LUMEN-PRO: Automating Multi-Task Learning on Optical Neural Networks with Weight Sharing and Physical Rotation

LUMEN-PRO: Automating Multi-Task Learning on Optical Neural Networks with Weight Sharing and Physical Rotation | 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 LUMEN-PRO: Automating Multi-Task Learning on Optical Neural Networks with Weight Sharing and Physical Rotation Shanglin Zhou, Yingjie Li, Weilu Gao, Cunxi Yu, Caiwen Ding This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4733442/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 25 Apr, 2025 Read the published version in Scientific Reports → Version 1 posted 10 You are reading this latest preprint version Abstract The democratization of AI encourages multi-task learning (MTL), demanding more parameters and processing time. To achieve highly energy-efficient MTL, Diffractive Optical Neural Networks (DONNs) have garnered attention due to extremely low energy and high computation speed. However, implementing MTL on DONNs requires manually reconfiguring & replacing layers, and rebuilding & duplicating the physical optical systems. To overcome the challenges, we propose LUMEN-PRO, an automated MTL framework using DONNs. We first propose to automate MTL utilizing an arbitrary backbone DONN and a set of tasks, resulting in a high-accuracy multi-task DONN model with small memory footprint that surpasses existing MTL. Second, we leverage the rotability of the physical optical system and replace task-specific layers with rotation of the corresponding shared layers. This replacement eliminates the storage requirement of task-specific layers, further optimizing the memory footprint. LUMEN-PRO provides flexibility in identifying optimal sharing patterns across diverse datasets, facilitating the search for highly energy-efficient DONNs. Experiments show that LUMEN-PRO provides up to 49.58% higher accuracy and 4× better cost efficiency than single-task and existing DONN approaches. It achieves memory lower bound of MTL, with memory efficiency matching single-task models. Compared to IBM-TrueNorth and Nanophotonic, LUMEN-PRO achieves 597× and 68× energy efficiency gain, respectively. Physical sciences/Physics/Optical physics Physical sciences/Optics and photonics Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 25 Apr, 2025 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 06 Jan, 2025 Reviews received at journal 30 Nov, 2024 Reviewers agreed at journal 20 Nov, 2024 Reviews received at journal 25 Aug, 2024 Reviewers agreed at journal 08 Aug, 2024 Reviewers invited by journal 08 Aug, 2024 Editor assigned by journal 08 Aug, 2024 Editor invited by journal 25 Jul, 2024 Submission checks completed at journal 19 Jul, 2024 First submitted to journal 13 Jul, 2024 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-4733442","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":338717779,"identity":"09cbed54-a54d-42e6-bd45-1292ae5f7411","order_by":0,"name":"Shanglin Zhou","email":"","orcid":"","institution":"University of Connecticut","correspondingAuthor":false,"prefix":"","firstName":"Shanglin","middleName":"","lastName":"Zhou","suffix":""},{"id":338717780,"identity":"ee8ecc1a-aef9-46f9-82dd-753a8a8520c6","order_by":1,"name":"Yingjie Li","email":"","orcid":"","institution":"University of Maryland","correspondingAuthor":false,"prefix":"","firstName":"Yingjie","middleName":"","lastName":"Li","suffix":""},{"id":338717781,"identity":"be165b6c-7fa9-4782-bc00-53dd222f1e86","order_by":2,"name":"Weilu Gao","email":"","orcid":"","institution":"University of Utah","correspondingAuthor":false,"prefix":"","firstName":"Weilu","middleName":"","lastName":"Gao","suffix":""},{"id":338717782,"identity":"4c4d3a67-ddbc-4dea-97ba-8c074dc5c59e","order_by":3,"name":"Cunxi Yu","email":"","orcid":"","institution":"University of Maryland","correspondingAuthor":false,"prefix":"","firstName":"Cunxi","middleName":"","lastName":"Yu","suffix":""},{"id":338717783,"identity":"96516592-00ea-4da8-b147-b0c5fec9e3cc","order_by":4,"name":"Caiwen Ding","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAoUlEQVRIiWNgGAWjYDCCwyCiAsKWIEHLGZK0HABixjZStPAd5z384ue8w9HmDMwHb/MQo0XyMF+aZe+2w7k7G9iSrYnSYnCYx8yYEahlwwEeM2kStMwBaeH/RrQW48eMDWBb2IjTIgm0hbHnWHruhsNsxpZziNHCd/6M8YcfNda5G443P7zxhhgtQMAGiQ5mIpWD1X4gQfEoGAWjYBSMRAAAFaQyLJW1kD4AAAAASUVORK5CYII=","orcid":"","institution":"University of Minnesota Twin Cities","correspondingAuthor":true,"prefix":"","firstName":"Caiwen","middleName":"","lastName":"Ding","suffix":""}],"badges":[],"createdAt":"2024-07-13 04:53:24","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4733442/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4733442/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-97262-2","type":"published","date":"2025-04-25T15:57:38+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":81569640,"identity":"b613e99f-fbe4-4b02-bb83-7fc38884b7d3","added_by":"auto","created_at":"2025-04-28 16:08:53","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1653361,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4733442/v1_covered_287dbbdf-5733-4dbf-a68a-d61a2f2275b2.