{"id":"https://openalex.org/W7204207599","doi":"https://doi.org/10.48550/arxiv.2608.21496","title":"The geometry of AI validation: From structural blindness to reusable audits","display_name":"The geometry of AI validation: From structural blindness to reusable audits","publication_year":2026,"publication_date":"2026-08-21","ids":{"openalex":"https://openalex.org/W7204207599","doi":"https://doi.org/10.48550/arxiv.2608.21496"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.21496","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.21496","pdf_url":"https://arxiv.org/pdf/2608.21496","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/2608.21496","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5049418177","display_name":"Ricardo Fitas","orcid":"https://orcid.org/0000-0001-5137-2451","observed_orcids":["https://orcid.org/0000-0001-5137-2451"]},"institutions":[{"id":"https://openalex.org/I31512782","display_name":"Technische Universit\u00e4t Darmstadt","ror":"https://ror.org/05n911h24","country_code":"DE","type":"education","lineage":["https://openalex.org/I31512782"]}],"countries":["DE"],"is_corresponding":true,"raw_author_name":"Fitas, Ricardo","raw_affiliation_strings":["Technical University of Darmstadt , Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Technical University of Darmstadt , Germany","institution_ids":["https://openalex.org/I31512782"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":["https://openalex.org/A5049418177"],"corresponding_institution_ids":["https://openalex.org/I31512782"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":1,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.3571999967098236,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.3571999967098236,"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/T12072","display_name":"Machine Learning and Algorithms","score":0.31470000743865967,"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":0.06589999794960022,"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/ai-validation","display_name":"AI validation","score":0.8899999856948853},{"id":"https://openalex.org/keywords/syntactic-ambiguity","display_name":"syntactic ambiguity","score":0.5320000052452087},{"id":"https://openalex.org/keywords/reliability-estimation","display_name":"reliability estimation","score":0.7400000095367432},{"id":"https://openalex.org/keywords/information-gap","display_name":"information gap","score":0.3959999978542328},{"id":"https://openalex.org/keywords/minimax-lower-bounds","display_name":"minimax lower bounds","score":0.6880000233650208},{"id":"https://openalex.org/keywords/candidate-ranking","display_name":"candidate ranking","score":0.3709999918937683},{"id":"https://openalex.org/keywords/code-generation","display_name":"code generation","score":0.4830000102519989},{"id":"https://openalex.org/keywords/mathematical-reasoning","display_name":"mathematical reasoning","score":0.843999981880188},{"id":"https://openalex.org/keywords/held-out-evaluation","display_name":"held-out evaluation","score":0.39800000190734863}],"concepts":[{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.545199990272522},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5038999915122986},{"id":"https://openalex.org/C34388435","wikidata":"https://www.wikidata.org/wiki/Q2267362","display_name":"Bounded function","level":2,"score":0.5033000111579895},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.42100000381469727},{"id":"https://openalex.org/C2780522230","wikidata":"https://www.wikidata.org/wiki/Q1140419","display_name":"Ambiguity","level":2,"score":0.40709999203681946},{"id":"https://openalex.org/C22324862","wikidata":"https://www.wikidata.org/wiki/Q652707","display_name":"Lipschitz continuity","level":2,"score":0.3637000024318695},{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.3528999984264374},{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.3513999879360199},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.34220001101493835},{"id":"https://openalex.org/C72169020","wikidata":"https://www.wikidata.org/wiki/Q194404","display_name":"Monotonic function","level":2,"score":0.3402000069618225},{"id":"https://openalex.org/C79337645","wikidata":"https://www.wikidata.org/wiki/Q779824","display_name":"Outlier","level":2,"score":0.31630000472068787},{"id":"https://openalex.org/C118615104","wikidata":"https://www.wikidata.org/wiki/Q121416","display_name":"Discrete mathematics","level":1,"score":0.31369999051094055},{"id":"https://openalex.org/C122041747","wikidata":"https://www.wikidata.org/wiki/Q838611","display_name":"Ball (mathematics)","level":2,"score":0.2978000044822693},{"id":"https://openalex.org/C2834757","wikidata":"https://www.wikidata.org/wiki/Q4925424","display_name":"Monotone