{"id":"https://openalex.org/W7168271527","doi":"https://doi.org/10.48550/arxiv.2607.09773","title":"EvoCUA-1.5: Online Reinforcement Learning for Multi-turn Computer-Use Agents","display_name":"EvoCUA-1.5: Online Reinforcement Learning for Multi-turn Computer-Use Agents","publication_year":2026,"publication_date":"2026-07-07","ids":{"openalex":"https://openalex.org/W7168271527","doi":"https://doi.org/10.48550/arxiv.2607.09773"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2607.09773","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.09773","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.2607.09773","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5140586334","display_name":"Mianqiu Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Mianqiu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140648115","display_name":"Taofeng Xue","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xue, Taofeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140737880","display_name":"Chong Peng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Peng, Chong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140656241","display_name":"Jinrui Ding","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ding, Jinrui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140634898","display_name":"Sicheng Fan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fan, Sicheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140744827","display_name":"Jiale Hong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hong, Jiale","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140728826","display_name":"Yufei Gao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gao, Yufei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140750508","display_name":"Xiaocheng Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Xiaocheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140725355","display_name":"Linsen Guo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Linsen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140648855","display_name":"Xin Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Xin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140616709","display_name":"Dengchang Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Dengchang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140656664","display_name":"Yuchen Xie","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xie, Yuchen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140640538","display_name":"Peng Pei","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pei, Peng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5140657649","display_name":"Xunliang Xie","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xie, Xunliang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5140666250","display_name":"Xipeng Qiu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qiu, Xipeng","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":0.5054000020027161,"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":0.5054000020027161,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.05460000038146973,"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":0.03869999945163727,"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/reinforcement-learning","display_name":"Reinforcement learning","score":0.7767000198364258},{"id":"https://openalex.org/keywords/reuse","display_name":"Reuse","score":0.5612999796867371},{"id":"https://openalex.org/keywords/asynchronous-communication","display_name":"Asynchronous communication","score":0.5580000281333923},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.5156000256538391},{"id":"https://openalex.org/keywords/executable","display_name":"Executable","score":0.5049999952316284},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.4528000056743622},{"id":"https://openalex.org/keywords/sandbox","display_name":"Sandbox (software development)","score":0.4440999925136566},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.42910000681877136},{"id":"https://openalex.org/keywords/trajectory","display_name":"Trajectory","score":0.3905999958515167},{"id":"https://openalex.org/keywords/robotics","display_name":"Robotics","score":0.3874000012874603}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8159000277519226},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7767000198364258},{"id":"https://openalex.org/C206588197","wikidata":"https://www.wikidata.org/wiki/Q846574","display_name":"Reuse","level":2,"score":0.5612999796867371},{"id":"https://openalex.org/C151319957","wikidata":"https://www.wikidata.org/wiki/Q752739","display_name":"Asynchronous communication","level":2,"score":0.5580000281333923},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.5156000256538391},{"id":"https://openalex.org/C160145156","wikidata":"https://www.wikidata.org/wiki/Q778586","display_name":"Executable","level":2,"score":0.5049999952316284},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5038999915122986},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.4528000056743622},{"id":"https://openalex.org/C167981075","wikidata":"https://www.wikidata.org/wiki/Q2667186","display_name":"Sandbox (software development)","level":2,"score":0.4440999925136566},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.42910000681877136},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4018999934196472},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.3905999958515167},{"id":"https://openalex.org/C34413123","wikidata":"https://www.wikidata.org/wiki/Q170978","display_name":"Robotics","level":3,"score":0.3874000012874603},{"id":"https://openalex.org/C2780490138","wikidata":"https://www.wikidata.org/wiki/Q7079636","display_name":"Offline