Temporal Modeling with Reversible Transformers

Temporal Modeling with Reversible Transformers | 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Download PDF Research Article Temporal Modeling with Reversible Transformers Leonid Kulyk This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6293520/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 Memory efficiency is a critical bottleneck in deep learning models for sequence processing, particularly in long-range dependencies and continuous data streams. In the effort to solve this, we introduce a new architecture denoted as Reversible Temporal Transformer (TempVerseFormer). TempVerseFormer integrates reversible transformer blocks uniquely with a time-agnostic backpropagation strategy that decouples the memory footprint from the temporal depth and enables efficient training on long prediction time ranges. We have tested the model on a procedurally generated dataset involving the rotation of 2D shapes and show that the predictive accuracy of TempVerseFormer is competitive compared to other tested baselines, with memory consumption being practically independent of the time-to-predict. This substantial gain in memory efficiency, achieved in a controlled synthetic environment while not dropping performance on our dataset, places the TempVerseFormer as an indicative candidate for scalable temporal sequence modeling, allowing for real-time adaptation or video analysis at edge devices, and thereby leading to more fiscally responsible and temporally wise resource AI systems that are capable of working with changing and evolving environments. Artificial intelligence Deep learning Memory efficiency Reversible architectures Transformer Backpropagation Full Text Additional Declarations No competing interests reported. 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-6293520","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":433088772,"identity":"c7263925-5751-4ec7-ab0e-4d0de6e97186","order_by":0,"name":"Leonid Kulyk","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABDElEQVRIie3Pv0oDMRzA8Uggt4R2TSlenkDIcVC3PobzLxxclis4lRsjhaw+wA2+gkVwTincVPoCLieuDjcKCno5kA6Sq24F8x0yhN8nfxAKhU6wkVuAABphcvbyIT5jt2GbAUK+CcEEC1ritCdwjCBHECGM7rDU/SFDJFrVTbPMpYkIsIkh6u5q+9zdMo8vtIfQWgnYF9JgYkVi6GL9lIuOZOnMeggrZkyasiORBmnYYl2BI1Y++gh/PRC7MUIllWqHCaOO9A/b3OgdAJ8WR26hec5gn6cdyTAqbXI/La4tCP9fxtG2nrwts/PbcX35joTlvFIPbVvOYx/5megnxW/HXVz/ZToUCoX+Q1/vhFvQRyH3swAAAABJRU5ErkJggg==","orcid":"","institution":"Vinnytsia National Technical University","correspondingAuthor":true,"prefix":"","firstName":"Leonid","middleName":"","lastName":"Kulyk","suffix":""}],"badges":[],"createdAt":"2025-03-24 09:08:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6293520/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6293520/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":79479832,"identity":"f524a61a-ba3e-4276-b49f-2d8eccd03c17","added_by":"auto","created_at":"2025-03-29 05:32:38","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2137474,"visible":true,"origin":"","legend":"","description":"","filename":"TemporalModelingwithReversibleTransformers.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6293520/v1_covered_7fa47f64-c452-46f2-894b-14f6666c2d6d.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Temporal Modeling with Reversible Transformers","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":"Artificial intelligence, Deep learning, Memory efficiency, Reversible architectures, Transformer, Backpropagation","lastPublishedDoi":"10.21203/rs.3.rs-6293520/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6293520/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMemory efficiency is a critical bottleneck in deep learning models for sequence processing, particularly in long-range dependencies and continuous data streams. In the effort to solve this, we introduce a new architecture denoted as Reversible Temporal Transformer (TempVerseFormer). TempVerseFormer integrates reversible transformer blocks uniquely with a time-agnostic backpropagation strategy that decouples the memory footprint from the temporal depth and enables efficient training on long prediction time ranges. We have tested the model on a procedurally generated dataset involving the rotation of 2D shapes and show that the predictive accuracy of TempVerseFormer is competitive compared to other tested baselines, with memory consumption being practically independent of the time-to-predict. This substantial gain in memory efficiency, achieved in a controlled synthetic environment while not dropping performance on our dataset, places the TempVerseFormer as an indicative candidate for scalable temporal sequence modeling, allowing for real-time adaptation or video analysis at edge devices, and thereby leading to more fiscally responsible and temporally wise resource AI systems that are capable of working with changing and evolving environments.\u003c/p\u003e","manuscriptTitle":"Temporal Modeling with Reversible Transformers","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-25 02:55:11","doi":"10.21203/rs.3.rs-6293520/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":"d0f1ddf2-0461-4bd6-b088-ac8ae2fde14b","owner":[],"postedDate":"March 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-03-29T05:08:27+00:00","versionOfRecord":[],"versionCreatedAt":"2025-03-25 02:55:11","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6293520","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6293520","identity":"rs-6293520","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}