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"LUMEN-PRO: Automating Multi-Task Learning on Optical Neural Networks with Weight Sharing and Physical Rotation","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"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-4733442/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4733442/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The democratization of AI encourages multi-task learning (MTL), demanding more parameters and processing time. To achieve highly energy-efficient MTL, Diffractive Optical Neural Networks (DONNs) have garnered attention due to extremely low energy and high computation speed. However, implementing MTL on DONNs requires manually reconfiguring \u0026 replacing layers, and rebuilding \u0026 duplicating the physical optical systems. To overcome the challenges, we propose LUMEN-PRO, an automated MTL framework using DONNs. We first propose to automate MTL utilizing an arbitrary backbone DONN and a set of tasks, resulting in a high-accuracy multi-task DONN model with small memory footprint that surpasses existing MTL. Second, we leverage the rotability of the physical optical system and replace task-specific layers with rotation of the corresponding shared layers. This replacement eliminates the storage requirement of task-specific layers, further optimizing the memory footprint. LUMEN-PRO provides flexibility in identifying optimal sharing patterns across diverse datasets, facilitating the search for highly energy-efficient DONNs. Experiments show that LUMEN-PRO provides up to 49.58% higher accuracy and 4× better cost efficiency than single-task and existing DONN approaches. It achieves memory lower bound of MTL, with memory efficiency matching single-task models. Compared to IBM-TrueNorth and Nanophotonic, LUMEN-PRO achieves 597× and 68× energy efficiency gain, respectively.","manuscriptTitle":"LUMEN-PRO: Automating Multi-Task Learning on Optical Neural Networks with Weight Sharing and Physical Rotation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-08-13 02:55:42","doi":"10.21203/rs.3.rs-4733442/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-01-06T05:09:23+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-12-01T04:55:27+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"181591115497841668185439901137332874594","date":"2024-11-20T06:04:49+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-08-25T08:31:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"120346256900904699593028374283958735418","date":"2024-08-08T17:39:29+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-08-08T17:34:51+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-08-08T17:33:34+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2024-07-25T17:28:56+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-07-19T07:40:04+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2024-07-13T04:51:51+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":"cef8589a-b67b-4cbd-906d-245f301d6343","owner":[],"postedDate":"August 13th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":35880879,"name":"Physical sciences/Physics/Optical physics"},{"id":35880880,"name":"Physical sciences/Optics and photonics"}],"tags":[],"updatedAt":"2025-04-28T16:01:56+00:00","versionOfRecord":{"articleIdentity":"rs-4733442","link":"https://doi.org/10.1038/s41598-025-97262-2","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2025-04-25 15:57:38","publishedOnDateReadable":"April 25th, 2025"},"versionCreatedAt":"2024-08-13 02:55:42","video":"","vorDoi":"10.1038/s41598-025-97262-2","vorDoiUrl":"https://doi.org/10.1038/s41598-025-97262-2","workflowStages":[]},"version":"v1","identity":"rs-4733442","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4733442","identity":"rs-4733442","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}