{"id":"https://openalex.org/W7092292012","doi":"https://doi.org/10.48550/arxiv.2510.14969","title":"LLMs as Scalable, General-Purpose Simulators For Evolving Digital Agent Training","display_name":"LLMs as Scalable, General-Purpose Simulators For Evolving Digital Agent Training","publication_year":2025,"publication_date":"2025-10-16","ids":{"openalex":"https://openalex.org/W7092292012","doi":"https://doi.org/10.48550/arxiv.2510.14969"},"language":null,"primary_location":{"id":"pmh:oai:arXiv.org:2510.14969","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2510.14969","pdf_url":"https://arxiv.org/pdf/2510.14969","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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/2510.14969","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Wang, Yiming","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yiming","raw_affiliation_strings":["Harvard University \u22c6 Co-First Authors \u2660 Co-Lead"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harvard University \u22c6 Co-First Authors \u2660 Co-Lead","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Yin, Da","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yin, Da","raw_affiliation_strings":["Harvard University \u22c6 Co-First Authors \u2660 Co-Lead"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harvard University \u22c6 Co-First Authors \u2660 Co-Lead","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Cui, Yuedong","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cui, Yuedong","raw_affiliation_strings":["Harvard University \u22c6 Co-First Authors \u2660 Co-Lead"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harvard University \u22c6 Co-First Authors \u2660 Co-Lead","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Zheng, Ruichen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Ruichen","raw_affiliation_strings":["Harvard University \u22c6 Co-First Authors \u2660 Co-Lead"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harvard University \u22c6 Co-First Authors \u2660 Co-Lead","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Li, Zhiqian","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Zhiqian","raw_affiliation_strings":["Harvard University \u22c6 Co-First Authors \u2660 Co-Lead"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harvard University \u22c6 Co-First Authors \u2660 Co-Lead","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Lin, Zongyu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Zongyu","raw_affiliation_strings":["Harvard University \u22c6 Co-First Authors \u2660 Co-Lead"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harvard University \u22c6 Co-First Authors \u2660 Co-Lead","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Wu, Di","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Di","raw_affiliation_strings":["Harvard University \u22c6 Co-First Authors \u2660 Co-Lead"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harvard University \u22c6 Co-First Authors \u2660 Co-Lead","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Wu, Xueqing","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Xueqing","raw_affiliation_strings":["Harvard University \u22c6 Co-First Authors \u2660 Co-Lead"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harvard University \u22c6 Co-First Authors \u2660 Co-Lead","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Ye, Chenchen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ye, Chenchen","raw_affiliation_strings":["Harvard University \u22c6 Co-First Authors \u2660 Co-Lead"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harvard University \u22c6 Co-First Authors \u2660 Co-Lead","institution_ids":[]}]},{"author_position":"middle","author":{"id":null,"display_name":"Zhou, Yu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Yu","raw_affiliation_strings":["Harvard University \u22c6 Co-First Authors \u2660 Co-Lead"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harvard University \u22c6 Co-First Authors \u2660 Co-Lead","institution_ids":[]}]},{"author_position":"last","author":{"id":null,"display_name":"Chang, Kai-Wei","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chang, Kai-Wei","raw_affiliation_strings":["Harvard University \u22c6 Co-First Authors \u2660 Co-Lead"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Harvard University \u22c6 Co-First Authors \u2660 Co-Lead","institution_ids":[]}]}],"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":true,"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":true,"primary_topic":{"id":"https://openalex.org/T12016","display_name":"Web Data Mining and Analysis","score":0.4771000146865845,"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"}},"topics":[{"id":"https://openalex.org/T12016","display_name":"Web Data Mining and