","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5151017676","display_name":"Jiaju Yin","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yin, Jiaju","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5000604178","display_name":"\u5f20\u632f\u8f89","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Zhenhui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5150905274","display_name":"Lixin Xu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Lixin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5150874202","display_name":"Heng Zhang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Heng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5040489898","display_name":"Jun Shao","orcid":"https://orcid.org/0000-0001-8352-0973","observed_orcids":["https://orcid.org/0000-0001-8352-0973"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shao, Jun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134486162","display_name":"Yating Feng","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Feng, Yating","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5070028115","display_name":"Arash Ajoudani","orcid":"https://orcid.org/0000-0002-1261-737X","observed_orcids":["https://orcid.org/0000-0002-1261-737X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ajoudani, Arash","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5100616654","display_name":"Renjing Xu","orcid":"https://orcid.org/0000-0002-0792-8974","observed_orcids":["https://orcid.org/0000-0002-0792-8974"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Renjing","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":0.8881999850273132,"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":0.8881999850273132,"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/T10653","display_name":"Robot Manipulation and Learning","score":0.1111999973654747,"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/T11023","display_name":"Prosthetics and Rehabilitation Robotics","score":0.0,"subfield":{"id":"https://openalex.org/subfields/2204","display_name":"Biomedical Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/dexterous-manipulation","display_name":"dexterous manipulation","score":0.9049999713897705},{"id":"https://openalex.org/keywords/human-in-the-loop-reinforcement-learning","display_name":"human-in-the-loop reinforcement learning","score":0.8700000047683716},{"id":"https://openalex.org/keywords/multi-fingered-robotic-hand","display_name":"multi-fingered hand","score":0.6299999952316284},{"id":"https://openalex.org/keywords/object-grasping","display_name":"object grasping","score":0.7369999885559082},{"id":"https://openalex.org/keywords/whirl","display_name":"whirl","score":0.5550000071525574}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.6590999960899353},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.58160001039505},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.57669997215271},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.557200014591217},{"id":"https://openalex.org/C132010649","wikidata":"https://www.wikidata.org/wiki/Q189222","display_name":"Intuition","level":2,"score":0.48570001125335693},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.4050000011920929},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.3644999861717224},{"id":"https://openalex.org/C17020691","wikidata":"https://www.wikidata.org/wiki/Q139677","display_name":"Operator (biology)","level":5,"score":0.3562999963760376},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.31060001254081726},{"id":"https://openalex.org/C172205157","wikidata":"https://www.wikidata.org/wiki/Q1782962","display_name":"Model predictive control","level":3,"score":0.29840001463890076},{"id":"https://openalex.org/C64876066","wikidata":"https://www.wikidata.org/wiki/Q5141226","display_name":"Cognitive neuroscience of visual object recognition","level":3,"score":0.29429998993873596},{"id":"https://openalex.org/C2780626000","wikidata":"https://www.wikidata.org/wiki/Q5936775","display_name":"Human-in-the-loop","level":2,"score":0.28299999237060547},{"id":"https://openalex.org/C2987834672","wikidata":"https://www.wikidata.org/wiki/Q4677630","display_name":"Action recognition","level":3,"score":0.28060001134872437},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2709999978542328},{"id":"https://openalex.org/C66905080","wikidata":"https://www.wikidata.org/wiki/Q17005494","display_name":"Binary classification","level":3,"score":0.26649999618530273},{"id":"https://openalex.org/C206588197","wikidata":"https://www.wikidata.org/wiki/Q846574","display_name":"Reuse","level":2,"score":0.26350000500679016},{"id":"https://openalex.org/C2780665704","wikidata":"https://www.wikidata.org/wiki/Q959298","display_name":"Intervention (counseling)","level":2,"score":0.26330000162124634},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.26170000433921814},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.25949999690055847},{"id":"https://openalex.org/C27415008","wikidata":"https://www.wikidata.org/wiki/Q7256382","display_name":"Psychological intervention","level":2,"score":0.25519999861717224},{"id":"https://openalex.org/C2781238097","wikidata":"https://www.wikidata.org/wiki/Q175026","display_name":"Object (grammar)","level":2,"score":0.2515000104904175},{"id":"https://openalex.org/C17511633","wikidata":"https://www.wikidata.org/wiki/Q830694","display_name":"SMT placement equipment","level":3,"score":0.2500999867916107}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.06009","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.06009","pdf_url":"https://arxiv.org/pdf/2609.06009","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.06009","doi":"https://doi.org/10.48550/arxiv.2609.06009","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.06009","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.06009","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.06009","pdf_url":"https://arxiv.org/pdf/2609.06009","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7164367896","https://openalex.org/W2949760349","https://openalex.org/W7117987030","https://openalex.org/W7213305879","https://openalex.org/W4285102237","https://openalex.org/W4366460430","https://openalex.org/W4412376632","https://openalex.org/W7203711345","https://openalex.org/W3210326730","https://openalex.org/W2990747716"],"abstract_inverted_index":{"Multi-fingered":[0],"dexterous":[1,225],"manipulation":[2],"remains":[3],"a":[4,68,75,108,120,152,181,218],"frontier":[5],"for":[6,87,222],"real-world":[7],"reinforcement":[8],"learning":[9],"(RL)":[10],"due":[11],"to":[12,30,54,115,179,198],"the":[13,18,45,51,117,135,142,177,192,206,228],"high-dimensional":[14],"action":[15],"space":[16],"and":[17,107,160,166,211,235],"prohibitive":[19],"cost":[20],"of":[21,119],"hardware":[22],"failures.":[23],"While":[24],"human-in-the-loop":[25],"(HIL)":[26],"RL":[27,77],"allows":[28],"operators":[29],"intervene":[31],"before":[32],"failures":[33],"occur,":[34],"current":[35],"pipelines":[36],"often":[37],"treat":[38],"these":[39],"interventions":[40,83],"as":[41],"reactive":[42],"corrections,":[43],"discarding":[44],"rich":[46],"safety":[47,145],"signal":[48],"inherent":[49],"in":[50,201,227],"operator's":[52,143],"decision":[53],"take":[55],"control.":[56],"In":[57],"this":[58,215],"paper,":[59],"we":[60,64],"ask:":[61],"How":[62],"can":[63],"learn":[65],"from":[66,137],"what":[67],"human":[69,82,121,209],"would":[70],"avoid?":[71],"We":[72,147],"present":[73],"WHIRL,":[74],"safety-aware":[76,220],"framework":[78,150],"that":[79,113,133,173],"transforms":[80],"binary":[81],"into":[84],"forward-predictive":[85],"signals":[86],"proactive":[88],"risk":[89],"avoidance.":[90],"Our":[91,170],"approach":[92],"centers":[93],"on":[94,151,186],"an":[95,129],"intervention-aware":[96],"latent":[97],"world":[98,213,230],"model":[99],"with":[100],"four":[101],"prediction":[102],"heads:":[103],"dynamics,":[104],"reward,":[105],"termination,":[106],"novel":[109],"per-state":[110],"intervention-probability":[111],"head":[112,127],"learns":[114],"predict":[116],"likelihood":[118],"takeover":[122],"at":[123],"future":[124],"states.":[125],"This":[126],"provides":[128,217],"actor-side":[130],"risk-shaping":[131,175],"term":[132],"discourages":[134],"policy":[136],"entering":[138],"\"intervention-prone\"":[139],"regions,":[140],"modeling":[141],"internal":[144],"threshold.":[146],"evaluate":[148],"our":[149],"16-DoF":[153],"LEAP":[154],"Hand":[155],"across":[156],"tasks":[157,189],"spanning":[158],"convex":[159],"irregular":[161],"object":[162],"grasping,":[163],"prismatic":[164],"manipulation,":[165],"long-horizon":[167],"multi-stage":[168],"tasks.":[169],"results":[171],"show":[172],"predictive":[174,212],"enables":[176],"system":[178],"achieve":[180],"96.7":[182],"percent":[183,200],"success":[184],"rate":[185],"complex":[187,224],"grasping":[188],"while":[190,231],"reducing":[191,232],"operator":[193,233],"intervention":[194],"burden":[195],"by":[196],"up":[197],"84":[199],"step-weighted":[202],"terms.":[203],"By":[204],"closing":[205],"loop":[207],"between":[208],"intuition":[210],"modeling,":[214],"work":[216],"practical":[219],"recipe":[221],"training":[223],"agents":[226],"real":[229],"fatigue":[234],"hardware-risk":[236],"exposure.":[237]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-09-10T00:00:00"}