learning","level":3,"score":0.36410000920295715},{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.3617999851703644},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.3617999851703644},{"id":"https://openalex.org/C2986087404","wikidata":"https://www.wikidata.org/wiki/Q15946010","display_name":"Online learning","level":2,"score":0.3506999909877777},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.34869998693466187},{"id":"https://openalex.org/C203479927","wikidata":"https://www.wikidata.org/wiki/Q5165939","display_name":"Controller (irrigation)","level":2,"score":0.34790000319480896},{"id":"https://openalex.org/C774472","wikidata":"https://www.wikidata.org/wiki/Q6760393","display_name":"Margin (machine learning)","level":2,"score":0.3391999900341034},{"id":"https://openalex.org/C2779038628","wikidata":"https://www.wikidata.org/wiki/Q7248497","display_name":"Programming by demonstration","level":3,"score":0.3278999924659729},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.3273000121116638},{"id":"https://openalex.org/C138673069","wikidata":"https://www.wikidata.org/wiki/Q322229","display_name":"Tracing","level":2,"score":0.3215000033378601},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.30970001220703125},{"id":"https://openalex.org/C196921405","wikidata":"https://www.wikidata.org/wiki/Q786431","display_name":"Online algorithm","level":2,"score":0.30480000376701355},{"id":"https://openalex.org/C2779182362","wikidata":"https://www.wikidata.org/wiki/Q17126187","display_name":"Session (web analytics)","level":2,"score":0.3012000024318695},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.2854999899864197},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.275299996137619},{"id":"https://openalex.org/C124681953","wikidata":"https://www.wikidata.org/wiki/Q339062","display_name":"Decomposition","level":2,"score":0.27410000562667847},{"id":"https://openalex.org/C90509273","wikidata":"https://www.wikidata.org/wiki/Q11012","display_name":"Robot","level":2,"score":0.27090001106262207},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.260699987411499},{"id":"https://openalex.org/C2777851325","wikidata":"https://www.wikidata.org/wiki/Q7094102","display_name":"Online model","level":2,"score":0.2581000030040741}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2607.09773","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.09773","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.2607.09773","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.09773","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":[{"display_name":"Industry, innovation and infrastructure","score":0.619147539138794,"id":"https://metadata.un.org/sdg/9"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Computer-use":[0],"agents":[1,53],"must":[2],"solve":[3],"long-horizon":[4],"tasks":[5],"through":[6],"repeated":[7],"interaction":[8,89],"with":[9,65,106,151,183],"partially":[10],"observable,":[11],"multimodal":[12],"desktop":[13],"environments.":[14],"Although":[15],"imitation":[16],"learning":[17,57],"and":[18,46,69,98,123,141,145,154,165,179],"offline":[19,55],"trajectory":[20],"refinement":[21],"provide":[22],"strong":[23],"priors,":[24],"static":[25],"traces":[26],"cannot":[27],"cover":[28],"the":[29,40],"causal":[30],"feedback":[31],"loop":[32],"of":[33],"real":[34],"computer":[35],"use:":[36],"each":[37],"action":[38,44],"changes":[39],"screen":[41],"state,":[42],"future":[43],"space,":[45],"recovery":[47],"options.":[48],"EvoCUA-1.5":[49,102,168,189],"extends":[50],"self-evolving":[51],"computer-use":[52,200],"from":[54,71],"experience":[56],"to":[58],"online":[59,196],"reinforcement":[60],"learning,":[61],"where":[62],"policies":[63],"interact":[64],"executable":[66],"sandbox":[67],"environments":[68],"improve":[70,162],"verifiable":[72,127],"task":[73],"outcomes.":[74],"Online":[75],"RL":[76,149,197],"in":[77,198],"this":[78],"setting":[79],"requires":[80],"more":[81],"than":[82],"directly":[83],"reusing":[84],"single-turn":[85],"language-RL":[86],"recipes.":[87],"Multi-turn":[88],"introduces":[90],"context-managed":[91],"observations,":[92],"sparse":[93],"terminal":[94],"rewards,":[95],"variable-length":[96],"trajectories,":[97],"slow":[99],"environment":[100],"feedback.":[101],"addresses":[103],"these":[104,160],"challenges":[105],"Step-Level":[107],"Policy":[108],"Optimization":[109],"(STEPO),":[110],"which":[111,134],"preserves":[112],"trajectory-level":[113],"advantage":[114],"balance":[115],"after":[116],"decomposition":[117],"into":[118],"step-level":[119],"samples;":[120],"policy-aware":[121],"filtering":[122],"pass-rate":[124],"calibration":[125],"over":[126],"synthesized":[128],"tasks;":[129],"Dynamic":[130],"Tri-Adaptive":[131],"Curriculum":[132],"(DTAC),":[133],"combines":[135],"learnable":[136],"tasks,":[137],"difficult":[138],"positive":[139],"replay,":[140],"controlled":[142],"infeasible-task":[143],"exposure;":[144],"a":[146,191],"fully":[147],"asynchronous":[148],"infrastructure":[150],"staleness":[152],"control":[153],"mini-group":[155],"batching.":[156],"Experiments":[157],"show":[158],"that":[159],"components":[161],"training":[163],"stability":[164],"downstream":[166],"performance.":[167],"achieves":[169],"63.2\\%":[170],"success":[171],"on":[172],"OSWorld-Verified,":[173],"outperforming":[174],"comparable":[175],"32B/35B-scale":[176],"open-weight":[177],"baselines":[178],"even":[180],"approaching":[181],"models":[182],"significantly":[184],"larger":[185],"parameter":[186],"counts.":[187],"Overall,":[188],"provides":[190],"practical":[192],"framework":[193],"for":[194],"scaling":[195],"multi-turn":[199],"agents.":[201]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-07-15T00:00:00"}