Neuron-Inspired Leader–Follower Networks for AI with Local Error Signals

Neuron-Inspired Leader–Follower Networks for AI with Local Error Signals | 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 Neuron-Inspired Leader–Follower Networks for AI with Local Error Signals Paul Bogdan, Anzhe Cheng, Chenzhong Yin, Mingxi Cheng, Xiongye Xiao, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8313432/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 Artificial intelligence (AI) systems often operate under memory and energy limits, similar to biological networks of neurons. The collective behavior of a network with heterogeneous, resource-limited information processing units (e.g., a group of fish, a flock of birds, or a network of neurons) demonstrates a high degree of self-organization, emergence, and complexity. These properties arise from simple interaction rules, where certain individuals can exhibit leadership-like behavior and influence the collective activity of the group. Motivated by collective intelligence in biological neural systems and neuron-level plasticity, we introduce a worker concept to an artificial neural network (NN) for AI. This NN structure contains workers that encompass one or more information processing units (e.g., neurons, filters, layers, or blocks of layers). Workers are either leaders or followers, and we train a leader-follower neural network (LFNN) by leveraging local error signals. The LFNN does not require backpropagation (BP) or a global loss function to achieve optimal performance (we denote LFNN trained without BP as LFNN-ℓ). By investigating worker behavior and evaluating the LFNN and LFNN-ℓ architectures on a variety of image classification tasks(e.g., MNIST, CIFAR-10, ImageNet), we demonstrate that LFNN-ℓ trained with local error signals achieves lower error rates and superior scalability than state-of-the-art machine learning approaches. Furthermore, LFNN-ℓ can be conveniently embedded in classic convolutional NN architectures (e.g., VGG, ResNet, and Vision Transformer (ViT)), achieving a 2x speedup compared to BP-based methods and significantly outperforming models trained with end-to-end BP and other state-of-the-art local learning methods in terms of accuracy on CIFAR-10, Tiny-ImageNet, and ImageNet. Lastly, the proposed LFNN-based model outperforms deep learning counterparts in brain-age prediction from magnetic resonance imaging (MRI) data, while also achieving a 2× speedup. These results indicate that neuron-inspired, decentralized training rules can improve AI accuracy and efficiency. The LFNN-ℓ architecture provides practical tools and insights for addressing core challenges in deep learning and for designing more biologically plausible and sustainable AI. Physical sciences/Mathematics and computing/Computer science Biological sciences/Neuroscience/Computational neuroscience/Network models Full Text Additional Declarations There is NO Competing Interest. Supplementary Files LFNNsupplementary.pdf Supplementary Materials 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-8313432","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":557747898,"identity":"fc471408-1500-4a08-b0cd-52cc38bc9b6f","order_by":0,"name":"Paul Bogdan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAsElEQVRIiWNgGAWjYDACCSDmYWCQg/DYSNBiTLqWxAaitZhLNx9+8abmTvqG84cfMHwoO0xYi+WcY2lA/Cx3w400A8YZ54jQYnAjx8yYh+0wUAsPAzNvG9Fa/h1ONzh/hoH5L5FajB8DDU8wOJDDwMxIjBaQXxjn9h02nAn0y8Gec+mEtYBC7MObb4fl+c4ffvjgR5k1EQ4DxoUEjHOAsHqIFuYPRKkcBaNgFIyCkQsAZsA/qfale/QAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0003-2118-0816","institution":"University of Southern California","correspondingAuthor":true,"prefix":"","firstName":"Paul","middleName":"","lastName":"Bogdan","suffix":""},{"id":557747899,"identity":"b568ee03-d53d-4a1f-8dc9-ee500ca4c359","order_by":1,"name":"Anzhe Cheng","email":"","orcid":"https://orcid.org/0009-0007-4647-8631","institution":"Ming Hsieh Department of Electrical and Computer Engineering, University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Anzhe","middleName":"","lastName":"Cheng","suffix":""},{"id":557747900,"identity":"0e6ff577-c0c7-4a46-9857-6ace26fd840b","order_by":2,"name":"Chenzhong Yin","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Chenzhong","middleName":"","lastName":"Yin","suffix":""},{"id":557747901,"identity":"25fe00df-1942-44f0-83a1-a6df3d5a5fb0","order_by":3,"name":"Mingxi Cheng","email":"","orcid":"","institution":"Ming Hsieh Department of Electrical and Computer Engineering, University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Mingxi","middleName":"","lastName":"Cheng","suffix":""},{"id":557747902,"identity":"964a908b-5555-4170-b826-2271c3e85162","order_by":4,"name":"Xiongye Xiao","email":"","orcid":"","institution":"USC","correspondingAuthor":false,"prefix":"","firstName":"Xiongye","middleName":"","lastName":"Xiao","suffix":""},{"id":557747903,"identity":"24bcfb2d-9819-4133-9422-6e703c9def1b","order_by":5,"name":"Heng Ping","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Heng","middleName":"","lastName":"Ping","suffix":""},{"id":557747904,"identity":"b243e054-4b04-4e48-8104-637c562bbc63","order_by":6,"name":"Andrei Irimia","email":"","orcid":"https://orcid.org/0000-0002-9254-9388","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Andrei","middleName":"","lastName":"Irimia","suffix":""},{"id":557747905,"identity":"c1e57341-eae2-405c-8e27-379a1b47e2e0","order_by":7,"name":"Shahin Nazarian","email":"","orcid":"","institution":"University of Southern California","correspondingAuthor":false,"prefix":"","firstName":"Shahin","middleName":"","lastName":"Nazarian","suffix":""}],"badges":[],"createdAt":"2025-12-09 05:45:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8313432/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8313432/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":98437362,"identity":"cc6bc176-6e36-4f57-9e0b-4a0e2345ec13","added_by":"auto","created_at":"2025-12-17 16:57:15","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2431851,"visible":true,"origin":"","legend":"","description":"","filename":"LFNN.