{"id":"https://openalex.org/W7160876754","doi":"https://doi.org/10.48550/arxiv.2605.07002","title":"Adaptive auditing of AI systems with anytime-valid guarantees","display_name":"Adaptive auditing of AI systems with anytime-valid guarantees","publication_year":2026,"publication_date":"2026-05-07","ids":{"openalex":"https://openalex.org/W7160876754","doi":"https://doi.org/10.48550/arxiv.2605.07002"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2605.07002","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2605.07002","pdf_url":"https://arxiv.org/pdf/2605.07002","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","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2605.07002","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5135876995","display_name":"Siyu Zhou","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I180670191","display_name":"University of California, San Francisco","ror":"https://ror.org/043mz5j54","country_code":"US","type":"education","lineage":["https://openalex.org/I180670191"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zhou, Siyu","raw_affiliation_strings":["University of California , San Francisco"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of California , San Francisco","institution_ids":["https://openalex.org/I180670191"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123145892","display_name":"Patrick Vossler","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I180670191","display_name":"University of California, San Francisco","ror":"https://ror.org/043mz5j54","country_code":"US","type":"education","lineage":["https://openalex.org/I180670191"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Vossler, Patrick","raw_affiliation_strings":["University of California , San Francisco"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of California , San Francisco","institution_ids":["https://openalex.org/I180670191"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5010866431","display_name":"Venkatesh Sivaraman","orcid":"https://orcid.org/0000-0002-6965-3961","observed_orcids":["https://orcid.org/0000-0002-6965-3961"]},"institutions":[{"id":"https://openalex.org/I180670191","display_name":"University of California, San Francisco","ror":"https://ror.org/043mz5j54","country_code":"US","type":"education","lineage":["https://openalex.org/I180670191"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Sivaraman, Venkatesh","raw_affiliation_strings":["University of California , San Francisco"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of California , San Francisco","institution_ids":["https://openalex.org/I180670191"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006623234","display_name":"Yifan Mai","orcid":"https://orcid.org/0000-0002-4807-037X","observed_orcids":["https://orcid.org/0000-0002-4807-037X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mai, Yifan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5051492770","display_name":"Jean Feng","orcid":"https://orcid.org/0000-0003-2041-3104","observed_orcids":["https://orcid.org/0000-0003-2041-3104"]},"institutions":[{"id":"https://openalex.org/I180670191","display_name":"University of California, San Francisco","ror":"https://ror.org/043mz5j54","country_code":"US","type":"education","lineage":["https://openalex.org/I180670191"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Feng, Jean","raw_affiliation_strings":["University of California , San Francisco"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of California , San Francisco","institution_ids":["https://openalex.org/I180670191"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":3,"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/T10743","display_name":"Software Testing and Debugging Techniques","score":0.18529999256134033,"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/T10743","display_name":"Software Testing and Debugging Techniques","score":0.18529999256134033,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.10270000249147415,"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.06870000064373016,"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/anytime-valid-inference","display_name":"anytime-valid inference","score":0.9580000042915344},{"id":"https://openalex.org/keywords/e-processes","display_name":"e-processes","score":0.781000018119812},{"id":"https://openalex.org/keywords/hypothesis-testing","display_name":"hypothesis testing","score":0.6460000276565552},{"id":"https://openalex.org/keywords/type-i-error-control","display_name":"type I error control","score":0.7429999709129333}],"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/C191988596","wikidata":"https://www.wikidata.org/wiki/Q628374","display_name":"Null hypothesis","level":2,"score":0.5956000089645386},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5583999752998352},{"id":"https://openalex.org/C2780598303","wikidata":"https://www.wikidata.org/wiki/Q65921492","display_name":"Flexibility (engineering)","level":2,"score":0.5580999851226807},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.539900004863739},{"id":"https://openalex.org/C87007009","wikidata":"https://www.wikidata.org/wiki/Q210832","display_name":"Statistical hypothesis testing","level":2,"score":0.5101000070571899},{"id":"https://openalex.org/C199521495","wikidata":"https://www.wikidata.org/wiki/Q181487","display_name":"Audit","level":2,"score":0.48969998955726624},{"id":"https://openalex.org/C203763787","wikidata":"https://www.wikidata.org/wiki/Q371029","display_name":"Null (SQL)","level":2,"score":0.4819999933242798},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.38179999589920044},{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.3702000081539154},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3528999984264374},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.3222000002861023},{"id":"https://openalex.org/C134261354","wikidata":"https://www.wikidata.org/wiki/Q938438","display_name":"Statistical inference","level":2,"score":0.32199999690055847},{"id":"https://openalex.org/C183905921","wikidata":"https://www.wikidata.org/wiki/Q1038757","display_name":"Multiple comparisons problem","level":2,"score":0.3199999928474426},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.3149000108242035},{"id":"https://openalex.org/C120577422","wikidata":"https://www.wikidata.org/wiki/Q843558","display_name":"Alternative hypothesis","level":3,"score":0.29319998621940613},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.2838999927043915},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2678000032901764},{"id":"https://openalex.org/C144352353","wikidata":"https://www.wikidata.org/wiki/Q2920411","display_name":"Computerized adaptive testing","level":3,"score":0.2648000121116638},{"id":"https://openalex.org/C80478641","wikidata":"https://www.wikidata.org/wiki/Q195771","display_name":"Sequential analysis","level":2,"score":0.2563000023365021}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2605.07002","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2605.07002","pdf_url":"https://arxiv.org/pdf/2605.07002","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.2605.07002","doi":"https://doi.org/10.48550/arxiv.2605.07002","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.07002","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":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2605.07002","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2605.07002","pdf_url":"https://arxiv.org/pdf/2605.07002","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":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7160876754.pdf","grobid_xml":"https://content.openalex.org/works/W7160876754.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7223293283","https://openalex.org/W7203632227","https://openalex.org/W4400583532","https://openalex.org/W7212255617","https://openalex.org/W7164991709","https://openalex.org/W7157323413","https://openalex.org/W7223477483","https://openalex.org/W7221614561","https://openalex.org/W7107954976","https://openalex.org/W4416158014"],"abstract_inverted_index":{"A":[0],"major":[1],"bottleneck":[2],"in":[3,85,198],"characterizing":[4],"the":[5,13,65,86,120,138,161,175,181,206],"failure":[6,128,152],"modes":[7],"of":[8,17,67,88,195,201],"generative":[9],"AI":[10,207],"systems":[11],"is":[12,46,69,126,183],"cost":[14],"and":[15,19,34,77,81,230],"time":[16],"annotation":[18],"evaluation.":[20],"Consequently,":[21],"adaptive":[22,107],"testing":[23,113,174,219,228],"paradigms":[24],"have":[25,144],"gained":[26],"popularity,":[27],"where":[28],"one":[29],"opportunistically":[30],"decides":[31],"which":[32,168],"cases":[33],"how":[35],"many":[36],"to":[37,55,74],"annotate":[38],"based":[39,93],"on":[40],"past":[41],"results.":[42],"While":[43],"this":[44],"framework":[45,114],"highly":[47,106],"practical,":[48],"its":[49],"extreme":[50],"flexibility":[51],"makes":[52],"it":[53,61],"difficult":[54],"draw":[56],"statistically":[57,233],"rigorous":[58,234],"conclusions,":[59],"as":[60,163,209,238,240],"violates":[62],"classical":[63],"assumptions:":[64],"number":[66],"observations":[68],"typically":[70],"limited":[71],"(often":[72],"10":[73],"50":[75],"cases)":[76],"decisions":[78],"regarding":[79],"sampling":[80,146],"stopping":[82],"are":[83,192],"made":[84],"midst":[87],"data":[89],"collection":[90],"rather":[91],"than":[92],"a":[94,111,133,145,151],"pre-specified":[95,227],"rule.":[96],"To":[97],"characterize":[98],"what":[99],"statistical":[100],"inferences":[101],"can":[102,231],"be":[103],"drawn":[104],"from":[105,115],"audits,":[108],"we":[109,159,186,214],"introduce":[110],"hypothesis":[112],"two":[116,190],"\"dueling\"":[117],"perspectives:":[118],"(i)":[119],"model's":[121],"null":[122,140,177],"that":[123,141,148,188,199,216],"asserts":[124,142],"there":[125],"no":[127],"mode":[129],"with":[130,237],"performance":[131],"below":[132],"target":[134],"threshold":[135],"versus":[136],"(ii)":[137],"auditor's":[139],"they":[143],"strategy":[147],"will":[149],"uncover":[150],"mode.":[153],"Leveraging":[154],"Safe":[155],"Anytime-Valid":[156],"Inference":[157],"(SAVI),":[158],"formalize":[160],"auditor":[162,182],"conducting":[164],"\"testing":[165],"by":[166],"betting,\"":[167],"translates":[169],"into":[170],"simultaneous":[171],"e-processes":[172],"for":[173],"dueling":[176],"hypotheses.":[178],"Furthermore,":[179],"if":[180],"sufficiently":[184],"powerful,":[185],"prove":[187],"these":[189],"hypotheses":[191],"asymptotically":[193],"inverses":[194],"each":[196],"other,":[197],"passage":[200],"such":[202],"an":[203],"audit":[204],"certifies":[205],"system":[208],"being":[210],"globally":[211],"robust.":[212],"Empirically,":[213],"demonstrate":[215],"our":[217],"proposed":[218],"procedures":[220],"maintain":[221],"anytime-valid":[222],"Type-I":[223],"error":[224],"control,":[225],"outperform":[226],"methods,":[229],"reach":[232],"conclusions":[235],"sometimes":[236],"few":[239],"20":[241],"observations.":[242]},"counts_by_year":[{"year":2026,"cited_by_count":3}],"updated_date":"2026-11-08T12:12:28.789561","created_date":"2026-05-12T00:00:00"}