Epistemic Frontiers: Distinguishing Causality, Information, and Predictability in Pattern Recognition

Epistemic Frontiers: Distinguishing Causality, Information, and Predictability in Pattern Recognition | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Epistemic Frontiers: Distinguishing Causality, Information, and Predictability in Pattern Recognition Pablo Neirz, Héctor Allende, Carolina Saavedra This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8712176/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 12 May, 2026 Read the published version in Scientific Reports → Version 1 posted 10 You are reading this latest preprint version Abstract High predictive accuracy is frequently misinterpreted as evidence of causal understanding or population-level signal. Models can exploit spurious correlations, confounding, or protocol-induced artefacts, while post-hoc explanations may faithfully describe model behaviour yet remain misleading about the underlying phenomenon. We propose a framework that separates three layers of evidence: (i)~causal relations in the phenomenon, (ii)~population-level statistical dependence, and (iii)~finite-sample, protocol-dependent predictive effects. This separation clarifies why predictive success and feature attributions do not license mechanistic interpretations without additional assumptions. Under log-loss and Bayes-risk-consistent protocols, the population predictive value of adding a feature equals the conditional mutual information, providing a principled reference for ''true signal.'' Using controlled simulations, we illustrate that bootstrap resampling can create false positives by amplifying chance correlations, and that SHAP can assign high importance to confounded variables while remaining faithful to the fitted model. These results suggest that ''feature importance'' is best treated as protocol-bounded evidence, and that interpretation benefits from reporting the protocol, robustness checks, and the intended inferential scope. Biological sciences/Computational biology and bioinformatics Physical sciences/Mathematics and computing causal machine learning interpretable machine learning feature attribution conditional mutual information Rashomon effect confounding trustworthy AI Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 12 May, 2026 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 20 Mar, 2026 Reviews received at journal 17 Mar, 2026 Reviews received at journal 01 Mar, 2026 Reviewers agreed at journal 18 Feb, 2026 Reviewers agreed at journal 17 Feb, 2026 Reviewers invited by journal 17 Feb, 2026 Editor invited by journal 17 Feb, 2026 Editor assigned by journal 28 Jan, 2026 Submission checks completed at journal 28 Jan, 2026 First submitted to journal 27 Jan, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8712176","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":593399771,"identity":"52c670c4-b356-4f49-8c50-8813ae8ee8af","order_by":0,"name":"Pablo Neirz","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAoklEQVRIiWNgGAWjYJCCAwwMNhDWAxK0pEFYCSRYdJgELebSzQcPV/w5nzh/2gHGD0RpsZxzLOHgGZ7biRtuJzBLEKXF4EaOwcEGCaAW6QQ24hxmcCP/w8EGg3OJ82cTryWH4WBDwoHEhtvEarGckQZ02IFk4w23E5uJ84u5RPLjjw1/7GTnz04++OEDUQ5DMBkbiNGAomUUjIJRMApGAQ4AANWEN4jMm50pAAAAAElFTkSuQmCC","orcid":"","institution":"Federico Santa María Technical University","correspondingAuthor":true,"prefix":"","firstName":"Pablo","middleName":"","lastName":"Neirz","suffix":""},{"id":593399775,"identity":"b4d207ac-f0b8-4492-b6db-7ea9886ad6a1","order_by":1,"name":"Héctor Allende","email":"","orcid":"","institution":"Federico Santa María Technical University","correspondingAuthor":false,"prefix":"","firstName":"Héctor","middleName":"","lastName":"Allende","suffix":""},{"id":593399780,"identity":"765e3122-0d94-4d22-a5ca-61ee0cf78702","order_by":2,"name":"Carolina Saavedra","email":"","orcid":"","institution":"Federico Santa María Technical University","correspondingAuthor":false,"prefix":"","firstName":"Carolina","middleName":"","lastName":"Saavedra","suffix":""}],"badges":[],"createdAt":"2026-01-27 15:23:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8712176/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8712176/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-026-52883-z","type":"published","date":"2026-05-13T00:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":109210295,"identity":"ae2e2050-9c14-4b3d-9aab-47a9e2dbdc23","added_by":"auto","created_at":"2026-05-13 15:44:54","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":565485,"visible":true,"origin":"","legend":"","description":"","filename":"EpistemicFrontiers.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8712176/v1_covered_935cc0c6-3688-484e-a2c6-576b5e40beff.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Epistemic Frontiers: Distinguishing Causality, Information, and Predictability in Pattern Recognition","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"causal machine learning, interpretable machine learning, feature attribution, conditional mutual information, Rashomon effect, confounding, trustworthy AI","lastPublishedDoi":"10.21203/rs.3.rs-8712176/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8712176/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"High predictive accuracy is frequently misinterpreted as evidence of causal understanding or population-level signal. Models can exploit spurious correlations, confounding, or protocol-induced artefacts, while post-hoc explanations may faithfully describe model behaviour yet remain misleading about the underlying phenomenon. We propose a framework that separates three layers of evidence: (i)~causal relations in the phenomenon, (ii)~population-level statistical dependence, and (iii)~finite-sample, protocol-dependent predictive effects. This separation clarifies why predictive success and feature attributions do not license mechanistic interpretations without additional assumptions. Under log-loss and Bayes-risk-consistent protocols, the population predictive value of adding a feature equals the conditional mutual information, providing a principled reference for ''true signal.'' Using controlled simulations, we illustrate that bootstrap resampling can create false positives by amplifying chance correlations, and that SHAP can assign high importance to confounded variables while remaining faithful to the fitted model. These results suggest that ''feature importance'' is best treated as protocol-bounded evidence, and that interpretation benefits from reporting the protocol, robustness checks, and the intended inferential scope.","manuscriptTitle":"Epistemic Frontiers: Distinguishing Causality, Information, and Predictability in Pattern Recognition","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-20 02:34:45","doi":"10.21203/rs.3.rs-8712176/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-20T12:24:52+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-17T23:24:05+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-01T13:35:27+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"9075675819823980582437340273139362758","date":"2026-02-18T14:54:12+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"192810935299330432316902581820246083876","date":"2026-02-17T21:12:29+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-17T19:27:29+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-02-17T16:12:44+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-01-28T12:23:38+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-01-28T12:19:21+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2026-01-27T15:10:57+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"ddda7ba6-58ec-42f2-83cd-afaee28f1159","owner":[],"postedDate":"February 20th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":63144735,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":63144736,"name":"Physical sciences/Mathematics and computing"}],"tags":[],"updatedAt":"2026-05-13T15:44:49+00:00","versionOfRecord":{"articleIdentity":"rs-8712176","link":"https://doi.org/10.1038/s41598-026-52883-z","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2026-05-13 00:00:00","publishedOnDateReadable":"May 13th, 2026"},"versionCreatedAt":"2026-02-20 02:34:45","video":"","vorDoi":"10.1038/s41598-026-52883-z","vorDoiUrl":"https://doi.org/10.1038/s41598-026-52883-z","workflowStages":[]},"version":"v1","identity":"rs-8712176","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8712176","identity":"rs-8712176","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}