Dual-tower Feature Fusion for Student Ontology and Explainable Knowledge Tracing

Dual-tower Feature Fusion for Student Ontology and Explainable Knowledge Tracing | 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 Dual-tower Feature Fusion for Student Ontology and Explainable Knowledge Tracing Li Yang, Yujia Huo, Xue Tan, Yao Wang, Changxiao Yang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5264599/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 Knowledge Tracing (KT) aims to model students' knowledge states through their learning interactions. While Deep Knowledge Tracing (DKT) leverages deep learning to capture complex learning patterns, existing methods often fail to jointly exploit structured student features and semantic relationships of textual knowledge components (KCs), limiting both performance and explainability. {\color{blue}To address these, we propose Feature Fusion of Dual-tower Knowledge Tracing (FFTKT) that consists of three key components: the student tower, the knowledge tower, and the fusion module.The student tower employs a Transformer encoder to process sequential student behavioral features. The knowledge tower, built upon a BERT-based encoder, models textual features of KCs. The Fusion Module dynamically aligns student and KCs interactions through a self-attention mechanism augmented with learnable memory tokens. Experimental results on ASSISTment2012/2017 demonstrate that FFTKT achieves superior AUC performance, outperforming DKT+ by 1.3 \(%\) to 6.7 $ % $ . Through feature visualization, FFTKT reliably explains the correlations between student behaviors and KC mastery states.} Knowledge Tracing Dual-Tower Architecture Self-Attention Mechanism Transformer Encoder Memory Tokens Full Text 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-5264599","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":447810626,"identity":"eb795731-ace6-4318-9b99-6ca8f7f44225","order_by":0,"name":"Li Yang","email":"","orcid":"","institution":"Guizhou Minzu University","correspondingAuthor":false,"prefix":"","firstName":"Li","middleName":"","lastName":"Yang","suffix":""},{"id":447810627,"identity":"008c037a-9146-4727-bf25-8c904bcb76fc","order_by":1,"name":"Yujia Huo","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4ElEQVRIiWNgGAWjYBACAwjFxsPA3gAVOkC0Fh6YUiK1AIFEApFazCWSDzDztvHJ8M98fvnTzTYGOb4bCYyfC/BosZyRlgDUwsYjcTunwDi3jcFY8kYCs/QMfA67kWMA1sJwOychGaglccONBDZmHmK0yN88k3AYqKWeeC0GN9gPNgO1JBgQ1HLmWcLBOefYeAzP5DAz55yTMJx55mGzNF4tx5MPPnhTdsxe7vjxx59zymzk+YAin/FpAYFDPAzHgBQPKI4kgJixgYAGoJIfDDVAiv0BQZWjYBSMglEwMgEAelRJ0iDaMDoAAAAASUVORK5CYII=","orcid":"https://orcid.org/0009-0003-5307-7474","institution":"Guizhou Minzu University","correspondingAuthor":true,"prefix":"","firstName":"Yujia","middleName":"","lastName":"Huo","suffix":""},{"id":447810628,"identity":"010d21ea-403e-4f99-8dbd-9daf06be7ffe","order_by":2,"name":"Xue Tan","email":"","orcid":"","institution":"Guizhou Minzu University","correspondingAuthor":false,"prefix":"","firstName":"Xue","middleName":"","lastName":"Tan","suffix":""},{"id":447810629,"identity":"32e1e3b5-0d5b-47bd-85bd-d0d2df9513e4","order_by":3,"name":"Yao Wang","email":"","orcid":"","institution":"Guizhou Minzu University","correspondingAuthor":false,"prefix":"","firstName":"Yao","middleName":"","lastName":"Wang","suffix":""},{"id":447810630,"identity":"56c0e5bb-bdb5-4238-8138-ac087bee38bc","order_by":4,"name":"Changxiao Yang","email":"","orcid":"","institution":"Guizhou Minzu University","correspondingAuthor":false,"prefix":"","firstName":"Changxiao","middleName":"","lastName":"Yang","suffix":""}],"badges":[],"createdAt":"2024-10-15 03:15:55","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5264599/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5264599/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":94612284,"identity":"1c4e45c8-d146-402a-b29c-6a1ae680825d","added_by":"auto","created_at":"2025-10-29 02:08:20","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1175209,"visible":true,"origin":"","legend":"","description":"","filename":"SOCOD2404563R1reviewer.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5264599/v1_covered_589e6758-52c2-4a96-84fa-2503f4cb76c3.pdf"}],"financialInterests":"","formattedTitle":"Dual-tower Feature Fusion for Student Ontology and Explainable Knowledge Tracing","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"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":"Knowledge Tracing, Dual-Tower Architecture, Self-Attention Mechanism, Transformer Encoder, Memory Tokens","lastPublishedDoi":"10.21203/rs.3.rs-5264599/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5264599/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eKnowledge Tracing (KT) aims to model students' knowledge states through their learning interactions. While Deep Knowledge Tracing (DKT) leverages deep learning to capture complex learning patterns, existing methods often fail to jointly exploit structured student features and semantic relationships of textual knowledge components (KCs), limiting both performance and explainability. {\\color{blue}To address these, we propose Feature Fusion of Dual-tower Knowledge Tracing (FFTKT) that consists of three key components: the student tower, the knowledge tower, and the fusion module.The student tower employs a Transformer encoder to process sequential student behavioral features. The knowledge tower, built upon a BERT-based encoder, models textual features of KCs. The Fusion Module dynamically aligns student and KCs interactions through a self-attention mechanism augmented with learnable memory tokens. Experimental results on ASSISTment2012/2017 demonstrate that FFTKT achieves superior AUC performance, outperforming DKT+ by 1.3\u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e \\(%\\) \u003c/span\u003e \u003c/span\u003e to 6.7\u003cspan\u003e$\u003c/span\u003e%\u003cspan\u003e$\u003c/span\u003e. Through feature visualization, FFTKT reliably explains the correlations between student behaviors and KC mastery states.}\u003c/p\u003e","manuscriptTitle":"Dual-tower Feature Fusion for Student Ontology and Explainable Knowledge Tracing","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-30 09:06:38","doi":"10.21203/rs.3.rs-5264599/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":"7a4fa5eb-5574-442c-9913-ec6c9a4aa21e","owner":[],"postedDate":"April 30th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-10-28T21:42:30+00:00","versionOfRecord":[],"versionCreatedAt":"2025-04-30 09:06:38","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5264599","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5264599","identity":"rs-5264599","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}