{"id":"https://openalex.org/W7168254859","doi":"https://doi.org/10.48550/arxiv.2607.09776","title":"HELP: Human-Efficient Large-Scale Robot Post-Training with Rollout Segmentation","display_name":"HELP: Human-Efficient Large-Scale Robot Post-Training with Rollout Segmentation","publication_year":2026,"publication_date":"2026-07-08","ids":{"openalex":"https://openalex.org/W7168254859","doi":"https://doi.org/10.48550/arxiv.2607.09776"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2607.09776","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.09776","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,"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.2607.09776","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5140628473","display_name":"Shaopeng Zhai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhai, Shaopeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140635948","display_name":"Qi Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Qi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140587059","display_name":"Tianyi Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Tianyi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140610744","display_name":"Haoran Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Haoran","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140701586","display_name":"Fuxian Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Fuxian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140704155","display_name":"Zhanhui Lin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Zhanhui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140583202","display_name":"Zijun Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Zijun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5146750932","display_name":"Weinan Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Weinan","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.7702000141143799,"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":0.7702000141143799,"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/T10653","display_name":"Robot Manipulation and Learning","score":0.08529999852180481,"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/T10709","display_name":"Social Robot Interaction and HRI","score":0.026000000536441803,"subfield":{"id":"https://openalex.org/subfields/3207","display_name":"Social Psychology"},"field":{"id":"https://openalex.org/fields/32","display_name":"Psychology"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.8241000175476074},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.6855999827384949},{"id":"https://openalex.org/keywords/robot","display_name":"Robot","score":0.6804999709129333},{"id":"https://openalex.org/keywords/throughput","display_name":"Throughput","score":0.5870000123977661},{"id":"https://openalex.org/keywords/reuse","display_name":"Reuse","score":0.4864000082015991},{"id":"https://openalex.org/keywords/human\u2013robot-interaction","display_name":"Human\u2013robot interaction","score":0.42669999599456787},{"id":"https://openalex.org/keywords/operator","display_name":"Operator (biology)","score":0.4133000075817108},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.40790000557899475},{"id":"https://openalex.org/keywords/task-analysis","display_name":"Task analysis","score":0.40619999170303345}],"concepts":[{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.8241000175476074},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.6855999827384949},{"id":"https://openalex.org/C90509273","wikidata":"https://www.wikidata.org/wiki/Q11012","display_name":"Robot","level":2,"score":0.6804999709129333},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6126000285148621},{"id":"https://openalex.org/C157764524","wikidata":"https://www.wikidata.org/wiki/Q1383412","display_name":"Throughput","level":3,"score":0.5870000123977661},{"id":"https://openalex.org/C206588197","wikidata":"https://www.wikidata.org/wiki/Q846574","display_name":"Reuse","level":2,"score":0.4864000082015991},{"id":"https://openalex.org/C79403827","wikidata":"https://www.wikidata.org/wiki/Q3988","display_name":"Real-time computing","level":1,"score":0.47620001435279846},{"id":"https://openalex.org/C145460709","wikidata":"https://www.wikidata.org/wiki/Q859951","display_name":"Human\u2013robot interaction","level":3,"score":0.42669999599456787},{"id":"https://openalex.org/C17020691","wikidata":"https://www.wikidata.org/wiki/Q139677","display_name":"Operator (biology)","level":5,"score":0.4133000075817108},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.40790000557899475},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.40619999170303345},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3919000029563904},{"id":"https://openalex.org/C2780440489","wikidata":"https://www.wikidata.org/wiki/Q5227278","display_name":"Data-driven","level":2,"score":0.37290000915527344},{"id":"https://openalex.org/C115901376","wikidata":"https://www.wikidata.org/wiki/Q184199","display_name":"Automation","level":2,"score":0.33739998936653137},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.3353999853134155},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.3190999925136566},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.313400000333786},{"id":"https://openalex.org/C134765980","wikidata":"https://www.wikidata.org/wiki/Q879126","display_name":"Bitwise