{"id":"https://openalex.org/W7139027933","doi":"https://doi.org/10.48550/arxiv.2603.17041","title":"When Marginals Match but Structure Fails: Covariance Fidelity in Generative Models","display_name":"When Marginals Match but Structure Fails: Covariance Fidelity in Generative Models","publication_year":2026,"publication_date":"2026-03-17","ids":{"openalex":"https://openalex.org/W7139027933","doi":"https://doi.org/10.48550/arxiv.2603.17041"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.17041","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.17041","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,"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.2603.17041","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5130152488","display_name":"Nazia Riasat","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Riasat, Nazia","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5130152488"],"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.6621999740600586,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.6621999740600586,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T11883","display_name":"Embodied and Extended Cognition","score":0.1062999963760376,"subfield":{"id":"https://openalex.org/subfields/2805","display_name":"Cognitive Neuroscience"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T12090","display_name":"Language and cultural evolution","score":0.02160000056028366,"subfield":{"id":"https://openalex.org/subfields/3316","display_name":"Cultural Studies"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/marginal-distribution","display_name":"Marginal distribution","score":0.6157000064849854},{"id":"https://openalex.org/keywords/univariate","display_name":"Univariate","score":0.6105999946594238},{"id":"https://openalex.org/keywords/covariance","display_name":"Covariance","score":0.570900022983551},{"id":"https://openalex.org/keywords/fidelity","display_name":"Fidelity","score":0.5440000295639038},{"id":"https://openalex.org/keywords/divergence","display_name":"Divergence (linguistics)","score":0.5430999994277954},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5303999781608582},{"id":"https://openalex.org/keywords/downstream","display_name":"Downstream (manufacturing)","score":0.5250999927520752},{"id":"https://openalex.org/keywords/generative-grammar","display_name":"Generative grammar","score":0.5220999717712402},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.5180000066757202},{"id":"https://openalex.org/keywords/generative-model","display_name":"Generative model","score":0.5069000124931335}],"concepts":[{"id":"https://openalex.org/C165216359","wikidata":"https://www.wikidata.org/wiki/Q670653","display_name":"Marginal distribution","level":3,"score":0.6157000064849854},{"id":"https://openalex.org/C199163554","wikidata":"https://www.wikidata.org/wiki/Q1681619","display_name":"Univariate","level":3,"score":0.6105999946594238},{"id":"https://openalex.org/C178650346","wikidata":"https://www.wikidata.org/wiki/Q201984","display_name":"Covariance","level":2,"score":0.570900022983551},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5473999977111816},{"id":"https://openalex.org/C2776459999","wikidata":"https://www.wikidata.org/wiki/Q2119376","display_name":"Fidelity","level":2,"score":0.5440000295639038},{"id":"https://openalex.org/C207390915","wikidata":"https://www.wikidata.org/wiki/Q1230525","display_name":"Divergence (linguistics)","level":2,"score":0.5430999994277954},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5303999781608582},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.5250999927520752},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.5220999717712402},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.5180000066757202},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.5069000124931335},{"id":"https://openalex.org/C18653775","wikidata":"https://www.wikidata.org/wiki/Q1333358","display_name":"Joint probability distribution","level":2,"score":0.49570000171661377},{"id":"https://openalex.org/C61224824","wikidata":"https://www.wikidata.org/wiki/Q2260434","display_name":"Mixture model","level":2,"score":0.40059998631477356},{"id":"https://openalex.org/C197656967","wikidata":"https://www.wikidata.org/wiki/Q17058458","display_name":"Marginal model","level":3,"score":0.38679999113082886},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3763999938964844},{"id":"https://openalex.org/C177384507","wikidata":"https://www.wikidata.org/wiki/Q1149000","display_name":"Multivariate normal distribution","level":3,"score":0.36629998683929443},{"id":"https://openalex.org/C95923904","wikidata":"https://www.wikidata.org/wiki/Q6760420","display_name":"Marginal likelihood","level":3,"score":0.36570000648498535},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3652999997138977},{"id":"https://openalex.org/C161584116","wikidata":"https://www.wikidata.org/wiki/Q1952580","display_name":"Multivariate statistics","level":2,"score":0.36070001125335693},{"id":"https://openalex.org/C2777606061","wikidata":"https://www.wikidata.org/wiki/Q16991115","display_name":"Tail dependence","level":3,"score":0.34700000286102295},{"id":"https://openalex.org/C32230216","wikidata":"https://www.wikidata.org/wiki/Q7882499","display_name":"Uncertainty