Signed-Weight Projection for Robust In-Sensor Computing Neural Classifiers under Amplitude-Dependent Noise and Stuck-at Faults

Signed-Weight Projection for Robust In-Sensor Computing Neural Classifiers under Amplitude-Dependent Noise and Stuck-at Faults | 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 Signed-Weight Projection for Robust In-Sensor Computing Neural Classifiers under Amplitude-Dependent Noise and Stuck-at Faults Difei Zhong This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7366045/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract In-sensor computing (ISC) monolithically integrates photodetection and neural-network inference on the same focal plane, avoiding analog-to-digital conversion and off-chip data movement. Despite its unrivalled latency and energy benefits, ISC remains limited by the fragility of analog weights. Amplitude-dependent drift, random-telegraph noise, and irreversible stuck-at faults can each trigger abrupt accuracy collapse. We introduce a training-free hardening strategy—signed-weight projection—that clips every weight to a symmetric range, enforcing a zero-mean distribution without changing network topology or requiring retraining. Firstorder perturbation analysis shows that the resulting balanced weights self-cancel the mean shift produced by multiplicative noise, while the magnitude cap limits the worst-case impact of saturated faults. Hardware measurements on a 10 × 10 ISC prototype confirm a full optical-toelectrical inference path in ∼ 2 µs, underscoring the need for on-array robustness. Device-level simulations on a 784−100−10 classifier further demonstrate graceful accuracy degradation. Noise tolerance and stuck-fault tolerance expand by well over an order of magnitude compared with an unprotected baseline, yet clean-data accuracy is preserved. The method offers a low-cost algorithm device co-design guideline, which can be retro-fitted to existing ISC pipelines and scaled to future high-resolution vision sensors. in-sensor computing analog neural networks robustness signed-weight projection stuck-at faults Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 22 Sep, 2025 Reviews received at journal 03 Sep, 2025 Reviews received at journal 27 Aug, 2025 Reviewers agreed at journal 26 Aug, 2025 Reviewers agreed at journal 23 Aug, 2025 Reviewers invited by journal 20 Aug, 2025 Editor invited by journal 16 Aug, 2025 Editor assigned by journal 15 Aug, 2025 Submission checks completed at journal 15 Aug, 2025 First submitted to journal 13 Aug, 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-7366045","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":506258291,"identity":"25daa333-7905-4260-92b7-cbfff16e35b3","order_by":0,"name":"Difei Zhong","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABDElEQVRIie3QMUsDMRTA8RcOziVH11ewFb9BShanfpeQVVRwuUkChXQpdq0I+hVOCs6RB9elH0GwcuBc0eGQG8wdIjhcWzfB/HlDAvlBeACh0J9MrWAtsN/xR9fM9rRgs/RIdu1vSMSXqcry5rYDORhrgMSimj+Nn+ns41FNTZS/vsPwpI2wSQHQtSjvcy7o6vJFzVysr/dBn7eRCE8dDCz2PAFKJqSM4zJCcMq0kBj9x5RFNrd7q4bcbiO8Jm6Jh1kMgnhJKvOErTcQ5AU8mBQl5seCEkPyjmIdgdCtxG+MFZW46HdGi+KNV9S7WYyIlemwldS57xOz9Ub8cLHh/c+qL1ruLEKhUOgf9AkcGl07yVk3PwAAAABJRU5ErkJggg==","orcid":"","institution":"Tsinghua University","correspondingAuthor":true,"prefix":"","firstName":"Difei","middleName":"","lastName":"Zhong","suffix":""}],"badges":[],"createdAt":"2025-08-13 14:38:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7366045/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7366045/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":90067469,"identity":"4af943c2-b8c7-4027-9a5d-4a7a1b5c325c","added_by":"auto","created_at":"2025-08-28 05:53:31","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1691606,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7366045/v1_covered_e148c726-e5d8-4f16-b7ac-ef63e080b0a2.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Signed-Weight Projection for Robust In-Sensor Computing Neural Classifiers under Amplitude-Dependent Noise and Stuck-at Faults","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":"discover-electronics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Discover Electronics](https://www.springer.com/journal/44291)","snPcode":"44291","submissionUrl":"https://submission.nature.com/new-submission/44291","title":"Discover Electronics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"in-sensor computing, analog neural networks, robustness, signed-weight projection, stuck-at faults","lastPublishedDoi":"10.21203/rs.3.rs-7366045/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7366045/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn-sensor computing (ISC) monolithically integrates photodetection and neural-network inference on the same focal plane, avoiding analog-to-digital conversion and off-chip data movement. Despite its unrivalled latency and energy benefits, ISC remains limited by the fragility of analog weights. Amplitude-dependent drift, random-telegraph noise, and irreversible stuck-at faults can each trigger abrupt accuracy collapse. We introduce a training-free hardening strategy—signed-weight projection—that clips every weight to a symmetric range, enforcing a zero-mean distribution without changing network topology or requiring retraining. Firstorder perturbation analysis shows that the resulting balanced weights self-cancel the mean shift produced by multiplicative noise, while the magnitude cap limits the worst-case impact of saturated faults. Hardware measurements on a 10 × 10 ISC prototype confirm a full optical-toelectrical inference path in ∼ 2 µs, underscoring the need for on-array robustness. Device-level simulations on a 784−100−10 classifier further demonstrate graceful accuracy degradation. Noise tolerance and stuck-fault tolerance expand by well over an order of magnitude compared with an unprotected baseline, yet clean-data accuracy is preserved. The method offers a low-cost algorithm device co-design guideline, which can be retro-fitted to existing ISC pipelines and scaled to future high-resolution vision sensors.\u003c/p\u003e","manuscriptTitle":"Signed-Weight Projection for Robust In-Sensor Computing Neural Classifiers under Amplitude-Dependent Noise and Stuck-at Faults","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-28 05:37:21","doi":"10.21203/rs.3.rs-7366045/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-09-22T13:04:14+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-03T17:43:33+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-27T08:38:01+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"295313839155834459101003950038797597984","date":"2025-08-26T07:15:57+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"296924191787274641782171342711002506","date":"2025-08-23T12:44:21+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-08-20T08:28:10+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-08-16T07:26:48+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-08-16T01:35:04+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-08-16T01:33:59+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Electronics","date":"2025-08-13T14:30:19+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-electronics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [Discover Electronics](https://www.springer.com/journal/44291)","snPcode":"44291","submissionUrl":"https://submission.nature.com/new-submission/44291","title":"Discover Electronics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"96b13c51-6ae0-4246-93db-0a8ba8628c18","owner":[],"postedDate":"August 28th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2025-11-25T12:23:45+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-28 05:37:21","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7366045","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7366045","identity":"rs-7366045","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}