Single-Thinking-Process-Based Learning of Complete Issue-Related Materials for Autonomous Agents

Single-Thinking-Process-Based Learning of Complete Issue-Related Materials for Autonomous Agents | 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 Single-Thinking-Process-Based Learning of Complete Issue-Related Materials for Autonomous Agents Hong Su This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9615104/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 Autonomous agents operating in open and long-term environments should be able to learn from their own observations, actions, reasoning processes, external model interactions, and feedback with reduced human guidance. However, many existing agent learning frameworks treat perception, reasoning, memory, action, and feedback as relatively separated processes, which may cause useful learning materials generated around the same issue to be fragmented or ignored. This paper proposes a Single-Thinking-Process-Based Learning of Complete Issue-Related Materials model for autonomous agents. The core idea is that all issue-related information is processed through a unified thinking process, including sensory observations, action contexts, LLM prompts, LLM responses, recalled cases, comparison results, judgments, and feedback. Since learning materials are generated through processing, deeper or repeated thinking can produce richer materials for the same issue, while unprocessed data remain raw and do not form structured training materials. The collected materials are organized into named sequences to support recall, comparison, reuse, merging, branching, and learning carrier update. The model also introduces active processing for unmastered issues, allowing the agent to obtain additional observations, reasoning results, or feedback when the current material is insufficient. Controlled verification results show that the proposed model improves recall coverage from about 0.328 to 0.965, improves reuse success rate from 0.919 to 0.991, and increases mastery degree from 0.742 to 0.808 compared with LLM-only processing. Large Language Models (LLMs) Single-Thinking-Process-Based Learning Complete Issue-Related Materials learning materials 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-9615104","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":634592127,"identity":"7fc9759e-25b6-42b3-a1ae-cdc710a08853","order_by":0,"name":"Hong Su","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAt0lEQVRIiWNgGAWjYBADOTb25gMk6TAw5uM5lkCalsR5EjkKxKmVn5F+TeLDnz/pbQw5DAw/KrYRYfyNnDLJmW0GuW0MZw8w9py5TYQWiZw0ad4GoBbGvgRmxjYitMjPAGr588cgnY2Zx4A4LQw30o9JM7AZJLCxEavF4MwbZsveNmPDNh62hINE+UW+Pf3hjR9/5OTl5z8++OBHBTEOY+AxgDMPEKMeCNgfEKlwFIyCUTAKRiwAAJtVOMUT72qXAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-2750-7808","institution":"Chengdu University of Information Technology","correspondingAuthor":true,"prefix":"","firstName":"Hong","middleName":"","lastName":"Su","suffix":""}],"badges":[],"createdAt":"2026-05-05 07:22:57","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-9615104/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9615104/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108805359,"identity":"13a8e550-e723-440b-a55f-d2afcb763725","added_by":"auto","created_at":"2026-05-08 15:25:41","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":599820,"visible":true,"origin":"","legend":"","description":"","filename":"mv18.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9615104/v1_covered_38cd2703-5c58-490f-b427-a5f8a814a44a.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eSingle-Thinking-Process-Based Learning of Complete Issue-Related Materials for Autonomous Agents\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"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":"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":"Large Language Models (LLMs), Single-Thinking-Process-Based Learning, Complete Issue-Related Materials, learning materials","lastPublishedDoi":"10.21203/rs.3.rs-9615104/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9615104/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAutonomous agents operating in open and long-term environments should be able to learn from their own observations, actions, reasoning processes, external model interactions, and feedback with reduced human guidance. However, many existing agent learning frameworks treat perception, reasoning, memory, action, and feedback as relatively separated processes, which may cause useful learning materials generated around the same issue to be fragmented or ignored. This paper proposes a Single-Thinking-Process-Based Learning of Complete Issue-Related Materials model for autonomous agents. The core idea is that all issue-related information is processed through a unified thinking process, including sensory observations, action contexts, LLM prompts, LLM responses, recalled cases, comparison results, judgments, and feedback. Since learning materials are generated through processing, deeper or repeated thinking can produce richer materials for the same issue, while unprocessed data remain raw and do not form structured training materials. The collected materials are organized into named sequences to support recall, comparison, reuse, merging, branching, and learning carrier update. The model also introduces active processing for unmastered issues, allowing the agent to obtain additional observations, reasoning results, or feedback when the current material is insufficient. Controlled verification results show that the proposed model improves recall coverage from about 0.328 to 0.965, improves reuse success rate from 0.919 to 0.991, and increases mastery degree from 0.742 to 0.808 compared with LLM-only processing.\u003c/p\u003e","manuscriptTitle":"Single-Thinking-Process-Based Learning of Complete Issue-Related Materials for Autonomous Agents","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-06 06:19:50","doi":"10.21203/rs.3.rs-9615104/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":"5041796e-91cb-4d63-b746-c21c5d938a0a","owner":[],"postedDate":"May 6th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-05-06T06:19:50+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-06 06:19:50","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9615104","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9615104","identity":"rs-9615104","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}