{"id":"https://openalex.org/W7162074997","doi":"https://doi.org/10.48550/arxiv.2605.21731","title":"I-SAFE: Wasserstein Coherence Metrics for Structural Auditing of Scientific AI Models","display_name":"I-SAFE: Wasserstein Coherence Metrics for Structural Auditing of Scientific AI Models","publication_year":2026,"publication_date":"2026-05-20","ids":{"openalex":"https://openalex.org/W7162074997","doi":"https://doi.org/10.48550/arxiv.2605.21731"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.21731","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.21731","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.21731","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5059863311","display_name":"Barbara Tarantino","orcid":"https://orcid.org/0000-0001-6384-2561"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tarantino, Barbara","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136730189","display_name":"Gennaro Auricchio","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Auricchio, Gennaro","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5136811022","display_name":"Paolo Giudici","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Giudici, Paolo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":3,"corresponding_author_ids":[],"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.18690000474452972,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.18690000474452972,"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/T12859","display_name":"Cell Image Analysis Techniques","score":0.181099995970726,"subfield":{"id":"https://openalex.org/subfields/1304","display_name":"Biophysics"},"field":{"id":"https://openalex.org/fields/13","display_name":"Biochemistry, Genetics and Molecular Biology"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T10211","display_name":"Computational Drug Discovery Methods","score":0.15600000321865082,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/coherence","display_name":"Coherence (philosophical gambling strategy)","score":0.6546000242233276},{"id":"https://openalex.org/keywords/metric","display_name":"Metric (unit)","score":0.6265000104904175},{"id":"https://openalex.org/keywords/exploit","display_name":"Exploit","score":0.5595999956130981},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.49380001425743103},{"id":"https://openalex.org/keywords/domain","display_name":"Domain (mathematical analysis)","score":0.4839000105857849},{"id":"https://openalex.org/keywords/audit","display_name":"Audit","score":0.43149998784065247},{"id":"https://openalex.org/keywords/domain-knowledge","display_name":"Domain knowledge","score":0.42879998683929443},{"id":"https://openalex.org/keywords/interpretation","display_name":"Interpretation (philosophy)","score":0.4147999882698059},{"id":"https://openalex.org/keywords/raw-data","display_name":"Raw data","score":0.3935999870300293}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6790000200271606},{"id":"https://openalex.org/C2781181686","wikidata":"https://www.wikidata.org/wiki/Q4226068","display_name":"Coherence (philosophical gambling strategy)","level":2,"score":0.6546000242233276},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.6265000104904175},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5976999998092651},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.5595999956130981},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5435000061988831},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.49380001425743103},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.4839000105857849},{"id":"https://openalex.org/C199521495","wikidata":"https://www.wikidata.org/wiki/Q181487","display_name":"Audit","level":2,"score":0.43149998784065247},{"id":"https://openalex.org/C207685749","wikidata":"https://www.wikidata.org/wiki/Q2088941","display_name":"Domain knowledge","level":2,"score":0.42879998683929443},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.4198000133037567},{"id":"https://openalex.org/C527412718","wikidata":"https://www.wikidata.org/wiki/Q855395","display_name":"Interpretation (philosophy)","level":2,"score":0.4147999882698059},{"id":"https://openalex.org/C132964779","wikidata":"https://www.wikidata.org/wiki/Q2110223","display_name":"Raw data","level":2,"score":0.3935999870300293},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.3662000000476837},{"id":"https://openalex.org/C45804977","wikidata":"https://www.wikidata.org/wiki/Q7239673","display_name":"Predictive modelling","level":2,"score":0.3549000024795532},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.31200000643730164},{"id":"https://openalex.org/C17231256","wikidata":"https://www.wikidata.org/wiki/Q5156540","display_name":"Completeness (order theory)","level":2,"score":0.301800012588501},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.30149999260902405},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.29739999771118164},{"id":"https://openalex.org/C2780898871","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Performance metric","level":2,"score":0.2775000035762787},{"id":"https://openalex.org/C2780522230","wikidata":"https://www.wikidata.org/wiki/Q1140419","display_name":"Ambiguity","level":2,"score":0.2759000062942505},{"id":"https://openalex.org/C2776436953","wikidata":"https://www.wikidata.org/wiki/Q5163215","display_name":"Consistency (knowledge bases)","level":2,"score":0.27079999446868896},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.26989999413490295},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.2615000009536743},{"id":"https://openalex.org/C107908354","wikidata":"https://www.wikidata.org/wiki/Q4781456","display_name":"Applicability domain","level":3,"score":0.25940001010894775},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2558000087738037},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.25529998540878296},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.251800000667572}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.21731","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.21731","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.21731","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.21731","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"sustainable_development_goals":[{"score":0.5085004568099976,"display_name":"Peace, Justice and strong institutions","id":"https://metadata.un.org/sdg/16"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Deep":[0],"learning":[1],"models":[2,28,73,172],"are":[3,39],"increasingly":[4],"used":[5],"in":[6,174],"scientific":[7,71],"prediction":[8,147],"tasks":[9],"where":[10,201],"strong":[11],"benchmark":[12],"performance":[13],"is":[14,25,194,211],"often":[15],"interpreted":[16],"as":[17,27],"evidence":[18],"of":[19,107],"scientifically":[20],"meaningful":[21],"behavior.":[22],"This":[23],"interpretation":[24],"fragile,":[26],"may":[29],"exploit":[30],"shortcut":[31],"features,":[32],"dataset-specific":[33],"regularities,":[34],"or":[35],"distributional":[36,67,183],"biases":[37],"that":[38],"predictive":[40,177],"on":[41,75,143],"held-out":[42],"data":[43],"but":[44],"not":[45],"aligned":[46],"with":[47],"domain-relevant":[48],"structure.":[49],"To":[50],"address":[51],"this":[52],"limitation,":[53],"we":[54],"introduce":[55],"the":[56,76,108,127,149,171],"\\textsc{I-SAFE}":[57,98,142,179],"(Interventional":[58],"Secure,":[59],"Accurate,":[60],"Fair":[61],"and":[62,86,132,157,161,168,196,207],"Explainable)":[63],"framework,":[64],"a":[65,82,120,133,175,186,204],"post-hoc":[66],"auditing":[68],"framework":[69,193],"for":[70,124,129,137],"AI":[72],"centered":[74],"Wasserstein":[77],"Coherence":[78],"Metric":[79,122],"(WCM).":[80],"Given":[81],"trained":[83],"black-box":[84],"predictor":[85],"an":[87,208],"external":[88,209],"structural":[89],"prior":[90,210],"encoding":[91],"domain":[92,200],"knowledge":[93],"about":[94],"task-relevant":[95],"input":[96],"structure,":[97],"evaluates":[99],"raw":[100],"model":[101],"outputs":[102],"under":[103],"structurally":[104],"guided":[105],"perturbations":[106],"input.":[109],"The":[110,192],"proposed":[111],"audit":[112],"measures":[113],"output-distribution":[114],"coherence":[115],"through":[116],"three":[117,162],"complementary":[118],"metrics:":[119],"Quantile-Based":[121],"(QBM)":[123],"location-level":[125],"coherence,":[126,131],"WCM":[128,135],"ordinal":[130],"translation-invariant":[134],"variant":[136],"shape":[138],"coherence.":[139],"We":[140],"instantiate":[141],"drug--target":[144],"interaction":[145],"(DTI)":[146],"using":[148],"Davis":[150],"kinase":[151],"benchmark,":[152],"KLIFS":[153],"(Kinase--Ligand":[154],"Interaction":[155],"Fingerprints":[156],"Structures)":[158],"binding-pocket":[159],"annotations,":[160],"sequence-based":[163],"DTI":[164],"models:":[165],"DeepConvDTI,":[166],"DeepDTA,":[167],"TAPB.":[169],"Although":[170],"operate":[173],"comparable":[176],"regime,":[178],"reveals":[180],"substantially":[181],"different":[182],"response":[184],"profiles,":[185],"distinction":[187],"invisible":[188],"to":[189,198],"accuracy-based":[190],"evaluation.":[191],"model-agnostic":[195],"applicable":[197],"any":[199],"inputs":[202],"admit":[203],"structured":[205],"decomposition":[206],"available.":[212]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-05-23T00:00:00"}