{"id":"https://openalex.org/W7169868577","doi":"https://doi.org/10.48550/arxiv.2609.17560","title":"Pay Only for Disagreement: Certified No-Regression Verdicts for Model Updates with Matching Label-Complexity Bounds","display_name":"Pay Only for Disagreement: Certified No-Regression Verdicts for Model Updates with Matching Label-Complexity Bounds","publication_year":2026,"publication_date":"2026-07-20","ids":{"openalex":"https://openalex.org/W7169868577","doi":"https://doi.org/10.48550/arxiv.2609.17560"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.17560","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.17560","pdf_url":"https://arxiv.org/pdf/2609.17560","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":null,"license_id":null,"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/2609.17560","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5141256250","display_name":"VISHNU BINDU BALACHANDRAN","orcid":"https://orcid.org/0000-0001-6903-6990","observed_orcids":["https://orcid.org/0000-0001-6903-6990"]},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Balachandran, Vishnu Bindu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5141256250"],"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.24940000474452972,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.24940000474452972,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.15710000693798065,"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":0.11069999635219574,"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/model-updating","display_name":"model updating","score":0.8579999804496765},{"id":"https://openalex.org/keywords/label-complexity","display_name":"label complexity","score":0.574999988079071},{"id":"https://openalex.org/keywords/confidence-sequences","display_name":"confidence sequences","score":0.9649999737739563},{"id":"https://openalex.org/keywords/anytime-valid-inference","display_name":"anytime-valid inference","score":0.8600000143051147},{"id":"https://openalex.org/keywords/low-rank-adaptation","display_name":"low-rank adaptation","score":0.8460000157356262},{"id":"https://openalex.org/keywords/language-model-fine-tuning","display_name":"language model fine-tuning","score":0.5730000138282776},{"id":"https://openalex.org/keywords/post-deployment-monitoring","display_name":"post-deployment monitoring","score":0.6959999799728394},{"id":"https://openalex.org/keywords/eu-ai-act","display_name":"EU AI Act","score":0.8980000019073486}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.692799985408783},{"id":"https://openalex.org/C199521495","wikidata":"https://www.wikidata.org/wiki/Q181487","display_name":"Audit","level":2,"score":0.647599995136261},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.6122999787330627},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.5393999814987183},{"id":"https://openalex.org/C46304622","wikidata":"https://www.wikidata.org/wiki/Q374814","display_name":"Certification","level":2,"score":0.5080999732017517},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.4474000036716461},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.38280001282691956},{"id":"https://openalex.org/C2780813799","wikidata":"https://www.wikidata.org/wiki/Q3274237","display_name":"Zero (linguistics)","level":2,"score":0.3790000081062317},{"id":"https://openalex.org/C80478641","wikidata":"https://www.wikidata.org/wiki/Q195771","display_name":"Sequential analysis","level":2,"score":0.3587999939918518},{"id":"https://openalex.org/C2777338717","wikidata":"https://www.wikidata.org/wiki/Q1762621","display_name":"Vendor","level":2,"score":0.35760000348091125},{"id":"https://openalex.org/C2778348673","wikidata":"https://www.wikidata.org/wiki/Q739302","display_name":"Production (economics)","level":2,"score":0.32899999618530273},{"id":"https://openalex.org/C98147612","wikidata":"https://www.wikidata.org/wiki/Q215599","display_name":"Promotion (chess)","level":3,"score":0.323199987411499},{"id":"https://openalex.org/C40696583","wikidata":"https://www.wikidata.org/wiki/Q989120","display_name":"Type I and type II errors","level":2,"score":0.3174999952316284},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3052999973297119},{"id":"https://openalex.org/C34388435","wikidata":"https://www.wikidata.org/wiki/Q2267362","display_name":"Bounded function","level":2,"score":0.303600013256073},{"id":"https://openalex.org/C151201525","wikidata":"https://www.wikidata.org/wiki/Q177239","display_name":"Limit (mathematics)","level":2,"score":0.2824000120162964},{"id":"https://openalex.org/C80958533","wikidata":"https://www.wikidata.org/wiki/Q1047174","display_name":"Audit trail","level":3,"score":0.2782999873161316},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.2773999869823456},{"id":"https://openalex.org/C40969351","wikidata":"https://www.wikidata.org/wiki/Q3516228","display_name":"Word error rate","level":2,"score":0.2768000066280365},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.2754000127315521},{"id":"https://openalex.org/C15569618","wikidata":"https://www.wikidata.org/wiki/Q3561421","display_name":"Liveness","level":2,"score":0.2694000005722046},{"id":"https://openalex.org/C2776639384","wikidata":"https://www.wikidata.org/wiki/Q840396","display_name":"Ideal (ethics)","level":2,"score":0.2694000005722046},{"id":"https://openalex.org/C163258240","wikidata":"https://www.wikidata.org/wiki/Q25342","display_name":"Power (physics)","level":2,"score":0.26899999380111694},{"id":"https://openalex.org/C195818886","wikidata":"https://www.wikidata.org/wiki/Q5421724","display_name":"Expressive