operation","level":2,"score":0.2913999855518341},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.27810001373291016},{"id":"https://openalex.org/C34413123","wikidata":"https://www.wikidata.org/wiki/Q170978","display_name":"Robotics","level":3,"score":0.274399995803833},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.26649999618530273},{"id":"https://openalex.org/C89600930","wikidata":"https://www.wikidata.org/wiki/Q1423946","display_name":"Segmentation","level":2,"score":0.262800008058548},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.260699987411499},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.2590999901294708},{"id":"https://openalex.org/C74222875","wikidata":"https://www.wikidata.org/wiki/Q16000312","display_name":"Robot kinematics","level":4,"score":0.2563000023365021},{"id":"https://openalex.org/C2780626000","wikidata":"https://www.wikidata.org/wiki/Q5936775","display_name":"Human-in-the-loop","level":2,"score":0.25619998574256897},{"id":"https://openalex.org/C150415221","wikidata":"https://www.wikidata.org/wiki/Q40687","display_name":"Robotic arm","level":2,"score":0.2529999911785126},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.251800000667572}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2607.09776","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.09776","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,"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.2607.09776","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.09776","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,"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":[{"score":0.7228860855102539,"display_name":"Decent work and economic growth","id":"https://metadata.un.org/sdg/8"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"When":[0],"adapting":[1],"Vision":[2],"Language":[3],"Action":[4],"(VLA)":[5],"models":[6],"to":[7,17,133],"downstream":[8],"tasks,":[9,196],"multiple":[10],"rounds":[11],"of":[12,44,120],"post-training":[13,190],"are":[14,180],"often":[15],"required":[16],"progressively":[18],"address":[19],"policy":[20,37,121],"weaknesses.":[21],"In":[22],"this":[23,32,164],"report,":[24],"we":[25],"focus":[26],"on":[27],"maximizing":[28],"human":[29,45,95,126],"efficiency":[30,96],"during":[31],"iterative":[33],"process,":[34],"measured":[35],"by":[36,97,124,206,222,227],"improvement":[38],"and":[39,47,74,87,105,135,141,149,174,182,202,224],"task":[40,99,204],"throughput":[41,205,220],"per":[42],"unit":[43],"labor":[46],"time.":[48],"We":[49],"propose":[50],"HELP,":[51],"a":[52,78],"Human-Efficient":[53],"Large-scale":[54],"robot":[55,83,107],"Post-training":[56],"pipeline":[57],"in":[58],"which":[59],"two":[60],"specialized":[61],"operators":[62],"supervise":[63],"twelve":[64],"robots":[65],"concurrently.":[66],"A":[67],"trained":[68],"Teleoperator":[69],"provides":[70],"high-value":[71],"remote":[72],"interventions":[73],"recovery":[75,142,175,215],"demonstrations,":[76],"while":[77],"Floor":[79],"Operator":[80],"monitors":[81],"the":[82,118,125,147,188,209],"fleet,":[84],"triggers":[85],"takeovers,":[86,139],"performs":[88],"physical":[89],"resets.":[90],"This":[91],"role":[92],"specialization":[93],"improves":[94,203],"reducing":[98],"switching,":[100],"lowering":[101],"operator":[102],"training":[103],"costs,":[104],"expanding":[106],"interaction":[108],"coverage.":[109],"Beyond":[110],"increasing":[111],"rollout":[112,151,158,178],"volume,":[113],"concurrent":[114],"supervision":[115],"also":[116],"broadens":[117],"range":[119],"behaviors":[122],"observed":[123],"team,":[127],"making":[128],"recurring":[129],"failure":[130],"modes":[131],"easier":[132],"identify":[134],"enabling":[136],"more":[137],"targeted":[138],"resets,":[140],"demonstrations.":[143],"To":[144],"efficiently":[145],"utilize":[146],"large":[148],"mixed-quality":[150],"data,":[152],"HELP":[153,197],"incorporates":[154],"\\vlac,":[155],"an":[156],"automatic":[157],"segmentation":[159],"critic":[160],"specifically":[161],"designed":[162],"for":[163,187],"setting.":[165],"It":[166],"separates":[167],"autonomous":[168],"trajectories":[169],"into":[170],"progress-making,":[171],"idle,":[172],"failure-inducing,":[173],"segments.":[176],"Useful":[177],"segments":[179],"retained":[181],"combined":[183],"with":[184],"Human-in-the-Loop":[185],"data":[186],"next":[189],"round.":[191],"Across":[192],"four":[193],"real-world":[194],"manipulation":[195],"achieves":[198],"80\\%--95\\%":[199],"success":[200],"rates":[201],"1.7$\\times$--4.2$\\times$":[207],"over":[208,229],"base":[210],"model.":[211],"Under":[212],"matched":[213],"HITL":[214],"budgets,":[216],"VLAC-CUT":[217],"further":[218],"amplifies":[219],"gains":[221,226],"1.20$\\times$--3.43$\\times$":[223],"success-rate":[225],"1.50$\\times$--3.00$\\times$":[228],"HITL-only":[230],"updates.":[231]},"counts_by_year":[],"updated_date":"2026-08-12T21:12:35.861297","created_date":"2026-07-15T00:00:00"}