Multi-modal News Recommendation with Deep Text Modeling and Multi‑View Image Fusion

Multi-modal News Recommendation with Deep Text Modeling and Multi‑View Image Fusion | 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 Multi-modal News Recommendation with Deep Text Modeling and Multi‑View Image Fusion Le Zhao This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8223622/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 In the digital media era, the sheer volume of news content has led to information overload, making it challenging for users to find relevant articles efficiently. Existing methods often rely on collaborative filtering and click histories but lack deep semantic modeling of news text, overlooking both local and global word dependencies. Static attention mechanisms further limit the ability to dynamically highlight context‑specific key information, while most multimodal systems extract image features from a single perspective, failing to capture the full visual context. To overcome these limitations, we propose M‑GRU‑DHA (Multi‑view GRU with Dynamic Hybrid Attention), which integrates two modules: GRU‑DHA for text and MV‑NPIC (Multi‑View News Picture Information Capturing) for images. GRU‑DHA combines GRU to model global dependencies and CNN to extract local semantics, augmented by a dynamic hybrid attention mechanism that adaptively focuses on key words. MV‑NPIC enriches multimodal fusion by extracting cover‑image features from multiple viewpoints, thereby capturing comprehensive visual semantics. Extensive experiments on real‑world datasets demonstrate that M‑GRU‑DHA consistently outperforms mainstream methods on standard evaluation metrics, validating the effectiveness of our approach for personalized news recommendation. Theoretical Computer Science Personalized news recommendation Dynamic hybrid attention Deep semantic understanding Multi‑View News Picture Information Capturing Multimodal Fusion Full Text Additional Declarations The authors declare no competing interests. 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-8223622","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":551863557,"identity":"0f38a349-aaed-4e9a-aa3b-553ab917a656","order_by":0,"name":"Le Zhao","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2ElEQVRIiWNgGAWjYBACA4YEhoMNDBYMDOwNIK4F0VokGBh4DoC4EsRpYQRrkUgA8YnQYs6ee/DgjAoJOX7J51c3/CiQYOBv707Aq8Wy513CwQ1nJIwlZ+eU3ewBOkzizNkN+B12I8fg4MM2icQNt3PSbvAAtRhI5BKj5Z9E/f6bZ9Ju/iFay8YGiQQDCfZjt4myxbLnjcHBGcckDGecyWG7LWMgwUPQL+bsOcYfe2ps5Pnbjz+7+eaPjRx/ey9+LUiAxwBMEqscBNgfkKJ6FIyCUTAKRhAAAM4ITGPO9o+tAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0009-0009-6032-4473","institution":"SouthWest Petroleum University","correspondingAuthor":true,"prefix":"","firstName":"Le","middleName":"","lastName":"Zhao","suffix":""}],"badges":[],"createdAt":"2025-11-27 15:48:27","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-8223622/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8223622/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97424224,"identity":"c711030d-f4a7-42ec-b5c0-1571bdf48c8a","added_by":"auto","created_at":"2025-12-04 08:54:27","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":686748,"visible":true,"origin":"","legend":"","description":"","filename":"MultimodalNewsRecommendationwithDeepTextModelingandMultiViewImageFusion.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8223622/v1_covered_fc521911-3e28-4277-8b4b-bd3677380e07.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eMulti-modal News Recommendation with Deep Text Modeling and Multi‑View Image Fusion\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Southwest Petroleum University","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":"Personalized news recommendation, Dynamic hybrid attention, Deep semantic understanding, Multi‑View News Picture Information Capturing, Multimodal Fusion","lastPublishedDoi":"10.21203/rs.3.rs-8223622/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8223622/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn the digital media era, the sheer volume of news content has led to information overload, making it challenging for users to find relevant articles efficiently. Existing methods often rely on collaborative filtering and click histories but lack deep semantic modeling of news text, overlooking both local and global word dependencies. Static attention mechanisms further limit the ability to dynamically highlight context‑specific key information, while most multimodal systems extract image features from a single perspective, failing to capture the full visual context.\u003c/p\u003e\n\u003cp\u003eTo overcome these limitations, we propose M‑GRU‑DHA (Multi‑view GRU with Dynamic Hybrid Attention), which integrates two modules: GRU‑DHA for text and MV‑NPIC (Multi‑View News Picture Information Capturing) for images. GRU‑DHA combines GRU to model global dependencies and CNN to extract local semantics, augmented by a dynamic hybrid attention mechanism that adaptively focuses on key words. MV‑NPIC enriches multimodal fusion by extracting cover‑image features from multiple viewpoints, thereby capturing comprehensive visual semantics. Extensive experiments on real‑world datasets demonstrate that M‑GRU‑DHA consistently outperforms mainstream methods on standard evaluation metrics, validating the effectiveness of our approach for personalized news recommendation.\u003c/p\u003e","manuscriptTitle":"Multi-modal News Recommendation with Deep Text Modeling and Multi‑View Image Fusion","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-04 08:54:07","doi":"10.21203/rs.3.rs-8223622/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":"04b39de7-d685-4dbf-9bdb-b7dfc1104276","owner":[],"postedDate":"December 4th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":58727672,"name":"Theoretical Computer Science"}],"tags":[],"updatedAt":"2025-12-04T08:54:07+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-04 08:54:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8223622","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8223622","identity":"rs-8223622","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}