{"id":"https://openalex.org/W7166810388","doi":"https://doi.org/10.18653/v1/2026.acl-long.21","title":"WebSTAR: Scalable Data Synthesis for Computer Use Agents with Step-Level Filtering","display_name":"WebSTAR: Scalable Data Synthesis for Computer Use Agents with Step-Level Filtering","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166810388","doi":"https://doi.org/10.18653/v1/2026.acl-long.21"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.acl-long.21","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.21","pdf_url":"https://aclanthology.org/2026.acl-long.21.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.acl-long.21.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139764033","display_name":"Yifei He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yifei He","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051720557","display_name":"Pranit Chawla","orcid":"https://orcid.org/0000-0001-8058-340X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pranit Chawla","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139832972","display_name":"Yaser Souri","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yaser Souri","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5020732660","display_name":"Subhojit Som","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Subhojit Som","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139815369","display_name":"Xia Song","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xia Song","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":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.85991506,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"516","last_page":"533"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10444","display_name":"Context-Aware Activity Recognition Systems","score":0.06129999831318855,"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/T10444","display_name":"Context-Aware Activity Recognition Systems","score":0.06129999831318855,"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/T12016","display_name":"Web Data Mining and Analysis","score":0.05180000141263008,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T12031","display_name":"Speech and dialogue systems","score":0.038600001484155655,"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/scalability","display_name":"Scalability","score":0.45719999074935913},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.3025999963283539},{"id":"https://openalex.org/keywords/filter","display_name":"Filter (signal processing)","score":0.3001999855041504},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.29919999837875366},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.29330000281333923}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6628999710083008},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.45719999074935913},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3181999921798706},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.3025999963283539},{"id":"https://openalex.org/C106131492","wikidata":"https://www.wikidata.org/wiki/Q3072260","display_name":"Filter (signal processing)","level":2,"score":0.3001999855041504},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.29919999837875366},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2946999967098236},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.29330000281333923},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.27869999408721924},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.25839999318122864}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.acl-long.21","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.21","pdf_url":"https://aclanthology.org/2026.acl-long.21.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.acl-long.21","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.21","pdf_url":"https://aclanthology.org/2026.acl-long.21.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166810388.pdf","grobid_xml":"https://content.openalex.org/works/W7166810388.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Computer":[0],"use":[1],"agents":[2],"(CUAs)":[3],"can":[4],"operate":[5],"realworld":[6],"digital":[7],"interfaces":[8],"but":[9],"remain":[10],"difficult":[11],"to":[12,15,42,99,190,225],"train":[13,131,169],"due":[14],"the":[16,26,64],"high":[17],"cost":[18],"of":[19,28,63,118,164],"graphical":[20],"user":[21],"interface":[22],"(GUI)":[23],"interaction":[24],"and":[25,121,134,168,206,228],"scarcity":[27],"high-quality":[29],"trajectory":[30],"data.Existing":[31],"datasets":[32,211],"rely":[33],"on":[34,136,155],"human":[35,88],"demonstrations,":[36],"limiting":[37],"scalability.A":[38],"natural":[39],"alternative":[40],"is":[41,92],"synthesize":[43],"data":[44,77],"from":[45,127,178],"strong":[46],"CUAs,":[47],"yet":[48],"their":[49],"rollouts":[50,83],"are":[51],"highly":[52],"noisy,":[53],"with":[54,151,215],"incorrect":[55],"or":[56],"suboptimal":[57],"actions":[58,97],"consisting":[59],"a":[60,75,116,162,171,199,216],"large":[61],"proportion":[62],"steps,":[65,103],"making":[66],"naive":[67],"imitation":[68],"ineffective.To":[69],"tackle":[70],"this":[71,111],"challenge,":[72],"we":[73,113,158,207],"introduce":[74,208],"scalable":[76,203],"synthesis":[78],"pipeline":[79],"that":[80],"transforms":[81],"noisy":[82],"into":[84],"reliable":[85],"supervision":[86],"without":[87],"annotation.The":[89],"core":[90],"idea":[91],"step-level":[93,156,166,196],"filtering,":[94],"which":[95,180],"evaluates":[96],"individually":[98],"retain":[100],"only":[101,152],"correct":[102],"complemented":[104],"by":[105],"reasoning":[106],"augmentation":[107],"for":[108,202],"improved":[109],"planning.Using":[110],"pipeline,":[112],"construct":[114],"WebSTAR,":[115],"dataset":[117,163],"13.3K":[119],"trajectories":[120],"267K":[122],"graded,":[123],"reasoning-rich":[124],"steps":[125],"synthesized":[126],"Ope-nAI's":[128],"computer-use-preview":[129],"model.We":[130],"Qwen-2.5-VL-Instructmodels":[132],"(7B":[133],"32B)":[135],"WebSTAR.On":[137],"WebVoyager,":[138],"our":[139],"7B":[140,172],"model":[141,146,176,220],"surpasses":[142],"SoTA":[143],"open-source":[144],"CUA":[145,204],"UI-TARS-1.5-7Bby":[147],"more":[148,188],"than":[149],"15%":[150],"supervised":[153],"finetuning.Building":[154],"grading,":[157],"further":[159],"create":[160],"WebSCORE,":[161],"graded":[165],"actions,":[167],"StepRM,":[170],"multimodal":[173],"process":[174,218],"reward":[175,219],"distilled":[177],"o4-mini,":[179],"matches":[181],"its":[182],"grading":[183],"quality":[184],"while":[185],"being":[186],"far":[187],"efficient":[189,229],"deploy":[191],"at":[192],"scale.Our":[193],"results":[194],"establish":[195],"filtering":[197],"as":[198,222],"key":[200],"principle":[201],"training,":[205],"two":[209],"new":[210],"(WebSTAR,":[212],"WebSCORE)":[213],"along":[214],"lightweight":[217],"(StepRM)":[221],"practical":[223],"tools":[224],"advance":[226],"robust":[227],"CUAs.":[230]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}