quantification","level":2,"score":0.3386000096797943},{"id":"https://openalex.org/C110121322","wikidata":"https://www.wikidata.org/wiki/Q865811","display_name":"Distribution (mathematics)","level":2,"score":0.3172000050544739},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.31619998812675476},{"id":"https://openalex.org/C113364801","wikidata":"https://www.wikidata.org/wiki/Q26674","display_name":"High fidelity","level":2,"score":0.3070000112056732},{"id":"https://openalex.org/C98763669","wikidata":"https://www.wikidata.org/wiki/Q176645","display_name":"Markov chain","level":2,"score":0.2955999970436096},{"id":"https://openalex.org/C64341305","wikidata":"https://www.wikidata.org/wiki/Q4919225","display_name":"Bivariate analysis","level":2,"score":0.2928999960422516},{"id":"https://openalex.org/C17618745","wikidata":"https://www.wikidata.org/wiki/Q207509","display_name":"Copula (linguistics)","level":2,"score":0.28029999136924744},{"id":"https://openalex.org/C185142706","wikidata":"https://www.wikidata.org/wiki/Q1134404","display_name":"Covariance matrix","level":2,"score":0.2750999927520752},{"id":"https://openalex.org/C149441793","wikidata":"https://www.wikidata.org/wiki/Q200726","display_name":"Probability distribution","level":2,"score":0.27309998869895935},{"id":"https://openalex.org/C114289077","wikidata":"https://www.wikidata.org/wiki/Q3284399","display_name":"Statistical model","level":2,"score":0.27230000495910645},{"id":"https://openalex.org/C139676723","wikidata":"https://www.wikidata.org/wiki/Q1193832","display_name":"Sign (mathematics)","level":2,"score":0.26989999413490295},{"id":"https://openalex.org/C134261354","wikidata":"https://www.wikidata.org/wiki/Q938438","display_name":"Statistical inference","level":2,"score":0.2662000060081482},{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.26350000500679016},{"id":"https://openalex.org/C83546350","wikidata":"https://www.wikidata.org/wiki/Q1139051","display_name":"Regression","level":2,"score":0.26019999384880066},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.25920000672340393},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2517000138759613}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.17041","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.17041","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,"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.2603.17041","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.17041","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,"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":{"Generative":[0],"models":[1],"are":[2],"increasingly":[3],"deployed":[4],"as":[5,66,138],"substitutes":[6],"for":[7,71,134],"real":[8],"data":[9,82,149],"in":[10,123,181],"downstream":[11,32,118],"scientific":[12],"workflows,":[13],"yet":[14],"standard":[15,184],"evaluation":[16,199],"criteria":[17],"remain":[18],"focused":[19],"on":[20,98],"marginal":[21,37,93,185],"distribution":[22],"matching.":[23],"We":[24,54],"argue":[25],"that":[26,42,173,191],"this":[27,90],"represents":[28],"a":[29,36,40,67,74,162],"fundamental":[30],"gap:":[31],"inference":[33],"is":[34],"rarely":[35],"operation,":[38],"and":[39,161,203],"model":[41,76],"passes":[43],"every":[44],"univariate":[45,85,109],"diagnostic":[46],"can":[47,102],"still":[48],"produce":[49],"structurally":[50],"unreliable":[51],"synthetic":[52],"data.":[53],"introduce":[55],"covariance-level":[56,192],"dependence":[57,99],"fidelity,":[58],"measured":[59],"by":[60],"D_Sigma(P,Q)":[61],"=":[62,153,159,170],"||Sigma_P":[63],"-":[64],"Sigma_Q||_F,":[65],"principled,":[68],"computable":[69],"criterion":[70],"evaluating":[72],"whether":[73],"generative":[75],"preserves":[77],"the":[78],"joint":[79],"structure":[80],"of":[81],"beyond":[83],"its":[84],"marginals.":[86],"Three":[87],"results":[88],"formalise":[89],"criterion.":[91],"First,":[92],"fidelity":[94,193],"provides":[95,130,194],"no":[96],"constraint":[97],"structure:":[100],"D_Sigma":[101,129],"be":[103],"made":[104],"arbitrarily":[105],"large":[106],"while":[107],"all":[108],"marginals":[110],"match":[111],"exactly.":[112],"Second,":[113],"covariance":[114],"divergence":[115],"induces":[116],"quantifiable":[117],"instability,":[119],"including":[120],"sign":[121],"reversals":[122],"population":[124],"regression":[125],"coefficients.":[126],"Third,":[127],"bounding":[128],"positive":[131],"stability":[132],"guarantees":[133],"dependence-sensitive":[135],"procedures":[136],"such":[137],"PCA":[139],"via":[140],"Davis-Kahan-type":[141],"bounds.":[142],"Empirical":[143],"validation":[144],"across":[145,201],"three":[146],"domains,":[147],"image":[148],"(Fashion-MNIST":[150],"VAE,":[151],"n":[152,158,169],"60,000),":[154],"bulk":[155],"RNA-seq":[156],"(TCGA-BRCA,":[157],"1,111),":[160],"small-sample":[163],"stress":[164],"test":[165],"(Alzheimer's":[166],"gene":[167],"expression,":[168],"113),":[171],"shows":[172],"D_Sigma/delta":[174],"consistently":[175],"distinguishes":[176],"structure-discarding":[177],"from":[178],"structure-preserving":[179],"generators":[180],"cases":[182],"where":[183],"diagnostics":[186],"show":[187],"little":[188],"separation,":[189],"confirming":[190],"information":[195],"orthogonal":[196],"to":[197],"existing":[198],"metrics":[200],"domains":[202],"sample":[204],"sizes.":[205]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-03-20T00:00:00"}