KN-VLM: KNowledge-guided Vision-and-Language Model for Visual Abductive Reasoning

KN-VLM: KNowledge-guided Vision-and-Language Model for Visual Abductive Reasoning | 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 KN-VLM: KNowledge-guided Vision-and-Language Model for Visual Abductive Reasoning Kuo Tan, Zhaobo Qi, Jianping Zhong, Yuanrong Xu, Weigang Zhang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4934011/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 09 Mar, 2025 Read the published version in Multimedia Systems → Version 1 posted 9 You are reading this latest preprint version Abstract Visual abductive reasoning strives to deduce the most suitable hypothesis that effectively explains the underlying visual context, garnering considerable attention in the academic community. However, recent efforts are inherently limited by their exclusive reliance on visual information, overlooking the invaluable com-monsense and the semantic/causal relationships, leading to inaccurate abductive reasoning outcomes. To tackle the above issue, we propose a simple but powerful KNowledge-guided Vision-and-Language Model (KN-VLM), which primarily consists of a visual reasoning branch and a knowledge reasoning branch. The visual reasoning branch utilizes a powerful visual embedding model followed by a visual-Qformer to capture visual features. The knowledge reasoning branch aims to acquire two complementary types of knowledge commonsense knowledge and complemented knowledge. The former aims to extract the intricate and detailed conceptual knowledge embedded within the observed video, which deepens the model’s comprehension of the presented video content. The latter utilizes the external knowledge base to further augment the understanding of the interconnec-tions and causal relationships among these concepts, thereby strengthening the model’s abductive reasoning capability. After that, the effective fusion of the two branches completes the abductive reasoning task, which generates descriptions for the observed and explanation events. Experimental results on the VAR and CookReasoning dataset show that our model achieves promising performance. Commonsense Knowledge Visual Language Model Visual Abductive Reasoning Dense Video Captioning Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 09 Mar, 2025 Read the published version in Multimedia Systems → Version 1 posted Editorial decision: Revision requested 04 Dec, 2024 Reviews received at journal 27 Nov, 2024 Reviews received at journal 22 Nov, 2024 Reviewers agreed at journal 22 Nov, 2024 Reviewers agreed at journal 21 Nov, 2024 Reviewers invited by journal 23 Aug, 2024 Editor assigned by journal 19 Aug, 2024 Submission checks completed at journal 19 Aug, 2024 First submitted to journal 18 Aug, 2024 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-4934011","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":347377677,"identity":"1d11c05d-7d5b-4c63-aebe-c6801413452d","order_by":0,"name":"Kuo Tan","email":"","orcid":"","institution":"Harbin Institute of Technology, Weihai","correspondingAuthor":false,"prefix":"","firstName":"Kuo","middleName":"","lastName":"Tan","suffix":""},{"id":347377680,"identity":"4e9999f9-a106-4d56-946a-20703d9f16cd","order_by":1,"name":"Zhaobo Qi","email":"","orcid":"","institution":"Harbin Institute of Technology, Weihai","correspondingAuthor":false,"prefix":"","firstName":"Zhaobo","middleName":"","lastName":"Qi","suffix":""},{"id":347377682,"identity":"fa23a18f-249f-4d61-9bba-e1d9f312f203","order_by":2,"name":"Jianping Zhong","email":"","orcid":"","institution":"Harbin Institute of Technology, Weihai","correspondingAuthor":false,"prefix":"","firstName":"Jianping","middleName":"","lastName":"Zhong","suffix":""},{"id":347377683,"identity":"c23f267e-ee0b-4a90-bff4-79f8df516148","order_by":3,"name":"Yuanrong Xu","email":"","orcid":"","institution":"Harbin Institute of Technology, Weihai","correspondingAuthor":false,"prefix":"","firstName":"Yuanrong","middleName":"","lastName":"Xu","suffix":""},{"id":347377684,"identity":"2bf971e0-5cbf-4c31-ada1-feb79484ce50","order_by":4,"name":"Weigang Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA00lEQVRIiWNgGAWjYHACNhAhZ8AMZxOpxdiAmZlELYkbGIjVYnB++bMHP3fUpm9n5z/A8KHsMAP/7AYCWm48SDfsPXM8d2czMwPjjHOHGSTuHMCvxezGgWMSvG3HcjccZmZg5m07zGAgkUBIy8E2yb9tx9INQFr+EqXlfDObNG9bTQJYCyMxWuxvsLFJy7YdMAQ6zOBgz7l0HokbBLRI9h9/Jvm2rU7e4PzBhw9+lFnL8c8goIUB4ozDYPYBIOYhoB4I+EHqGOoIKxwFo2AUjIKRCwCWtkRmgEH+RAAAAABJRU5ErkJggg==","orcid":"","institution":"Harbin Institute of Technology, Weihai","correspondingAuthor":true,"prefix":"","firstName":"Weigang","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2024-08-18 15:25:09","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4934011/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4934011/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s00530-025-01683-y","type":"published","date":"2025-03-09T15:57:08+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":78181315,"identity":"dcfe9ea2-1676-401f-9d3a-28f6ed9cb788","added_by":"auto","created_at":"2025-03-10 17:45:59","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1018342,"visible":true,"origin":"","legend":"","description":"","filename":"KNVLMmainpaper.