HALO-GNN: Hallucination-Resistant Temporal Graph Neural Networks for Dynamic Community Detection

HALO-GNN: Hallucination-Resistant Temporal Graph Neural Networks for Dynamic Community 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 Article HALO-GNN: Hallucination-Resistant Temporal Graph Neural Networks for Dynamic Community Detection Yanfei Ma, Daozheng Qu, Yibo Wang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9255695/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Temporal graph learning is inherently subject to illusory structural dynamics, which is a phenomenon in which transient noise and ephemeral perturbations are exaggerated into deceptive community movements. Consequently, this leads to community assignments that are unstable and inconsistent, which severely limits the effectiveness of temporal graph neural networks in situations that involve streaming and dynamic circumstances. We provide a learning paradigm that is resistant to hallucinations and stabilizes temporal representations through the use of memory-guided structural regularization. The framework that has been suggested stabilizes node embeddings by referencing historical structures. This helps to reduce oscillations that are brought on by high-frequency noise while also preserving low-frequency development that is compatible with the community. Thorough evaluations across a variety of temporal graph benchmarks demonstrate significant improvements in robustness, temporal consistency, and perturbation resilience. These findings highlight the importance of hallucination resistance for the purpose of achieving reliable dynamic community detection. Physical sciences/Mathematics and computing Biological sciences/Neuroscience Physical sciences/Physics Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 09 May, 2026 Reviewers agreed at journal 08 May, 2026 Reviews received at journal 19 Apr, 2026 Reviewers agreed at journal 07 Apr, 2026 Reviewers invited by journal 07 Apr, 2026 Editor invited by journal 02 Apr, 2026 Editor assigned by journal 30 Mar, 2026 Submission checks completed at journal 30 Mar, 2026 First submitted to journal 28 Mar, 2026 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-9255695","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":620408079,"identity":"59253406-d804-48b5-8bea-2e3fb40814fb","order_by":0,"name":"Yanfei Ma","email":"","orcid":"","institution":"Fairleigh Dickinson University","correspondingAuthor":false,"prefix":"","firstName":"Yanfei","middleName":"","lastName":"Ma","suffix":""},{"id":620408080,"identity":"6a8f48f6-4e0a-4b4f-8807-babf49170ce1","order_by":1,"name":"Daozheng Qu","email":"","orcid":"","institution":"Fairleigh Dickinson University","correspondingAuthor":false,"prefix":"","firstName":"Daozheng","middleName":"","lastName":"Qu","suffix":""},{"id":620408081,"identity":"a772f087-7141-42a8-bdef-3fd9cf5db6e0","order_by":2,"name":"Yibo Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyklEQVRIiWNgGAWjYNACgwMM/Mwka5FsJtGaAyBdxJp//Ozh1zwFd+SMjzMffPCBwSZf3oGQljN5adY8Bs+MzQ6zJRvOYEiz3EjIOrMDOWbGPAaHE7cd5jGT5mE4bGDYQEjL+TcQLZub+b///kOUlhs5xo9BWjYw87ABw/mwgTwBHQz2N96YMc4B+kXiMJuxZI9BmoEBIS2S/TnGH978uSPH33/44YcfFTYG8oQcBgRsUjxwtgEDURHE/PEHMpcYW0bBKBgFo2BkAQDpokDR+nWI2AAAAABJRU5ErkJggg==","orcid":"","institution":"University of California, San Diego","correspondingAuthor":true,"prefix":"","firstName":"Yibo","middleName":"","lastName":"Wang","suffix":""}],"badges":[],"createdAt":"2026-03-29 00:53:14","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9255695/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9255695/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106840249,"identity":"ccddeb1e-04b5-4d16-88b9-97592c98b7db","added_by":"auto","created_at":"2026-04-14 03:20:32","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1270946,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9255695/v1_covered_3f81c5ed-a6f8-4150-bade-23c67868860b.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eHALO-GNN: Hallucination-Resistant Temporal Graph Neural Networks for Dynamic Community Detection\u003c/p\u003e","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":"","lastPublishedDoi":"10.21203/rs.3.rs-9255695/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9255695/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Temporal graph learning is inherently subject to illusory structural dynamics, which is a phenomenon in which transient noise and ephemeral perturbations are exaggerated into deceptive community movements. Consequently, this leads to community assignments that are unstable and inconsistent, which severely limits the effectiveness of temporal graph neural networks in situations that involve streaming and dynamic circumstances. We provide a learning paradigm that is resistant to hallucinations and stabilizes temporal representations through the use of memory-guided structural regularization. The framework that has been suggested stabilizes node embeddings by referencing historical structures. This helps to reduce oscillations that are brought on by high-frequency noise while also preserving low-frequency development that is compatible with the community. Thorough evaluations across a variety of temporal graph benchmarks demonstrate significant improvements in robustness, temporal consistency, and perturbation resilience. These findings highlight the importance of hallucination resistance for the purpose of achieving reliable dynamic community detection.","manuscriptTitle":"HALO-GNN: Hallucination-Resistant Temporal Graph Neural Networks for Dynamic Community Detection","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-14 03:20:14","doi":"10.21203/rs.3.rs-9255695/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"13660908712790792027688130748249682427","date":"2026-05-09T10:16:54+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"274490097266781117072163583941823942305","date":"2026-05-09T00:46:05+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-20T01:12:59+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"48714548252800722080992138513863439574","date":"2026-04-08T03:49:53+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-07T12:49:22+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-04-03T02:47:34+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-30T04:58:48+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-30T04:58:02+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2026-03-29T00:44:12+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":"b3b62e07-ab5f-4af0-859d-b1415dd90079","owner":[],"postedDate":"April 14th, 2026","published":true,"recentEditorialEvents":[{"type":"reviewerAgreed","content":"13660908712790792027688130748249682427","date":"2026-05-09T10:16:54+00:00","index":54,"fulltext":""},{"type":"reviewerAgreed","content":"274490097266781117072163583941823942305","date":"2026-05-09T00:46:05+00:00","index":53,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":66258842,"name":"Physical sciences/Mathematics and computing"},{"id":66258843,"name":"Biological sciences/Neuroscience"},{"id":66258844,"name":"Physical sciences/Physics"}],"tags":[],"updatedAt":"2026-04-14T03:20:15+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-14 03:20:14","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9255695","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9255695","identity":"rs-9255695","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}