Efficient Knowledge Distillation for News Classification Based on ModernBERT

Efficient Knowledge Distillation for News Classification Based on ModernBERT | 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 Efficient Knowledge Distillation for News Classification Based on ModernBERT Xuyang Wang, Yuxi Zheng This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9229561/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 8 You are reading this latest preprint version Abstract Large-scale pretrained language models (e.g., BERT-large, RoBERTa-large) achieve strong performance in text classification; however, their substantial computational and energy costs hinder deployment, whereas base counterparts still lag in accuracy. ModernBERT-large and ModernBERT-base are adopted as a teacher–student pair to systematically explore three knowledge distillation (KD) strategies—full-sample output-layer distillation, selective output-layer distillation, and attention-layer distillation—for news classification. To assess sustainability, a carbon-efficiency metric (Accuracy per kWh) is introduced. Experiments demonstrate that the distilled student model improves ACC by 0.74% and boosts energy efficiency by 138.9% over the teacher on AG News; on 20 Newsgroups, ACC increases by 2.62% with a 133.6% efficiency gain. The results validate the effectiveness of a green distillation framework. Knowledge Distillation ModernBERT News Classification Energy Efficiency Carbon Efficiency Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 07 May, 2026 Reviewers agreed at journal 26 Apr, 2026 Reviews received at journal 02 Apr, 2026 Reviewers agreed at journal 02 Apr, 2026 Reviewers invited by journal 02 Apr, 2026 Editor assigned by journal 31 Mar, 2026 Submission checks completed at journal 27 Mar, 2026 First submitted to journal 26 Mar, 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-9229561","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":616898202,"identity":"981a38e2-361c-41e7-864c-2bd7c4bbb869","order_by":0,"name":"Xuyang Wang","email":"","orcid":"","institution":"Lanzhou University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Xuyang","middleName":"","lastName":"Wang","suffix":""},{"id":616898206,"identity":"1fcd408e-2572-4a17-b250-b348e278dab3","order_by":1,"name":"Yuxi Zheng","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCElEQVRIie3PMUvDQBTA8RcCl+W1WV9Q7Fc4CISChXyVyxKXDoIgHYpEhLj0A6Qfw0Ucrzy46cC1g0sWJ4cUFwdBbR3rRbs53H94w8GP9w7A5/uHxcBab94FxmGodXe+e5S9JLk2RbushsfJbV2sGgn0K5Fs03RQnUyktSnjXwgYnSXLhwxhrSRPJF9BdHNPMH9yimChS9rYEoNGKZ5KJkBzSWCenSQkbZJGGAxJ6W9C04yCip1EjNr6CMUHCioqHm/J6KWfIHCYDmqBiAwMuy3YT76uDtrGCqSohtVCniU1lhdjZdwk14+d7mYizzl+7d5mp3Ec8d26m7vJD7/bDnUA8Pl8Pt9+n4AgWOPXj7W/AAAAAElFTkSuQmCC","orcid":"","institution":"Lanzhou University of Technology","correspondingAuthor":true,"prefix":"","firstName":"Yuxi","middleName":"","lastName":"Zheng","suffix":""}],"badges":[],"createdAt":"2026-03-26 05:39:24","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9229561/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9229561/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106389369,"identity":"24fce4cd-2521-4248-a102-6e893c8e8d97","added_by":"auto","created_at":"2026-04-08 06:57:44","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":274812,"visible":true,"origin":"","legend":"","description":"","filename":"ModernBERTen.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9229561/v1_covered_7217fbbf-d16b-4f88-ab19-5899b416afc3.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Efficient Knowledge Distillation for News Classification Based on ModernBERT","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"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":"international-journal-of-machine-learning-and-cybernetics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jmlc","sideBox":"Learn more about [International Journal of Machine Learning and Cybernetics](http://actavetscand.biomedcentral.com/)","snPcode":"13042","submissionUrl":"https://submission.nature.com/new-submission/13042/3","title":"International Journal of Machine Learning and Cybernetics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Knowledge Distillation, ModernBERT, News Classification, Energy Efficiency, Carbon Efficiency","lastPublishedDoi":"10.21203/rs.3.rs-9229561/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9229561/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eLarge-scale pretrained language models (e.g., BERT-large, RoBERTa-large) achieve strong performance in text classification; however, their substantial computational and energy costs hinder deployment, whereas base counterparts still lag in accuracy. ModernBERT-large and ModernBERT-base are adopted as a teacher\u0026ndash;student pair to systematically explore three knowledge distillation (KD) strategies\u0026mdash;full-sample output-layer distillation, selective output-layer distillation, and attention-layer distillation\u0026mdash;for news classification. To assess sustainability, a carbon-efficiency metric (Accuracy per kWh) is introduced. Experiments demonstrate that the distilled student model improves ACC by 0.74% and boosts energy efficiency by 138.9% over the teacher on AG News; on 20 Newsgroups, ACC increases by 2.62% with a 133.6% efficiency gain. The results validate the effectiveness of a green distillation framework.\u003c/p\u003e","manuscriptTitle":"Efficient Knowledge Distillation for News Classification Based on ModernBERT","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-08 06:56:21","doi":"10.21203/rs.3.rs-9229561/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-05-07T11:36:21+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"311268180424449160057070638044751783525","date":"2026-04-26T14:57:54+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-03T02:41:37+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"255579218471352820687691487702742296933","date":"2026-04-03T00:58:39+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-02T16:32:45+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-31T13:51:50+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-27T07:39:08+00:00","index":"","fulltext":""},{"type":"submitted","content":"International Journal of Machine Learning and Cybernetics","date":"2026-03-26T05:32:47+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"international-journal-of-machine-learning-and-cybernetics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jmlc","sideBox":"Learn more about [International Journal of Machine Learning and Cybernetics](http://actavetscand.biomedcentral.com/)","snPcode":"13042","submissionUrl":"https://submission.nature.com/new-submission/13042/3","title":"International Journal of Machine Learning and Cybernetics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"0c727d1b-64c4-429e-90a6-f5b98ceb09b8","owner":[],"postedDate":"April 8th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-07T11:36:21+00:00","index":35,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-08T06:56:22+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-08 06:56:21","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9229561","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9229561","identity":"rs-9229561","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}