Hierarchical Multimodal Disinformation Detection System for Armed Conflict: Integration of Forensic and Geopolitical Intelligence

Hierarchical Multimodal Disinformation Detection System for Armed Conflict: Integration of Forensic and Geopolitical Intelligence | 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 Hierarchical Multimodal Disinformation Detection System for Armed Conflict: Integration of Forensic and Geopolitical Intelligence Lakhan Singh, Debasish Pradhan This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9455363/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 This paper presents a specialized hierarchical multimodal system for detecting disinformation in armed conflict scenarios. The proposed two-layer fusion framework—combining geometric-mean forensic fusion and a validated decision-rule refinement (Layer1: 0.65, 0.55, 0.40; GeoScore ≥ 0.50)—achieves 100% accuracy with 0% false positives on a 30-video benchmark, ensuring zero false accusations. The system addresses a critical gap: general-purpose models experience a 22–44% degradation on military-domain media, whereas our domain-specific architecture is designed to keep degradation minimal. On the benchmark, the method achieves perfect separability, and extended dataset evaluation is planned to quantify real-world domain shift. The 8-component hierarchical fusion framework uniquely integrates: (1) Deepfake Detection Ensemble, (2) DistilBERT Linguistic Analysis with real-world corroboration, (3) LLM Ensemble Verification (Gemini 2.0 Flash and other LLMs, e.g., GPT-4, Claude), (4) AASIST Audio Forensics, (5) CLIP+CLAP Audio-Video Mismatch Detection, (6) Social Media Disinformation Scoring with geopolitical penalties, (7) novel 93-feature Geopolitical Intelligence Module (GeoScore), and (8) Cross-Modal Consistency Verification. Key innovation: Hierarchical two-layer fusion: Layer 1 applies parameter-free geometric mean on six forensic modules; Layer 2 refines via decision tree incorporating audio-video mismatch and geopolitical context. Feature importance analysis reveals forensic detection contributes 71.6%, AV-mismatch 15.2%, and GeoScore 13.3%. Performance on 30-video benchmark: 100% Precision, 100% Recall, 100% Accuracy, F1-Score 1.0. Comparative evaluation across three hierarchical approaches validates optimal architecture design. Multimodal Disinformation Detection Armed Conflicts Deepfake Detection Geopolitical Intelligence and Trust Score Hierarchical Fusion Decision Tree LLM-based Verification Domain-Specific AI Full Text Additional Declarations No competing interests reported. Supplementary Files FinalThesisReportbyLakhanSingh.zip 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-9455363","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":626448487,"identity":"86e3b76e-c913-48a7-b88c-be12776a875d","order_by":0,"name":"Lakhan Singh","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA20lEQVRIiWNgGAWjYBACPgY2NhDNY9/eAKQMLAhrYYNqkTPgOQDSIkG8FmMDiQQQTYwW9mNpD37uqEvcLvn86oYfBRIM/O3dCfi18KQdN+w9czhx5+ycsps9QIdJnDm7gYDD0tskeNsOJDbczkm7wQPUYiCRS0AL//M2yb9tdYkNN8+k3fxDlBaJtGPSvG3MxgY32I/dJs4WiWdp0rJth+Uke3LYbssYSPAQ9As/f5qZ5Nu2Oh5+9uPPbr75YyPH396LXwsS4DEAk8QqBwH2B6SoHgWjYBSMghEEAOXLQ//rlC9uAAAAAElFTkSuQmCC","orcid":"","institution":"Defence Institute of Advanced Technology","correspondingAuthor":true,"prefix":"","firstName":"Lakhan","middleName":"","lastName":"Singh","suffix":""},{"id":626448488,"identity":"4ce7e212-25ce-4297-a553-5e1b1c69d267","order_by":1,"name":"Debasish Pradhan","email":"","orcid":"","institution":"Defence Institute of Advanced Technology","correspondingAuthor":false,"prefix":"","firstName":"Debasish","middleName":"","lastName":"Pradhan","suffix":""}],"badges":[],"createdAt":"2026-04-18 08:08:15","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9455363/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9455363/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108215574,"identity":"7c4bbe94-c142-442b-81cd-005e03c4d040","added_by":"auto","created_at":"2026-04-30 14:25:58","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":408048,"visible":true,"origin":"","legend":"","description":"","filename":"MultimodalDisinformationDetectionSystemforArmedConflictsAComprehensiveFrameworkwithGeopoliticalIntelligenceIntegrationfundingupdated.