Multi-Agent based Dynamic Anchors for Interpretation of Deep Learning Classifiers

Multi-Agent based Dynamic Anchors for Interpretation of Deep Learning Classifiers | 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 Multi-Agent based Dynamic Anchors for Interpretation of Deep Learning Classifiers Supreeth Suresh, Suresh Muknahallipatna This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9390427/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 4 You are reading this latest preprint version Abstract Explainable Artificial Intelligence (XAI) provides insights into how black-box models make decisions. Among existing approaches, anchors provide high-precision, human-interpretable rules in the form of simple if-then conditions over input features. Classical anchors compute discrete instance-wise rules using a bandit-guided beam search without learning across instances or coordinating rules across classes. Consequently, they are fundamentally local and do not yield a coherent picture of the model's decision regions.We propose Reinforcement Learning Dynamic Anchors (RLDA) , a reinforcement learning (RL) formulation of anchor discovery, in which a policy learns to refine an axis-aligned box around an instance through a sequence of continuous actions, directly optimizing interpretable quantities such as precision and coverage. We then extend this framework to Multi-Agent Dynamic Anchors (MADA) , a cooperative game with one or more agents per class, where agents jointly learn class-wise anchor regions under shared rewards that encourage both local fidelity and a global structure, operating under defined equilibrium conditions.The trained policies were applied to data samples to generate both instance- and class-level rules, which were then tested globally across all classes. Experiments on standard tabular datasets showed that, first, RLDA provides more precise rules and the performance is comparable to classical anchors while producing reusable policies; and second, MADA yields class-wise rules with high precision, useful coverage, and reduced cross-class overlap, thereby providing a more global and structured explanation of the classifier. Explainable AI Interpretability Reinforcement Learning Multi-Agent Reinforcement Learning Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 24 Apr, 2026 Editor assigned by journal 15 Apr, 2026 Submission checks completed at journal 15 Apr, 2026 First submitted to journal 11 Apr, 2026 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-9390427","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":622531610,"identity":"cf767696-6f3b-4172-9662-450b5685b240","order_by":0,"name":"Supreeth Suresh","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAt0lEQVRIiWNgGAWjYPCCA3JQBjPxWoxJ15LYQLQW+f4zZh8+/LmTvuFGjuEDhgprmF7cwOBGjvHMmW3PcoFajA0YzqQToUWCx5iZt+Fw7obbOWYSjG2HCWsBOsyY+c+fw+kGYC3/iNDCcCDHmJmB7XACREsDEVoMbqQVM/a2HTacef9ZsUHCsXRjIhx2eDPDjz+H5fnOHN744EONtSxhhyEAhwFDAgnKQYD9AYkaRsEoGAWjYKQAAMlUQnzyLJ87AAAAAElFTkSuQmCC","orcid":"","institution":"University of Wyoming","correspondingAuthor":true,"prefix":"","firstName":"Supreeth","middleName":"","lastName":"Suresh","suffix":""},{"id":622531612,"identity":"ff889138-4530-42c9-a125-f97257fcd69c","order_by":1,"name":"Suresh Muknahallipatna","email":"","orcid":"","institution":"University of Wyoming","correspondingAuthor":false,"prefix":"","firstName":"Suresh","middleName":"","lastName":"Muknahallipatna","suffix":""}],"badges":[],"createdAt":"2026-04-11 20:38:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9390427/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9390427/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106961006,"identity":"5bf76328-8378-4412-a482-383f08df5075","added_by":"auto","created_at":"2026-04-15 09:23:57","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3077837,"visible":true,"origin":"","legend":"","description":"","filename":"DynAnchorsSpringerNCA.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9390427/v1_covered_66f3c82e-e6ed-4249-8ec0-42119346090c.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Multi-Agent based Dynamic Anchors for Interpretation of Deep Learning Classifiers","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":"discover-artificial-intelligence","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"diai","sideBox":"Learn more about [Discover Artificial Intelligence](https://www.springer.com/44163)","snPcode":"","submissionUrl":"","title":"Discover Artificial Intelligence","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Explainable AI, Interpretability, Reinforcement Learning, Multi-Agent Reinforcement Learning","lastPublishedDoi":"10.21203/rs.3.rs-9390427/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9390427/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eExplainable Artificial Intelligence (XAI) provides insights into how black-box models make decisions. Among existing approaches, anchors provide high-precision, human-interpretable rules in the form of simple \u003cb\u003eif-then\u003c/b\u003e conditions over input features. Classical anchors compute discrete instance-wise rules using a bandit-guided beam search without learning across instances or coordinating rules across classes. Consequently, they are fundamentally local and do not yield a coherent picture of the model's decision regions.We propose \u003cem\u003eReinforcement Learning Dynamic Anchors (RLDA)\u003c/em\u003e, a reinforcement learning (RL) formulation of anchor discovery, in which a policy learns to refine an axis-aligned box around an instance through a sequence of continuous actions, directly optimizing interpretable quantities such as precision and coverage. We then extend this framework to \u003cem\u003eMulti-Agent Dynamic Anchors (MADA)\u003c/em\u003e, a cooperative game with one or more agents per class, where agents jointly learn class-wise anchor regions under shared rewards that encourage both local fidelity and a global structure, operating under defined equilibrium conditions.The trained policies were applied to data samples to generate both instance- and class-level rules, which were then tested globally across all classes. Experiments on standard tabular datasets showed that, first, RLDA provides more precise rules and the performance is comparable to classical anchors while producing reusable policies; and second, MADA yields class-wise rules with high precision, useful coverage, and reduced cross-class overlap, thereby providing a more global and structured explanation of the classifier.\u003c/p\u003e","manuscriptTitle":"Multi-Agent based Dynamic Anchors for Interpretation of Deep Learning Classifiers","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-14 13:49:16","doi":"10.21203/rs.3.rs-9390427/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-04-24T18:54:23+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-15T07:17:46+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-15T07:17:03+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Artificial Intelligence","date":"2026-04-11T20:27:27+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-artificial-intelligence","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"diai","sideBox":"Learn more about [Discover Artificial Intelligence](https://www.springer.com/44163)","snPcode":"","submissionUrl":"","title":"Discover Artificial Intelligence","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"9711380a-412c-4255-b259-0621057dab98","owner":[],"postedDate":"April 14th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-29T03:08:15+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-14 13:49:16","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9390427","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9390427","identity":"rs-9390427","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}