power","level":2,"score":0.2637999951839447},{"id":"https://openalex.org/C2778067643","wikidata":"https://www.wikidata.org/wiki/Q166507","display_name":"Interval (graph theory)","level":2,"score":0.25920000672340393},{"id":"https://openalex.org/C96865113","wikidata":"https://www.wikidata.org/wiki/Q2946816","display_name":"Certificate","level":2,"score":0.2533000111579895}],"mesh":[],"locations_count":4,"locations":[{"id":"pmh:oai:arXiv.org:2609.17560","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.17560","pdf_url":"https://arxiv.org/pdf/2609.17560","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.17560","doi":"https://doi.org/10.48550/arxiv.2609.17560","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.17560","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"},{"id":"doi:10.5281/zenodo.21459379","doi":"https://doi.org/10.5281/zenodo.21459379","is_oa":true,"landing_page_url":"https://doi.org/10.5281/zenodo.21459379","pdf_url":null,"source":{"id":"https://openalex.org/S4306400562","display_name":"Zenodo (CERN European Organization for Nuclear Research)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I67311998","host_organization_name":"European Organization for Nuclear Research","host_organization_lineage":["https://openalex.org/I67311998"],"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"},{"id":"doi:10.5281/zenodo.21459380","doi":"https://doi.org/10.5281/zenodo.21459380","is_oa":true,"landing_page_url":"https://doi.org/10.5281/zenodo.21459380","pdf_url":null,"source":{"id":"https://openalex.org/S4306400562","display_name":"Zenodo (CERN European Organization for Nuclear Research)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I67311998","host_organization_name":"European Organization for Nuclear Research","host_organization_lineage":["https://openalex.org/I67311998"],"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:2609.17560","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.17560","pdf_url":"https://arxiv.org/pdf/2609.17560","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":null,"license_id":null,"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/W7169868577.pdf","grobid_xml":"https://content.openalex.org/works/W7169868577.grobid-xml"},"referenced_works_count":5,"referenced_works":["https://openalex.org/W3030671780","https://openalex.org/W3126216366","https://openalex.org/W3134428469","https://openalex.org/W4388461534","https://openalex.org/W4388520288"],"related_works":["https://openalex.org/W7214726705","https://openalex.org/W7123273274","https://openalex.org/W7116877536","https://openalex.org/W7153773467","https://openalex.org/W4416968721","https://openalex.org/W7151251863","https://openalex.org/W7214125621","https://openalex.org/W7204238461","https://openalex.org/W4417250589","https://openalex.org/W4415959394"],"abstract_inverted_index":{"Every":[0],"production":[1],"model":[2,198],"is":[3,36,72,166],"updated,":[4],"by":[5],"retraining,":[6],"fine-tuning,":[7],"quantization,":[8],"or":[9],"a":[10,37,59,125,188],"silent":[11],"vendor":[12],"swap,":[13],"and":[14,96,109,132,176],"each":[15],"update":[16,26,153],"risks":[17],"being":[18],"worse":[19],"than":[20],"what":[21],"it":[22],"replaced.":[23],"We":[24,56,105],"formalize":[25],"promotion":[27],"as":[28],"certified":[29],"paired":[30],"risk-difference":[31],"auditing.":[32],"Our":[33],"starting":[34],"point":[35],"support":[38],"identity.":[39],"The":[40],"risk":[41],"difference":[42],"between":[43],"two":[44],"models":[45,160],"lives":[46],"on":[47],"the":[48,116,133],"inputs":[49],"where":[50],"they":[51],"disagree,":[52],"observable":[53],"without":[54],"labels.":[55,184],"build":[57],"Discern,":[58],"sequential":[60],"two-tier":[61],"protocol.":[62],"A":[63],"zero-label":[64],"tier":[65,81],"certifies":[66],"benign":[67,179],"updates":[68,180],"whose":[69],"disagreement":[70,122],"rate":[71,117],"below":[73],"tolerance":[74],"from":[75,142],"unlabeled":[76],"traffic":[77],"alone.":[78],"An":[79],"audited":[80],"labels":[82],"only":[83],"sampled":[84],"disagreements":[85],"through":[86],"an":[87,102,137],"anytime-valid":[88,200],"confidence":[89,204],"sequence,":[90],"valid":[91],"at":[92,115],"every":[93],"stopping":[94],"time":[95],"under":[97],"any":[98,129],"label-routing":[99],"rule,":[100],"even":[101],"adversarial":[103],"judge.":[104],"prove":[106],"finite-sample":[107],"validity":[108],"matching":[110],"label-complexity":[111],"bounds":[112],"e\u0398":[113],"(\u03c12/\u03b52)":[114],"level,":[118],"so":[119],"exploiting":[120],"free":[121],"provably":[123],"saves":[124],"factor":[126],"\u0398(1/\u03c1)":[127],"over":[128,151],"pairing-blind":[130],"auditor,":[131],"guarantee":[134],"composes":[135],"across":[136],"unbounded":[138],"sequence":[139],"of":[140,158,178],"promotions":[141],"one":[143],"error":[144],"budget.":[145],"Across":[146],"14,000+":[147],"replayed":[148],"audit":[149,186],"streams":[150],"785":[152],"pairs,":[154],"including":[155],"LoRA":[156],"fine-tunes":[157],"language":[159],"up":[161],"to":[162],"1.4B":[163],"parameters,":[164],"miscoverage":[165],"0.0002":[167],"(nominal":[168],"5%),":[169],"power":[170],"0.986":[171],"with":[172,182],"zero":[173,183],"false":[174],"alarms,":[175],"56%":[177],"certify":[181],"Each":[185],"emits":[187],"machine-checkable":[189],"evidence":[190],"record":[191],"for":[192],"EU":[193],"AI":[194],"Act":[195],"post-market":[196],"monitoring.Keywords:":[197],"updates,":[199],"inference,":[201],"label":[202],"complexity,":[203],"sequences,":[205],"deployment":[206],"monitoring":[207]},"counts_by_year":[{"year":2026,"cited_by_count":3}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-07-22T00:00:00"}