Multimodal Conflict-Aware and Generative-Enhanced AI for Early Startup Survival and Risk Prediction

Multimodal Conflict-Aware and Generative-Enhanced AI for Early Startup Survival and Risk Prediction | 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 Article Multimodal Conflict-Aware and Generative-Enhanced AI for Early Startup Survival and Risk Prediction Jiaying Xi This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8365925/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Early-stage startups are central to innovation-driven economies, yet their failure rates remain persistently high, with more than half of new ventures not surviving their first three to five years. Accurately assessing the risk of young ventures is challenging because relevant signals are dispersed across heterogeneous sources, including entrepreneurial narratives, early financial indicators, and founders' positions in entrepreneurial networks. Most existing models either focus on a single modality or treat multimodal features as independent inputs, overlooking the informative role of cross-modal inconsistencies and struggling with small, imbalanced datasets. This paper proposes a conflict-aware, generative-enhanced multimodal framework for early-stage startup risk assessment. The model jointly encodes business plan and interview text, structured financial features, and founder-centric social network information via transformer, multilayer perceptron, and heterogeneous graph neural network encoders. Cross-modal contrastive learning aligns the three modalities into a shared representation space, while a modal conflict attention module explicitly quantifies inconsistencies between textual claims, financial realities, and network signals as additional risk features. To mitigate data scarcity and class imbalance, we further introduce a conditional GAN operating in the fused latent space to generate label- and conflict-conditioned synthetic representations that expose the classifier to both typical and strategically conflictual patterns. Experiments on a real-world dataset of 4,500 early-stage startups show that the proposed framework outperforms strong structured, text-based, graph-based, and multimodal baselines. Compared to the best multimodal fusion baseline, our model improves AUROC from 0.82 to 0.87 and AUPRC from 0.53 to 0.62, with consistent gains in macro F$_1$ and balanced accuracy. Ablation studies confirm the incremental contributions of contrastive alignment, modal conflict attention, and generative augmentation, while analyses of conflict features and case studies illustrate how the framework offers interpretable, conflict-aware decision support for investors and policymakers. Humanities/Complex networks Social science/Complex networks Physical sciences/Mathematics and computing early-stage startups risk assessment multimodal deep learning cross-modal conflict graph neural networks generative adversarial networks Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 08 Mar, 2026 Reviewers agreed at journal 05 Mar, 2026 Reviewers invited by journal 25 Dec, 2025 Editor invited by journal 23 Dec, 2025 Editor assigned by journal 18 Dec, 2025 Submission checks completed at journal 18 Dec, 2025 First submitted to journal 15 Dec, 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-8365925","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":565604072,"identity":"a531e0e3-f661-439a-b2b3-414b50e6adac","order_by":0,"name":"Jiaying Xi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAz0lEQVRIiWNgGAWjYFCCA2BShoG9sfHhBxK0GPAw8BxuNpYgwSqgFon0NgEeotQePGP44cefPzz8kg/bGCQY7OR0GwhokWw4YyzZ22bAIzk7se1BAUOysdkBAlr4Gc6YMfA2GPAY3E5sN5BgOJC4jZAWNqAWxj9/DHjsbx5sk+AhRgvIFmYeNqAtEoxEapFsOFYsLdtmzCNxJhEYyAZE+MXgxuGNH9/8kZPjbz/+8OGHCjs5gloYJE4YIJtASDkI8Lc/IEbZKBgFo2AUjGQAAExrQFvVVbSUAAAAAElFTkSuQmCC","orcid":"","institution":"Zhejiang Technical Institute of Economics","correspondingAuthor":true,"prefix":"","firstName":"Jiaying","middleName":"","lastName":"Xi","suffix":""}],"badges":[],"createdAt":"2025-12-15 12:08:36","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8365925/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8365925/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":99128422,"identity":"a62873c4-23e0-4113-918d-a88ec89c48f2","added_by":"auto","created_at":"2025-12-29 03:43:49","extension":"json","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4400,"visible":true,"origin":"","legend":"","description":"","filename":"d26d40e28a274c1499a2de4d4704f7a2.json","url":"https://assets-eu.researchsquare.com/files/rs-8365925/v1/cbab2c994b815d4e9dd6400d.json"},{"id":99315521,"identity":"52b7b034-47c4-422b-8c8e-1842d719ac1a","added_by":"auto","created_at":"2025-12-31 16:27:02","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":714394,"visible":true,"origin":"","legend":"","description":"","filename":"SRforEconomy.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8365925/v1_covered_9daa4053-5afd-4b71-b002-e547a6457a58.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Multimodal Conflict-Aware and Generative-Enhanced AI for Early Startup Survival and Risk Prediction","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"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":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"early-stage startups, risk assessment, multimodal deep learning, cross-modal conflict, graph neural networks, generative adversarial networks","lastPublishedDoi":"10.21203/rs.3.rs-8365925/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8365925/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eEarly-stage startups are central to innovation-driven economies, yet their failure rates remain persistently high, with more than half of new ventures not surviving their first three to five years.\u0026nbsp;Accurately assessing the risk of young ventures is challenging because relevant signals are dispersed across heterogeneous sources, including entrepreneurial narratives, early financial indicators, and founders' positions in entrepreneurial networks.\u0026nbsp;Most existing models either focus on a single modality or treat multimodal features as independent inputs, overlooking the informative role of cross-modal inconsistencies and struggling with small, imbalanced datasets.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThis paper proposes a conflict-aware, generative-enhanced multimodal framework for early-stage startup risk assessment.\u0026nbsp;The model jointly encodes business plan and interview text, structured financial features, and founder-centric social network information via transformer, multilayer perceptron, and heterogeneous graph neural network encoders.\u0026nbsp;Cross-modal contrastive learning aligns the three modalities into a shared representation space, while a modal conflict attention module explicitly quantifies inconsistencies between textual claims, financial realities, and network signals as additional risk features.\u0026nbsp;To mitigate data scarcity and class imbalance, we further introduce a conditional GAN operating in the fused latent space to generate label- and conflict-conditioned synthetic representations that expose the classifier to both typical and strategically conflictual patterns. Experiments on a real-world dataset of 4,500 early-stage startups show that the proposed framework outperforms strong structured, text-based, graph-based, and multimodal baselines.\u0026nbsp;Compared to the best multimodal fusion baseline, our model improves AUROC from 0.82 to 0.87 and AUPRC from 0.53 to 0.62, with consistent gains in macro F$_1$ and balanced accuracy.\u0026nbsp;Ablation studies confirm the incremental contributions of contrastive alignment, modal conflict attention, and generative augmentation, while analyses of conflict features and case studies illustrate how the framework offers interpretable, conflict-aware decision support for investors and policymakers.\u003c/p\u003e","manuscriptTitle":"Multimodal Conflict-Aware and Generative-Enhanced AI for Early Startup Survival and Risk Prediction","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-29 03:43:45","doi":"10.21203/rs.3.rs-8365925/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-03-09T03:26:38+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"194680083700778081360754575355647109977","date":"2026-03-05T10:43:07+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-12-25T14:30:48+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-12-23T08:50:33+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-12-18T10:05:13+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-12-18T10:01:22+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-12-15T12:03:23+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"ce255e23-7eb9-49a4-aa8f-ce556c508878","owner":[],"postedDate":"December 29th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":60217240,"name":"Humanities/Complex networks"},{"id":60217241,"name":"Social science/Complex networks"},{"id":60217242,"name":"Physical sciences/Mathematics and computing"}],"tags":[],"updatedAt":"2025-12-29T03:43:45+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-29 03:43:45","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8365925","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8365925","identity":"rs-8365925","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}