{"id":"https://openalex.org/W7152505967","doi":"https://doi.org/10.48550/arxiv.2604.06240","title":"The Art of Building Verifiers for Computer Use Agents","display_name":"The Art of Building Verifiers for Computer Use Agents","publication_year":2026,"publication_date":"2026-04-05","ids":{"openalex":"https://openalex.org/W7152505967","doi":"https://doi.org/10.48550/arxiv.2604.06240"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.06240","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.06240","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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.2604.06240","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5026219235","display_name":"Corby Rosset","orcid":"https://orcid.org/0000-0001-9167-6214"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rosset, Corby","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133280787","display_name":"Pratyusha Sharma","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sharma, Pratyusha","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133309934","display_name":"Andrew Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Andrew","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133303290","display_name":"Miguel Gonzalez-Fernandez","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gonzalez-Fernandez, Miguel","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5021000040","display_name":"Ahmed Hassan Awadallah","orcid":"https://orcid.org/0000-0001-6426-3537"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Awadallah, Ahmed","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/T13197","display_name":"Spreadsheets and End-User Computing","score":0.11180000007152557,"subfield":{"id":"https://openalex.org/subfields/1712","display_name":"Software"},"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/T13197","display_name":"Spreadsheets and End-User Computing","score":0.11180000007152557,"subfield":{"id":"https://openalex.org/subfields/1712","display_name":"Software"},"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":0.059700001031160355,"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/T12607","display_name":"Personal Information Management and User Behavior","score":0.046799998730421066,"subfield":{"id":"https://openalex.org/subfields/1802","display_name":"Information Systems and Management"},"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/process","display_name":"Process (computing)","score":0.5548999905586243},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.5461999773979187},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.5457000136375427},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.5436000227928162},{"id":"https://openalex.org/keywords/outcome","display_name":"Outcome (game theory)","score":0.5432999730110168},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.4903999865055084},{"id":"https://openalex.org/keywords/replicate","display_name":"Replicate","score":0.4620000123977661}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6933000087738037},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.5548999905586243},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5461999773979187},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.5457000136375427},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.5436000227928162},{"id":"https://openalex.org/C148220186","wikidata":"https://www.wikidata.org/wiki/Q7111912","display_name":"Outcome (game theory)","level":2,"score":0.5432999730110168},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.4903999865055084},{"id":"https://openalex.org/C2781162219","wikidata":"https://www.wikidata.org/wiki/Q26250693","display_name":"Replicate","level":2,"score":0.4620000123977661},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4578000009059906},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.42899999022483826},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.41519999504089355},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3869999945163727},{"id":"https://openalex.org/C62230096","wikidata":"https://www.wikidata.org/wiki/Q275969","display_name":"Crowdsourcing","level":2,"score":0.3513000011444092},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.35109999775886536},{"id":"https://openalex.org/C111640148","wikidata":"https://www.wikidata.org/wiki/Q847349","display_name":"Rubric","level":2,"score":0.3416000008583069},{"id":"https://openalex.org/C2776502983","wikidata":"https://www.wikidata.org/wiki/Q690182","display_name":"Contrast (vision)","level":2,"score":0.31869998574256897},{"id":"https://openalex.org/C103278499","wikidata":"https://www.wikidata.org/wiki/Q254465","display_name":"Similarity (geometry)","level":3,"score":0.2973000109195709}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.06240","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.06240","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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.2604.06240","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.06240","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Verifying":[0],"the":[1,41,46,79,192,196,215,226],"success":[2],"of":[3,137,195,209,214],"computer":[4],"use":[5],"agent":[6,77,206],"(CUA)":[7],"trajectories":[8,139],"is":[9],"a":[10,33,100,109,120,134,165],"critical":[11],"challenge:":[12],"without":[13],"reliable":[14],"verification,":[15],"neither":[16],"evaluation":[17],"nor":[18],"training":[19],"signal":[20],"can":[21],"be":[22],"trusted.":[23],"In":[24],"this":[25],"paper,":[26],"we":[27,39],"present":[28],"lessons":[29],"learned":[30],"from":[31,191],"building":[32],"best-in-class":[34],"verifier":[35],"for":[36,103],"web":[37],"tasks":[38],"call":[40],"Universal":[42,47,150,227,232],"Verifier.":[43,228],"We":[44,128,163,185,200,229],"design":[45,197],"Verifier":[48,151,233],"around":[49],"four":[50],"key":[51],"principles:":[52],"1)":[53],"constructing":[54],"rubrics":[55],"with":[56,140,153,160,236],"meaningful,":[57],"non-overlapping":[58],"criteria":[59],"to":[60,116,171,175,219,224],"reduce":[61],"noise;":[62],"2)":[63],"separating":[64],"process":[65,142],"and":[66,95,107,143,181],"outcome":[67,144],"rewards":[68],"that":[69,114,148,187,203],"yield":[70],"complementary":[71],"signals,":[72],"capturing":[73],"cases":[74],"where":[75],"an":[76,88,204],"follows":[78],"right":[80],"steps":[81],"but":[82,217],"gets":[83],"blocked":[84],"or":[85],"succeeds":[86],"through":[87],"unexpected":[89],"path;":[90],"3)":[91],"distinguishing":[92],"between":[93],"controllable":[94],"uncontrollable":[96],"failures":[97],"scored":[98],"via":[99],"cascading-error-free":[101],"strategy":[102],"finer-grained":[104],"failure":[105],"understanding;":[106],"4)":[108],"divide-and-conquer":[110],"context":[111],"management":[112],"scheme":[113],"attends":[115],"all":[117,221],"screenshots":[118],"in":[119,167,212],"trajectory,":[121],"improving":[122],"reliability":[123],"on":[124,132],"longer":[125],"task":[126],"horizons.":[127],"validate":[129],"these":[130,188],"findings":[131],"CUAVerifierBench,":[133],"new":[135],"set":[136],"CUA":[138],"both":[141],"human":[145],"labels,":[146],"showing":[147],"our":[149,231],"agrees":[152],"humans":[154,158],"as":[155,157],"often":[156],"agree":[159],"each":[161],"other.":[162],"report":[164],"reduction":[166],"false":[168],"positive":[169],"rates":[170],"near":[172],"zero":[173],"compared":[174],"baselines":[176],"like":[177],"WebVoyager":[178],"($\\geq$":[179,183],"45\\%)":[180],"WebJudge":[182],"22\\%).":[184],"emphasize":[186],"gains":[189],"stem":[190],"cumulative":[193],"effect":[194],"choices":[198],"above.":[199],"also":[201],"find":[202],"auto-research":[205],"achieves":[207],"70\\%":[208],"expert":[210],"quality":[211],"5\\%":[213],"time,":[216],"fails":[218],"discover":[220],"strategies":[222],"required":[223],"replicate":[225],"open-source":[230],"system":[234],"along":[235],"CUAVerifierBench;":[237],"available":[238],"at":[239],"https://github.com/microsoft/fara.":[240]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-10T00:00:00"}