Tension Adaptation Versus Scheduler Effects on CIFAR-10: A Matched Ablation Study

Tension Adaptation Versus Scheduler Effects on CIFAR-10: A Matched Ablation Study | 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 Research Article Tension Adaptation Versus Scheduler Effects on CIFAR-10: A Matched Ablation Study Datorien L. Anderson This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9089751/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Adaptive training schemes often combine several control components at once, making it unclear which mechanism is actually responsible for improved learning. This paper isolates that question on CIFAR-10 with a matched 4-condition ablation over a fixed convolutional classifier and fixed training protocol. We vary only two components: a scheduler/warmup con troller and a per-batch tension-adaptation mechanism that adjusts optimization behavior from loss-tension signals between successive updates. Under the refreshed checkpoint-backed run set, the base condition reaches 68.89% final accuracy, scheduler-only improves to 71.46%, tension-only reaches 74.42%, and the combined condition reaches 74.66%. Per-sample trajec tory populations show the same pattern: relative to the base run, the stronger conditions reduce persistent uncertainty and increase uncertainty-to-correct conversion, but they do so in differ ent ways. The combined condition preserves the lowest persistent confident-error mass, while tension-only carries the largest UI→CC conversion count. A diagnostic exclusion sweep over CVS tail populations further shows that weak-specification cases explain only a modest part of the remaining ceiling, whereas the volatile cluster is load-bearing for learning rather than disposable noise. In this setting, scheduler effects are real but secondary, tension adaptation re mains the dominant mechanism, and certainty-validity diagnostics are most informative when read together with coverage and per-sample trajectory populations. Artificial Intelligence and Machine Learning CIFAR-10 uncertainty quantification diagnostic ablation per-sample trajectory analysis epistemic calibration learning-rate scheduling adaptive optimization confidence-error decomposition Full Text Additional Declarations The authors declare no competing interests. Supplementary Files cifarartifacts.zip CIFAR Artifacts for the Abalation Study Cite Share Download PDF Status: Posted Version 1 posted 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-9089751","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":604162296,"identity":"af6a6778-421a-4430-a8d5-418f83592cbd","order_by":0,"name":"Datorien L. Anderson","email":"data:image/png;base64,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","orcid":"https://orcid.org/0009-0004-7688-6829","institution":"Occybyte","correspondingAuthor":true,"prefix":"","firstName":"Datorien","middleName":"L.","lastName":"Anderson","suffix":""}],"badges":[],"createdAt":"2026-03-11 04:34:04","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-9089751/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9089751/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104781924,"identity":"4b3e1515-cb33-4e07-88dc-0982cf76a680","added_by":"auto","created_at":"2026-03-17 07:56:34","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2632696,"visible":true,"origin":"","legend":"","description":"","filename":"references.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9089751/v1_covered_360a2847-1331-4c86-a08f-0d7601c2e99f.pdf"},{"id":104560930,"identity":"86eab1fa-cb08-4a1f-9736-29c89a02711f","added_by":"auto","created_at":"2026-03-13 10:14:50","extension":"zip","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":2547250,"visible":true,"origin":"","legend":"\u003cp\u003eCIFAR Artifacts for the Abalation Study\u003c/p\u003e","description":"","filename":"cifarartifacts.zip","url":"https://assets-eu.researchsquare.com/files/rs-9089751/v1/3e95eb9985582c32dac1912d.zip"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eTension Adaptation Versus Scheduler Effects on CIFAR-10: A Matched Ablation Study\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"CIFAR-10, uncertainty quantification, diagnostic ablation, per-sample trajectory analysis, epistemic calibration, learning-rate scheduling, adaptive optimization, confidence-error decomposition","lastPublishedDoi":"10.21203/rs.3.rs-9089751/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9089751/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAdaptive training schemes often combine several control components at once, making it unclear which mechanism is actually responsible for improved learning. This paper isolates that question on CIFAR-10 with a matched 4-condition ablation over a fixed convolutional classifier and fixed training protocol. We vary only two components: a scheduler/warmup con troller and a per-batch tension-adaptation mechanism that adjusts optimization behavior from loss-tension signals between successive updates. Under the refreshed checkpoint-backed run set, the base condition reaches 68.89% final accuracy, scheduler-only improves to 71.46%, tension-only reaches 74.42%, and the combined condition reaches 74.66%. Per-sample trajec tory populations show the same pattern: relative to the base run, the stronger conditions reduce persistent uncertainty and increase uncertainty-to-correct conversion, but they do so in differ ent ways. The combined condition preserves the lowest persistent confident-error mass, while tension-only carries the largest UI→CC conversion count. A diagnostic exclusion sweep over CVS tail populations further shows that weak-specification cases explain only a modest part of the remaining ceiling, whereas the volatile cluster is load-bearing for learning rather than disposable noise. In this setting, scheduler effects are real but secondary, tension adaptation re mains the dominant mechanism, and certainty-validity diagnostics are most informative when read together with coverage and per-sample trajectory populations.\u003c/p\u003e","manuscriptTitle":"Tension Adaptation Versus Scheduler Effects on CIFAR-10: A Matched Ablation Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-13 10:14:42","doi":"10.21203/rs.3.rs-9089751/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0aefec0a-0554-495a-8345-70f7d5ac61bb","owner":[],"postedDate":"March 13th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":64290110,"name":"Artificial Intelligence and Machine Learning"}],"tags":[],"updatedAt":"2026-03-13T10:14:42+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-13 10:14:42","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9089751","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9089751","identity":"rs-9089751","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}