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8313432/v1/cdb363350db0389d2212764e.pdf"},{"id":98318200,"identity":"b595c7ac-495a-49c0-b3e4-6d194ce61dc3","added_by":"auto","created_at":"2025-12-16 13:40:33","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9932,"visible":true,"origin":"","legend":"","description":"","filename":"NCOMMS25099606.json","url":"https://assets-eu.researchsquare.com/files/rs-8313432/v1/129197f8119f8fa85c78943e.json"},{"id":98318201,"identity":"a97d6b61-fe63-4c8a-8e1d-4e23cb5a588e","added_by":"auto","created_at":"2025-12-16 13:40:33","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3107210,"visible":true,"origin":"","legend":"","description":"","filename":"LFNNsupplementary.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8313432/v1/09277fca0017187d885669da.pdf"},{"id":98623146,"identity":"09250014-5b1f-49f4-a294-cd7edd6bd005","added_by":"auto","created_at":"2025-12-19 17:04:53","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1976784,"visible":true,"origin":"","legend":"Article File","description":"","filename":"LFNN.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8313432/v1_covered_2babcac6-598d-4277-b5ba-4e5ae28dcadd.pdf"},{"id":98435514,"identity":"56a092f0-339f-4ac9-9255-f22d3baf5b84","added_by":"auto","created_at":"2025-12-17 16:53:59","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":3107210,"visible":true,"origin":"","legend":"Supplementary Materials","description":"","filename":"LFNNsupplementary.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8313432/v1/a4a201616f03239fc426a482.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Neuron-Inspired Leader–Follower Networks for AI with Local Error Signals","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-8313432/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8313432/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Artificial intelligence (AI) systems often operate under memory and energy limits, similar to biological networks of neurons. The collective behavior of a network with heterogeneous, resource-limited information processing units (e.g., a group of fish, a flock of birds, or a network of neurons) demonstrates a high degree of self-organization, emergence, and complexity. These properties arise from simple interaction rules, where certain individuals can exhibit leadership-like behavior and influence the collective activity of the group. Motivated by collective intelligence in biological neural systems and neuron-level plasticity, we introduce a worker concept to an artificial neural network (NN) for AI. This NN structure contains workers that encompass one or more information processing units (e.g., neurons, filters, layers, or blocks of layers). Workers are either leaders or followers, and we train a leader-follower neural network (LFNN) by leveraging local error signals. The LFNN does not require backpropagation (BP) or a global loss function to achieve optimal performance (we denote LFNN trained without BP as LFNN-ℓ). By investigating worker behavior and evaluating the LFNN and LFNN-ℓ architectures on a variety of image classification tasks(e.g., MNIST, CIFAR-10, ImageNet), we demonstrate that LFNN-ℓ trained with local error signals achieves lower error rates and superior scalability than state-of-the-art machine learning approaches. Furthermore, LFNN-ℓ can be conveniently embedded in classic convolutional NN architectures (e.g., VGG, ResNet, and Vision Transformer (ViT)), achieving a 2x speedup compared to BP-based methods and significantly outperforming models trained with end-to-end BP and other state-of-the-art local learning methods in terms of accuracy on CIFAR-10, Tiny-ImageNet, and ImageNet. Lastly, the proposed LFNN-based model outperforms deep learning counterparts in brain-age prediction from magnetic resonance imaging (MRI) data, while also achieving a 2× speedup. These results indicate that neuron-inspired, decentralized training rules can improve AI accuracy and efficiency. The LFNN-ℓ architecture provides practical tools and insights for addressing core challenges in deep learning and for designing more biologically plausible and sustainable AI.","manuscriptTitle":"Neuron-Inspired Leader–Follower Networks for AI with Local Error Signals","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-16 13:40:28","doi":"10.21203/rs.3.rs-8313432/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":"950e229c-8613-4d70-950e-2f74f0b28dd9","owner":[],"postedDate":"December 16th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":59363977,"name":"Physical sciences/Mathematics and computing/Computer science"},{"id":59363978,"name":"Biological sciences/Neuroscience/Computational neuroscience/Network models"}],"tags":[],"updatedAt":"2026-04-03T03:00:47+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-16 13:40:28","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8313432","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8313432","identity":"rs-8313432","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}