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9455363/v1_covered_b2315812-345e-4b3f-839d-fabbabba85a0.pdf"},{"id":107489722,"identity":"d81b3339-fcd2-4839-9d3f-65f34695ecf3","added_by":"auto","created_at":"2026-04-22 02:48:49","extension":"zip","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":12368633,"visible":true,"origin":"","legend":"","description":"","filename":"FinalThesisReportbyLakhanSingh.zip","url":"https://assets-eu.researchsquare.com/files/rs-9455363/v1/478e7f680cdbfd67d8ce1d38.zip"}],"financialInterests":"No competing interests reported.","formattedTitle":"Hierarchical Multimodal Disinformation Detection System for Armed Conflict: Integration of Forensic and Geopolitical Intelligence","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":"Multimodal Disinformation Detection, Armed Conflicts, Deepfake Detection, Geopolitical Intelligence and Trust Score, Hierarchical Fusion, Decision Tree, LLM-based Verification, Domain-Specific AI","lastPublishedDoi":"10.21203/rs.3.rs-9455363/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9455363/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis paper presents a specialized hierarchical multimodal system for detecting disinformation in armed conflict scenarios. The proposed two-layer fusion framework—combining geometric-mean forensic fusion and a validated decision-rule refinement (Layer1: 0.65, 0.55, 0.40; GeoScore ≥ 0.50)—achieves 100% accuracy with 0% false positives on a 30-video benchmark, ensuring zero false accusations. The system addresses a critical gap: general-purpose models experience a 22–44% degradation on military-domain media, whereas our domain-specific architecture is designed to keep degradation minimal. On the benchmark, the method achieves perfect separability, and extended dataset evaluation is planned to quantify real-world domain shift. The 8-component hierarchical fusion framework uniquely integrates: (1) Deepfake Detection Ensemble, (2) DistilBERT Linguistic Analysis with real-world corroboration, (3) LLM Ensemble Verification (Gemini 2.0 Flash and other LLMs, e.g., GPT-4, Claude), (4) AASIST Audio Forensics, (5) CLIP+CLAP Audio-Video Mismatch Detection, (6) Social Media Disinformation Scoring with geopolitical penalties, (7) novel 93-feature Geopolitical Intelligence Module (GeoScore), and (8) Cross-Modal Consistency Verification. Key innovation: Hierarchical two-layer fusion: Layer 1 applies parameter-free geometric mean on six forensic modules; Layer 2 refines via decision tree incorporating audio-video mismatch and geopolitical context. Feature importance analysis reveals forensic detection contributes 71.6%, AV-mismatch 15.2%, and GeoScore 13.3%. Performance on 30-video benchmark: 100% Precision, 100% Recall, 100% Accuracy, F1-Score 1.0. Comparative evaluation across three hierarchical approaches validates optimal architecture design.\u003c/p\u003e","manuscriptTitle":"Hierarchical Multimodal Disinformation Detection System for Armed Conflict: Integration of Forensic and Geopolitical Intelligence","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-21 11:42:12","doi":"10.21203/rs.3.rs-9455363/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":"0be59985-5887-40c1-a9c2-ef9f0d77146b","owner":[],"postedDate":"April 21st, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Rejected","date":"2026-04-30T14:11:32+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-29T15:54:43+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-30T14:25:18+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-21 11:42:12","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9455363","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9455363","identity":"rs-9455363","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}