Improving Multimodal Reasoning in Large Language Models via Federated Example Selection

Improving Multimodal Reasoning in Large Language Models via Federated Example Selection | 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 Improving Multimodal Reasoning in Large Language Models via Federated Example Selection Beatrice Fawcett, Felix Ashworth, Harriet Dunbar This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4745818/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 Large language models have experienced a significant transformation with the advent of powerful models capable of generating, translating, and summarizing text. However, effectively integrating and reasoning with multimodal data, which includes text, images, and audio, continues to pose a formidable challenge. The novel concept of federated example selection, as proposed in this paper, is significant because it addresses critical issues of data privacy, diversity, and representation, thus enhancing multimodal reasoning capabilities in AI models. By modifying the open-source LLM, Llama, with a federated example selection algorithm, this research demonstrates substantial improvements in model performance across various tasks. The methodology involves a comprehensive approach that includes secure data collection, algorithm design, and iterative model refinement, all within a federated learning framework. The experimental results indicate notable advancements in accuracy, precision, recall, and F1-score, outperforming state-of-the-art models and showcasing the algorithm's efficacy. Additionally, scalability analysis confirms the algorithm's ability to handle large-scale datasets efficiently, while ablation studies highlight the importance of each algorithmic component. The findings suggest that the proposed methodology not only enhances the integration of diverse data types but also paves the way for future developments in multimodal AI, demonstrating the transformative potential of federated learning approaches. Artificial Intelligence and Machine Learning Multimodal reasoning Federated learning Example selection Data privacy AI integration 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-4745818","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":327381407,"identity":"b2dbda6a-714a-4fd8-87f6-3764c94837fe","order_by":0,"name":"Beatrice Fawcett","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABG0lEQVRIie3PMUvDQBTA8ZNAsxxkvYKkX+EdAXEozVe5EmiXBurm5gWHDEZc67e4KeIWKNol7gkdHAS7KDZ0yXCDl2TQIVdwc7j/EMLxftw9hEymf5jffkH9ZN2BC93JuZZQjk54Q7pJBh4gqyFESyBrCPohU9ESpCc0vt69XS4nLtg53Ve1nD/Yd0+iXhLkxDesdxf8TKMcAs/HC2+4YhA+JutBmaiHkfxF9N6yYjTikKn3hKmFFRFFMCiwIkDCXgKvu6ohV8L5TC3JYA6KlPIYKXB7C1MDqaXWZw3ZHruFJouLew4BFeTrMExmHhX5+mx7CgTrdvHtjai4nIzAmU339dgdwSZ6Lz/k2HXi216iD/9t3GQymUy/+wbgRmR2Bj4UXAAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0009-0008-3026-0651","institution":"","correspondingAuthor":true,"prefix":"","firstName":"Beatrice","middleName":"","lastName":"Fawcett","suffix":""},{"id":327381408,"identity":"7b93335c-c127-478e-9d1a-c076a42df9e1","order_by":1,"name":"Felix Ashworth","email":"","orcid":"https://orcid.org/0009-0000-6673-622X","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Felix","middleName":"","lastName":"Ashworth","suffix":""},{"id":327381409,"identity":"d07af27b-6400-43d7-9369-3d8bed944aba","order_by":2,"name":"Harriet Dunbar","email":"","orcid":"https://orcid.org/0009-0003-7014-6494","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Harriet","middleName":"","lastName":"Dunbar","suffix":""}],"badges":[],"createdAt":"2024-07-15 22:43:26","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-4745818/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4745818/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":60462150,"identity":"931e6ca6-4d9e-4ab6-8aed-7974d164b278","added_by":"auto","created_at":"2024-07-17 04:11:28","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":190845,"visible":true,"origin":"","legend":"","description":"","filename":"file.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4745818/v1_covered_6255a34c-39d9-4ed6-9e73-3e3e51c88baa.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eImproving Multimodal Reasoning in Large Language Models via Federated Example Selection\u003c/p\u003e","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 reasoning, Federated learning, Example selection, Data privacy, AI integration","lastPublishedDoi":"10.21203/rs.3.rs-4745818/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4745818/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eLarge language models have experienced a significant transformation with the advent of powerful models capable of generating, translating, and summarizing text. However, effectively integrating and reasoning with multimodal data, which includes text, images, and audio, continues to pose a formidable challenge. The novel concept of federated example selection, as proposed in this paper, is significant because it addresses critical issues of data privacy, diversity, and representation, thus enhancing multimodal reasoning capabilities in AI models. By modifying the open-source LLM, Llama, with a federated example selection algorithm, this research demonstrates substantial improvements in model performance across various tasks. The methodology involves a comprehensive approach that includes secure data collection, algorithm design, and iterative model refinement, all within a federated learning framework. The experimental results indicate notable advancements in accuracy, precision, recall, and F1-score, outperforming state-of-the-art models and showcasing the algorithm's efficacy. Additionally, scalability analysis confirms the algorithm's ability to handle large-scale datasets efficiently, while ablation studies highlight the importance of each algorithmic component. The findings suggest that the proposed methodology not only enhances the integration of diverse data types but also paves the way for future developments in multimodal AI, demonstrating the transformative potential of federated learning approaches.\u003c/p\u003e","manuscriptTitle":"Improving Multimodal Reasoning in Large Language Models via Federated Example Selection","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-07-17 04:03:22","doi":"10.21203/rs.3.rs-4745818/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":"0bccee65-da98-4114-98f6-251cc6684130","owner":[],"postedDate":"July 17th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":34646695,"name":"Artificial Intelligence and Machine Learning"}],"tags":[],"updatedAt":"2024-07-17T04:03:22+00:00","versionOfRecord":[],"versionCreatedAt":"2024-07-17 04:03:22","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4745818","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4745818","identity":"rs-4745818","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}