Evaluating Lightweight Neural Models for Edge-Based Anomaly Detection: Performance and Efficiency Trade-offs

Evaluating Lightweight Neural Models for Edge-Based Anomaly Detection: Performance and Efficiency Trade-offs | 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 Evaluating Lightweight Neural Models for Edge-Based Anomaly Detection: Performance and Efficiency Trade-offs Isaac Kofi Nti, Lee Jo Ning, Clark Alex, Miriyala, Sai Manikanta, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7138288/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 In edge computing scenarios where constraints on memory, latency, and energy hinder the utilization of large-scale models, lightweight neural networks are increasingly favored for anomaly detection. However, uniform benchmarks for comparing commonly utilized lightweight models under these constraints remain absent. This research addresses the gap by evaluating three prominent lightweight neural architectures, pruned convolutional neural networks (CNNs), quantized long short-term memory networks (LSTMs), and distilled transformers, across two established intrusion detection datasets: CIC-IoT-DIAD 2024 and TON_IoT (TON_IoT_Modbus and TON_IoT_Thermostat). We evaluate each model utilizing standard detection measures (accuracy, precision, recall, and F1-score) and deployment metrics (model size, inference latency, and memory consumption) under simulated edge constraints. Our findings indicate significant trade-offs between model accuracy and efficiency, with performance varying based on the dataset utilized. Certain models perform more effectively with flow-based data compared to others with IoT telemetry. No single model excelled in all evaluation criteria. This research provides future edge-optimized anomaly detection studies with a reliable, reproducible foundation and valuable insights on selecting models for real-time edge deployment. The results also guide our attention in creating our forthcoming architecture, S3LiteNet, intended to enhance performance and deployment in information-centric networks. Artificial Intelligence and Machine Learning edge-based anomaly detection lightweight neural models edge-optimization resource constraints inference efficiency Full Text Additional Declarations The authors declare no competing interests. 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-7138288","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":486234482,"identity":"69a10864-adfd-40dd-9bf6-97f4a516ca0f","order_by":0,"name":"Isaac Kofi Nti","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAuklEQVRIiWNgGAWjYFACxgYgISEHYh54QJQONrAWC2OwlgTitIDJikSQRgaitJjLNzc+5vkjkT4/7PBDoC12croNBLRYtjE2G/O2SeRuvJ1mANSSbGx2gIAWg2OMbdK8DUAtsxNAWg4kbiNKC8hhhrPTP5CihU0iQV46h2hbEpsN57ZJGG6Qzik4kGBAjF8OH3/44M2fOnn52embP3yosJMjqAWhF6zSgFjlICDfQIrqUTAKRsEoGFEAALhfQyfNdO0xAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0001-9257-4295","institution":"School of Information Technology University of Cincinnati","correspondingAuthor":true,"prefix":"","firstName":"Isaac","middleName":"Kofi","lastName":"Nti","suffix":""},{"id":486234483,"identity":"3d9a32f9-4374-45df-bc6b-4dc6408930ed","order_by":1,"name":"Lee Jo Ning","email":"","orcid":"","institution":"National Chung Hsing University Taichung, Taiwan","correspondingAuthor":false,"prefix":"","firstName":"Lee","middleName":"Jo","lastName":"Ning","suffix":""},{"id":486234484,"identity":"84da3451-d5ad-4bd3-8509-a3c9c096981c","order_by":2,"name":"Clark Alex","email":"","orcid":"","institution":"School of Information Technology University of Cincinnati","correspondingAuthor":false,"prefix":"","firstName":"Clark","middleName":"","lastName":"Alex","suffix":""},{"id":486234485,"identity":"a94924fc-9fe9-4fd0-9e43-477f78868c84","order_by":3,"name":"Miriyala, Sai Manikanta","email":"","orcid":"","institution":"School of Information Technology University of Cincinnati","correspondingAuthor":false,"prefix":"","firstName":"Sai","middleName":"Manikanta","lastName":"Miriyala","suffix":""},{"id":486234486,"identity":"83fe4d17-0d33-4135-a658-1ae260f90ce4","order_by":4,"name":"Murat Ozer","email":"","orcid":"","institution":"School of Information Technology University of Cincinnati Cincinnati, USA","correspondingAuthor":false,"prefix":"","firstName":"Murat","middleName":"","lastName":"Ozer","suffix":""}],"badges":[],"createdAt":"2025-07-16 09:27:49","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-7138288/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7138288/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":86941755,"identity":"e0d091bb-637c-4ae3-a94c-73f2f04b406a","added_by":"auto","created_at":"2025-07-17 12:05:11","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":607371,"visible":true,"origin":"","legend":"","description":"","filename":"IEEEManuscriptdraftFinal.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7138288/v1_covered_dcf2375f-29da-45bd-b2a9-6d4f1f76dc9d.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eEvaluating Lightweight Neural Models for Edge-Based Anomaly Detection: Performance and Efficiency Trade-offs\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"University of Cincinnati","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":"edge-based anomaly detection, lightweight neural models, edge-optimization, resource constraints, inference efficiency","lastPublishedDoi":"10.21203/rs.3.rs-7138288/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7138288/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn edge computing scenarios where constraints on memory, latency, and energy hinder the utilization of large-scale models, lightweight neural networks are increasingly favored for anomaly detection. However, uniform benchmarks for comparing commonly utilized lightweight models under these constraints remain absent. This research addresses the gap by evaluating three prominent lightweight neural architectures, pruned convolutional neural networks (CNNs), quantized long short-term memory networks (LSTMs), and distilled transformers, across two established intrusion detection datasets: CIC-IoT-DIAD 2024 and TON_IoT (TON_IoT_Modbus and TON_IoT_Thermostat). We evaluate each model utilizing standard detection measures (accuracy, precision, recall, and F1-score) and deployment metrics (model size, inference latency, and memory consumption) under simulated edge constraints. Our findings indicate significant trade-offs between model accuracy and efficiency, with performance varying based on the dataset utilized. Certain models perform more effectively with flow-based data compared to others with IoT telemetry. No single model excelled in all evaluation criteria. This research provides future edge-optimized anomaly detection studies with a reliable, reproducible foundation and valuable insights on selecting models for real-time edge deployment. The results also guide our attention in creating our forthcoming architecture, S3LiteNet, intended to enhance performance and deployment in information-centric networks.\u003c/p\u003e","manuscriptTitle":"Evaluating Lightweight Neural Models for Edge-Based Anomaly Detection: Performance and Efficiency Trade-offs","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-17 11:57:04","doi":"10.21203/rs.3.rs-7138288/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":"e0bb9d2a-03ad-452e-bf65-de416489e521","owner":[],"postedDate":"July 17th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":51617093,"name":"Artificial Intelligence and Machine Learning"}],"tags":[],"updatedAt":"2025-07-17T11:57:04+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-17 11:57:04","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7138288","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7138288","identity":"rs-7138288","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}