Analysis","score":0.4771000146865845,"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.07540000230073929,"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/T10906","display_name":"AI-based Problem Solving and Planning","score":0.06960000097751617,"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/trajectory-generation","display_name":"trajectory generation","score":0.6359999775886536},{"id":"https://openalex.org/keywords/webarena","display_name":"WebArena","score":0.8399999737739563},{"id":"https://openalex.org/keywords/agent-robustness","display_name":"agent robustness","score":0.5740000009536743}],"concepts":[{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.7390999794006348},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.7296000123023987},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6837000250816345},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.6816999912261963},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.48910000920295715},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.4691999852657318},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.42410001158714294},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.4058000147342682},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.3917999863624573},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.29840001463890076},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2648000121116638},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.25859999656677246},{"id":"https://openalex.org/C44154836","wikidata":"https://www.wikidata.org/wiki/Q45045","display_name":"Simulation","level":1,"score":0.25760000944137573},{"id":"https://openalex.org/C71559656","wikidata":"https://www.wikidata.org/wiki/Q671298","display_name":"Divide and conquer algorithms","level":2,"score":0.25360000133514404}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2510.14969","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2510.14969","pdf_url":"https://arxiv.org/pdf/2510.14969","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2510.14969","doi":"https://doi.org/10.48550/arxiv.2510.14969","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2510.14969","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2510.14969","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2510.14969","pdf_url":"https://arxiv.org/pdf/2510.14969","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"sustainable_development_goals_aurora":[{"id":"https://metadata.un.org/sdg/4","display_name":"Quality education","score":0.4082072675228119}],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7092292012.pdf","grobid_xml":"https://content.openalex.org/works/W7092292012.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Digital":[0],"agents":[1,116],"require":[2],"diverse,":[3],"large-scale":[4],"UI":[5,39,58],"trajectories":[6,46,76],"to":[7,43,152],"generalize":[8],"across":[9],"real-world":[10],"tasks,":[11],"yet":[12],"collecting":[13],"such":[14],"data":[15],"is":[16],"prohibitively":[17],"expensive":[18],"in":[19],"both":[20],"human":[21],"annotation,":[22],"infra":[23],"and":[24,41,67,74,92,99,107,154],"engineering":[25],"perspectives.":[26],"To":[27],"this":[28],"end,":[29],"we":[30],"introduce":[31],"$\\textbf{UI-Simulator}$,":[32],"a":[33,52,60,68,84],"scalable":[34],"paradigm":[35,50,151],"that":[36,71,88,110],"generates":[37],"structured":[38],"states":[40],"transitions":[42],"synthesize":[44],"training":[45],"at":[47],"scale.":[48],"Our":[49],"integrates":[51],"digital":[53,158],"world":[54],"simulator":[55],"for":[56,64,77],"diverse":[57,75],"states,":[59],"guided":[61],"rollout":[62],"process":[63],"coherent":[65],"exploration,":[66],"trajectory":[69,102],"wrapper":[70],"produces":[72],"high-quality":[73],"agent":[78],"training.":[79],"We":[80],"further":[81],"propose":[82],"$\\textbf{UI-Simulator-Grow}$,":[83],"targeted":[85,148],"scaling":[86,94,150],"strategy":[87],"enables":[89],"more":[90],"rapid":[91],"data-efficient":[93],"by":[95],"prioritizing":[96],"high-impact":[97],"tasks":[98],"synthesizes":[100],"informative":[101],"variants.":[103],"Experiments":[104],"on":[105,118],"WebArena":[106],"AndroidWorld":[108],"show":[109],"UI-Simulator":[111],"rivals":[112],"or":[113],"surpasses":[114],"open-source":[115],"trained":[117],"real":[119],"UIs":[120],"with":[121],"significantly":[122],"better":[123],"robustness,":[124],"despite":[125],"using":[126,137],"weaker":[127],"teacher":[128],"models.":[129],"Moreover,":[130],"UI-Simulator-Grow":[131],"matches":[132],"the":[133,141,145,157],"performance":[134],"of":[135,147],"Llama-3-70B-Instruct":[136],"only":[138],"Llama-3-8B-Instruct":[139],"as":[140],"base":[142],"model,":[143],"highlighting":[144],"potential":[146],"synthesis":[149],"continuously":[153],"efficiently":[155],"enhance":[156],"agents.":[159]},"counts_by_year":[],"updated_date":"2026-09-30T09:27:27.532063","created_date":"2025-10-18T00:00:00"}