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4934011/v1_covered_5a35b2f3-b211-46ae-97ee-3d26fc2b944a.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eKN-VLM: KNowledge-guided Vision-and-Language Model for Visual Abductive Reasoning\u003c/p\u003e","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":"multimedia-systems","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"mmsj","sideBox":"Learn more about [Multimedia Systems](http://link.springer.com/journal/530)","snPcode":"530","submissionUrl":"https://submission.nature.com/new-submission/530/3","title":"Multimedia Systems","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Commonsense Knowledge, Visual Language Model, Visual Abductive Reasoning, Dense Video Captioning","lastPublishedDoi":"10.21203/rs.3.rs-4934011/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4934011/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Visual abductive reasoning strives to deduce the most suitable hypothesis that effectively explains the underlying visual context, garnering considerable attention in the academic community. However, recent efforts are inherently limited by their exclusive reliance on visual information, overlooking the invaluable com-monsense and the semantic/causal relationships, leading to inaccurate abductive reasoning outcomes. To tackle the above issue, we propose a simple but powerful KNowledge-guided Vision-and-Language Model (KN-VLM), which primarily consists of a visual reasoning branch and a knowledge reasoning branch. The visual reasoning branch utilizes a powerful visual embedding model followed by a visual-Qformer to capture visual features. The knowledge reasoning branch aims to acquire two complementary types of knowledge commonsense knowledge and complemented knowledge. The former aims to extract the intricate and detailed conceptual knowledge embedded within the observed video, which deepens the model’s comprehension of the presented video content. The latter utilizes the external knowledge base to further augment the understanding of the interconnec-tions and causal relationships among these concepts, thereby strengthening the model’s abductive reasoning capability. After that, the effective fusion of the two branches completes the abductive reasoning task, which generates descriptions for the observed and explanation events. Experimental results on the VAR and CookReasoning dataset show that our model achieves promising performance.","manuscriptTitle":"KN-VLM: KNowledge-guided Vision-and-Language Model for Visual Abductive Reasoning","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-09-16 07:55:25","doi":"10.21203/rs.3.rs-4934011/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-12-05T04:12:02+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-11-27T10:30:05+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-11-22T11:32:24+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"282659612954795029759629304024814829123","date":"2024-11-22T06:40:02+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"61973830033056528707282161514799462534","date":"2024-11-22T03:38:43+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-08-23T08:53:41+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-08-19T09:10:44+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-08-19T08:10:26+00:00","index":"","fulltext":""},{"type":"submitted","content":"Multimedia Systems","date":"2024-08-18T15:23:47+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"multimedia-systems","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"mmsj","sideBox":"Learn more about [Multimedia Systems](http://link.springer.com/journal/530)","snPcode":"530","submissionUrl":"https://submission.nature.com/new-submission/530/3","title":"Multimedia Systems","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"207b6248-585e-41cb-8e26-f5ff5ec1e7e6","owner":[],"postedDate":"September 16th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-03-10T17:09:17+00:00","versionOfRecord":{"articleIdentity":"rs-4934011","link":"https://doi.org/10.1007/s00530-025-01683-y","journal":{"identity":"multimedia-systems","isVorOnly":false,"title":"Multimedia Systems"},"publishedOn":"2025-03-09 15:57:08","publishedOnDateReadable":"March 9th, 2025"},"versionCreatedAt":"2024-09-16 07:55:25","video":"","vorDoi":"10.1007/s00530-025-01683-y","vorDoiUrl":"https://doi.org/10.1007/s00530-025-01683-y","workflowStages":[]},"version":"v1","identity":"rs-4934011","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4934011","identity":"rs-4934011","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}