DDCAF: Dynamic Dual Cross-Attention Fusion Framework for Multimodal Hate Speech Detection

DDCAF: Dynamic Dual Cross-Attention Fusion Framework for Multimodal Hate Speech Detection | 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 DDCAF: Dynamic Dual Cross-Attention Fusion Framework for Multimodal Hate Speech Detection Gauri Kitukale, Navneet Pratap Singh, Sidharth Quamara This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7506109/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 11 You are reading this latest preprint version Abstract The rapid spread of offensive and violent speech over social media platforms adds a significant threat to community harmony. In particular, hateful memes are multimodal artifacts that combine images with text to convey implicit or sarcastic hate cues and remain difficult to detect, as they often bypass traditional unimodal detection methods. To address this problem, we proposed DDCAF (Dynamic Dual Cross-Attention Fusion), a novel multimodal framework for detecting hate speech that integrates profound semantic comprehension from both visual and textual modalities. It uses a dual-stream architecture consisting of a RoBERTa-based text encoder and a Vision Transformer (ViT)-based image encoder. Using a bidirectional cross-attention mechanism, the model dynamically computes text-guided visual attention and visual-guided text attention, enabling it to prioritize semantically aligned features across modalities. The dynamic attention-driven fusion mechanism is able to identify subtle, context-dependent cues of hate intent.The proposed framework is primarily tested on multimodal benchmark datasets such as Hateful Memes and MMHS150K, while also assessing unimodal baselines (HateEval and OLID). The experimental findings reveal that it surpasses existing approaches and provides an accuracy of 89.35% on Hateful Memes and 91.20% on MMHS150K. Furthermore, ablation studies are carried out to demonstrate the impact of the subcomponent within the DDCAF, underscoring the importance of dynamic adaptive gating in capturing intermodal dependencies. Multimodal Learning Multimodal Hate Speech Detection Social Media Analysis Image-Text Fusion Vision-Language Models Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 01 Nov, 2025 Reviews received at journal 17 Oct, 2025 Reviews received at journal 20 Sep, 2025 Reviewers agreed at journal 10 Sep, 2025 Reviews received at journal 09 Sep, 2025 Reviewers agreed at journal 05 Sep, 2025 Reviewers agreed at journal 05 Sep, 2025 Reviewers invited by journal 05 Sep, 2025 Editor assigned by journal 04 Sep, 2025 Submission checks completed at journal 02 Sep, 2025 First submitted to journal 01 Sep, 2025 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-7506109","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":514481661,"identity":"766a9c12-a2a6-4886-acd5-afd4f3acb63e","order_by":0,"name":"Gauri Kitukale","email":"","orcid":"","institution":"Bennett University","correspondingAuthor":false,"prefix":"","firstName":"Gauri","middleName":"","lastName":"Kitukale","suffix":""},{"id":514481662,"identity":"808bd899-b9d4-44ed-962c-6ceb30183b82","order_by":1,"name":"Navneet Pratap Singh","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7UlEQVRIiWNgGAWjYFCDw0D8AYjZ2IlTbwDSwtg4A6SFmWgtBxgYm3lAbEJazNnPmD34UPGHge848/PHNr+2yfMxMzB++JiDW4tlT4654YwzBgySh9kMm3P7bhu2MTMwS87chsdFB9LSpHnbDBgMDjMAtfTcZgRqYWPmxafl/LM06b//QFrYPzZb9ty2J6zlRvIxacYGkBYew2aGH7cTidDy+JhkzzFjHsnDPIUzextuJ7cxMzbj98v5xDaJHzVycnznj2/48OPPbdv57c0HP3zEowUGwDHCwNgGJhsIq0eAP6QoHgWjYBSMgpECANyrT5vZrGrxAAAAAElFTkSuQmCC","orcid":"","institution":"Bennett University","correspondingAuthor":true,"prefix":"","firstName":"Navneet","middleName":"Pratap","lastName":"Singh","suffix":""},{"id":514481663,"identity":"0150064c-d3fc-4c56-a40f-c8020a8b20d8","order_by":2,"name":"Sidharth Quamara","email":"","orcid":"","institution":"Bennett University","correspondingAuthor":false,"prefix":"","firstName":"Sidharth","middleName":"","lastName":"Quamara","suffix":""}],"badges":[],"createdAt":"2025-09-01 08:38:15","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7506109/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7506109/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":91358833,"identity":"3381fe14-4107-4b5b-983f-e9894e5b89ff","added_by":"auto","created_at":"2025-09-15 15:57:07","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":735925,"visible":true,"origin":"","legend":"","description":"","filename":"MultimodalHateSpeechdetectionFinal.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7506109/v1_covered_734b7470-1bef-48d7-8d40-c03b2ab0574b.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"DDCAF: Dynamic Dual Cross-Attention Fusion Framework for Multimodal Hate Speech Detection","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":"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":"Multimodal Learning, Multimodal Hate Speech Detection, Social Media Analysis, Image-Text Fusion, Vision-Language Models","lastPublishedDoi":"10.21203/rs.3.rs-7506109/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7506109/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The rapid spread of offensive and violent speech over social media platforms adds a significant threat to community harmony. In particular, hateful memes are multimodal artifacts that combine images with text to convey implicit or sarcastic hate cues and remain difficult to detect, as they often bypass traditional unimodal detection methods. To address this problem, we proposed DDCAF (Dynamic Dual Cross-Attention Fusion), a novel multimodal framework for detecting hate speech that integrates profound semantic comprehension from both visual and textual modalities. It uses a dual-stream architecture consisting of a RoBERTa-based text encoder and a Vision Transformer (ViT)-based image encoder. Using a bidirectional cross-attention mechanism, the model dynamically computes text-guided visual attention and visual-guided text attention, enabling it to prioritize semantically aligned features across modalities. The dynamic attention-driven fusion mechanism is able to identify subtle, context-dependent cues of hate intent.The proposed framework is primarily tested on multimodal benchmark datasets such as Hateful Memes and MMHS150K, while also assessing unimodal baselines (HateEval and OLID). The experimental findings reveal that it surpasses existing approaches and provides an accuracy of 89.35% on Hateful Memes and 91.20% on MMHS150K. Furthermore, ablation studies are carried out to demonstrate the impact of the subcomponent within the DDCAF, underscoring the importance of dynamic adaptive gating in capturing intermodal dependencies.","manuscriptTitle":"DDCAF: Dynamic Dual Cross-Attention Fusion Framework for Multimodal Hate Speech Detection","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-15 15:49:02","doi":"10.21203/rs.3.rs-7506109/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-11-02T02:30:39+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-17T10:19:29+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-20T05:57:42+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"58090546401565360928863428176213182454","date":"2025-09-11T01:10:28+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-09T06:44:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"235338644548812777286984015258296324381","date":"2025-09-06T02:34:26+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"319122245859717385755367087369750857639","date":"2025-09-06T02:29:14+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-06T01:33:33+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-05T02:25:11+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-02T11:17:52+00:00","index":"","fulltext":""},{"type":"submitted","content":"International Journal of Machine Learning and Cybernetics","date":"2025-09-01T08:24:08+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":"e49777a0-e195-4929-a413-6d3c2d074b25","owner":[],"postedDate":"September 15th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-15T01:53:31+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-15 15:49:02","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7506109","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7506109","identity":"rs-7506109","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}