polygon","level":2,"score":0.2888999879360199},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.2822999954223633},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.27230000495910645},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.2639000117778778},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.26260000467300415},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2567000091075897},{"id":"https://openalex.org/C2776436953","wikidata":"https://www.wikidata.org/wiki/Q5163215","display_name":"Consistency (knowledge bases)","level":2,"score":0.2556999921798706},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.2554999887943268}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.21496","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.21496","pdf_url":"https://arxiv.org/pdf/2608.21496","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.2608.21496","doi":"https://doi.org/10.48550/arxiv.2608.21496","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.21496","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:2608.21496","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.21496","pdf_url":"https://arxiv.org/pdf/2608.21496","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/W7204207599.pdf","grobid_xml":"https://content.openalex.org/works/W7204207599.grobid-xml"},"referenced_works_count":19,"referenced_works":["https://openalex.org/W81897278","https://openalex.org/W1974231626","https://openalex.org/W2010060929","https://openalex.org/W2044728162","https://openalex.org/W2083109489","https://openalex.org/W2317610121","https://openalex.org/W2612994090","https://openalex.org/W3014740640","https://openalex.org/W4239987253","https://openalex.org/W4295022726","https://openalex.org/W4299521717","https://openalex.org/W4366783720","https://openalex.org/W4394988976","https://openalex.org/W4402733601","https://openalex.org/W4402952666","https://openalex.org/W4415142802","https://openalex.org/W4416043407","https://openalex.org/W7163209470","https://openalex.org/W7163346078"],"related_works":["https://openalex.org/W7170883127","https://openalex.org/W7171715373","https://openalex.org/W7202170911","https://openalex.org/W7160876754","https://openalex.org/W7122760613","https://openalex.org/W7223293283","https://openalex.org/W7154507511","https://openalex.org/W7096548374","https://openalex.org/W4412825577","https://openalex.org/W7223445378"],"abstract_inverted_index":{"AI":[0],"systems":[1],"increasingly":[2],"search":[3,13,66],"among":[4],"candidate":[5,71,122,137],"answers":[6],"and":[7,73,120,136,164,189],"deploy":[8],"the":[9,37,53,95,133,180],"highest-scoring":[10],"one.":[11],"Increasing":[12],"changes":[14],"which":[15],"errors":[16],"matter,":[17],"so":[18],"a":[19,77,111,150,158,176,201],"precise":[20],"evaluation":[21],"at":[22,110],"one":[23,80],"computation":[24],"budget":[25,29],"can":[26,82],"leave":[27],"another":[28],"unresolved.":[30],"We":[31],"connect":[32],"this":[33,142,148],"information":[34],"gap":[35],"to":[36,89,141,192],"cost":[38],"of":[39,116],"closing":[40],"it.":[41],"For":[42],"independent":[43],"best-of-n":[44],"search,":[45],"aggregate":[46],"reliability":[47,84],"measurements":[48],"identify":[49],"deployment":[50],"only":[51,64],"through":[52],"directions":[54],"they":[55],"observe;":[56],"we":[57],"derive":[58],"an":[59],"exact":[60],"ambiguity":[61,199],"frontier":[62],"when":[63],"small":[65],"widths":[67,87,186],"are":[68,139],"audited.":[69],"Retaining":[70],"ranks":[72],"truth":[74,118],"labels":[75,119],"enables":[76],"constructive":[78],"alternative:":[79],"audit":[81,178],"estimate":[83],"across":[85,184],"all":[86],"up":[88],"N.":[90],"With":[91],"known":[92,159],"score":[93,160],"percentiles,":[94],"minimax":[96],"worst-coordinate":[97],"mean":[98],"squared":[99],"error":[100,183],"scales":[101],"as":[102],"(1":[103],"+":[104,107],"log":[105],"N)/T":[106],"N/M,":[108],"capped":[109],"constant,":[112],"for":[113,204],"expected":[114],"budgets":[115],"T":[117],"M":[121],"observations.":[123],"Matching":[124],"lower":[125],"bounds":[126],"allow":[127],"adaptive":[128],"label":[129,135],"acquisition,":[130],"establishing":[131],"that":[132],"distinct":[134],"costs":[138],"intrinsic":[140],"experiment.":[143],"An":[144],"explicit":[145],"design":[146],"attains":[147],"order;":[149],"complementary":[151],"record-based":[152],"procedure":[153],"supplies":[154],"simultaneous":[155],"guarantees":[156],"without":[157],"distribution.":[161],"Retrospective":[162],"mathematical-reasoning":[163],"code-generation":[165],"analyses":[166],"show":[167],"why":[168],"search-dependent":[169],"validation":[170],"matters.":[171],"In":[172],"held-out":[173],"CodeRM":[174],"pools,":[175],"shared":[177],"reduces":[179],"95th-percentile":[181],"maximum":[182],"100":[185],"by":[187],"58%":[188],"40%":[190],"relative":[191],"uniform":[193],"labeling.":[194],"These":[195],"results":[196],"turn":[197],"structural":[198],"into":[200],"quantitative":[202],"prescription":[203],"reusable":[205],"validation.":[206]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-08-26T00:00:00"}