Non-equilibrium active noise enhances generative memory in diffusion models | 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 Non-equilibrium active noise enhances generative memory in diffusion models Suriyanarayanan Vaikuntanathan, Agnish Behera, Alexandra Lamtyugina, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8702780/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Generative diffusion models have emerged as powerful tools for sampling high-dimensional distributions, yet they typically rely on white gaussian noise and noise schedules to destroy and reconstruct information. Here, we demonstrate that driving the generative process out of equilibrium using active, temporally correlated noise sources fundamentally alters the information thermodynamics of the system. We show that coupling the data to an active non-Markovian bath creates a `memory effect' where high-level semantic information (such as class identity or molecular metastability) is stored in the temporal correlations of auxiliary degrees of freedom. Using Fisher information analysis, we prove that this active mechanism significantly retards the rate of information decay compared to passive Brownian motion. Crucially, this memory effect facilitates an earlier and more robust symmetry breaking (speciation) during the reverse generative process, allowing the system to resolve multi-scale structures, reminiscent of metastable states in molecular configurations that are washed out in the typical noising processes. Our results suggest that non-equilibrium protocols, inspired by active matter physics, offer a thermodynamically distinct and potentially advantageous pathway for recovering high-dimensional energy landscapes using generative diffusion. Physical sciences/Mathematics and computing/Computational science Physical sciences/Chemistry/Theoretical chemistry/Statistical mechanics Physical sciences/Physics/Information theory and computation Full Text Additional Declarations There is NO Competing Interest. Cite Share Download PDF Status: Under Review 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-8702780","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":585602753,"identity":"00ec7452-2da0-40a9-a70c-c381e1751a04","order_by":0,"name":"Suriyanarayanan Vaikuntanathan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA60lEQVRIiWNgGAWjYJCCgw1Agp+BgY0hASoiQYwWCckGUrQwgrQYHABqgQG8WuRnJD88OLPtTp3xjeRnDx7U2Nn1MzAfvM2DR4vBjTSDgxvbnkmY3UgzN0g4lpw8s4Et2RqvFp4DBgcfth0Gaslhk0hsYE42OMBjJo1Pi3zP8Q9gLcYzwFrqgVr4v+HVwnC8B+SwwxIGEmAth+2AtrDh1WJwvKfg4IxzhyVnnHlmJpFw7HiCZDObseUcfA5rZt/8safsMD9/e/IzyR811fb87M0Pb7zB5zB0AAwBUpSDgD2pGkbBKBgFo2D4AwCvz0/l8mBf6AAAAABJRU5ErkJggg==","orcid":"","institution":"University of Chicago","correspondingAuthor":true,"prefix":"","firstName":"Suriyanarayanan","middleName":"","lastName":"Vaikuntanathan","suffix":""},{"id":585602755,"identity":"1514b6e1-23dd-44f2-bfe7-1f25d9bbb97d","order_by":1,"name":"Agnish Behera","email":"","orcid":"","institution":"University of Chicago","correspondingAuthor":false,"prefix":"","firstName":"Agnish","middleName":"","lastName":"Behera","suffix":""},{"id":585602756,"identity":"6ee5d3fa-9aa8-4f22-ae97-66169caa410a","order_by":2,"name":"Alexandra Lamtyugina","email":"","orcid":"","institution":"University of Chicago","correspondingAuthor":false,"prefix":"","firstName":"Alexandra","middleName":"","lastName":"Lamtyugina","suffix":""},{"id":585602757,"identity":"41454f21-1fef-4207-a483-adf6a96f7162","order_by":3,"name":"Aditya Nandy","email":"","orcid":"https://orcid.org/0000-0001-7137-5449","institution":"University of California, Los Angeles","correspondingAuthor":false,"prefix":"","firstName":"Aditya","middleName":"","lastName":"Nandy","suffix":""},{"id":585602759,"identity":"5e327afa-3e8d-4136-9713-1232f470ae41","order_by":4,"name":"Daiki Goto","email":"","orcid":"","institution":"University of chicago","correspondingAuthor":false,"prefix":"","firstName":"Daiki","middleName":"","lastName":"Goto","suffix":""},{"id":585602760,"identity":"791d1757-1cf0-414e-853f-184e273ee1ce","order_by":5,"name":"Carlos Floyd","email":"","orcid":"https://orcid.org/0000-0002-6270-7250","institution":"University of Chicago","correspondingAuthor":false,"prefix":"","firstName":"Carlos","middleName":"","lastName":"Floyd","suffix":""}],"badges":[],"createdAt":"2026-01-26 17:29:50","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8702780/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8702780/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":101943454,"identity":"046f6433-6689-4df0-bce2-6659870a864e","added_by":"auto","created_at":"2026-02-05 09:42:01","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":6250548,"visible":true,"origin":"","legend":"Article File","description":"","filename":"Activediffusionarxiv26.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8702780/v1_covered_30668c6f-1655-4b24-b40d-c64106bf5295.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Non-equilibrium active noise enhances generative memory in diffusion models","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"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":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8702780/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8702780/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Generative diffusion models have emerged as powerful tools for sampling high-dimensional distributions, yet they typically rely on white gaussian noise and noise schedules to destroy and reconstruct information. Here, we demonstrate that driving the generative process out of equilibrium using active, temporally correlated noise sources fundamentally alters the information thermodynamics of the system. We show that coupling the data to an active non-Markovian bath creates a `memory effect' where high-level semantic information (such as class identity or molecular metastability) is stored in the temporal correlations of auxiliary degrees of freedom. Using Fisher information analysis, we prove that this active mechanism significantly retards the rate of information decay compared to passive Brownian motion. Crucially, this memory effect facilitates an earlier and more robust symmetry breaking (speciation) during the reverse generative process, allowing the system to resolve multi-scale structures, reminiscent of metastable states in molecular configurations that are washed out in the typical noising processes. Our results suggest that non-equilibrium protocols, inspired by active matter physics, offer a thermodynamically distinct and potentially advantageous pathway for recovering high-dimensional energy landscapes using generative diffusion.","manuscriptTitle":"Non-equilibrium active noise enhances generative memory in diffusion models","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-05 04:29:07","doi":"10.21203/rs.3.rs-8702780/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"
[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"6c7e1594-fb46-4171-8f89-287dcdfed891","owner":[],"postedDate":"February 5th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":62300644,"name":"Physical sciences/Mathematics and computing/Computational science"},{"id":62300645,"name":"Physical sciences/Chemistry/Theoretical chemistry/Statistical mechanics"},{"id":62300646,"name":"Physical sciences/Physics/Information theory and computation"}],"tags":[],"updatedAt":"2026-03-24T09:05:12+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-05 04:29:07","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8702780","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8702780","identity":"rs-8702780","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}
Text is read by the "Ask this paper" AI Q&A widget below.
Extraction quality varies by source — PMC NXML preserves structure
cleanly, OA-HTML may include some navigation residue, and OA-PDF can
have broken hyphenation. The publisher copy
(via DOI)
is the canonical version.