HIAT: Human-in-the-Loop Reinforcement Learning with Auxiliary Task

HIAT: Human-in-the-Loop Reinforcement Learning with Auxiliary Task | 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 HIAT: Human-in-the-Loop Reinforcement Learning with Auxiliary Task Bo Niu, Biao Luo, Yongzheng Cui, Xiaodong Xu, Yuqian Zhao, Yu Feng This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5823719/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 09 Jun, 2025 Read the published version in Artificial Intelligence Review → Version 1 posted 10 You are reading this latest preprint version Abstract Human-in-the-loop reinforcement learning (HIRL) tackles the challenges of inefficient exploration and low sample efficiency by incorporating human feedback into the learning process to guide exploration and speed up learning. Nevertheless, most recent approaches overlook the problem of sample imbalance caused by varying behavioral policy distributions in the early stages of training, which results in policy performance degradation due to biased learning toward dominant low-quality data. In this paper, we propose a human-in-the-loop reinforcement learning method by constructing an auxiliary task, namely Human-in-the-Loop Reinforcement Learning with Auxiliary Task (HIAT). Specifically, state-action pairs are utilized to create an auxiliary task that assesses the similarity between different policies, forcing the model to focus on the differences in data sources. The HIAT method guides the learning process of new policies during the policy evaluation phase by implicitly influencing the assessment of different actions, thereby compensating for sample imbalance. Subsequently, we introduce an adaptive weighting version, namely adaptive weighting based HIAT (AWHIAT), which adjusts the impact of the auxiliary task on the primary task based on their task-relevance and mitigates conflicts between different tasks caused by cognitive load. Through empirical evaluations on six continuous control benchmarks, the results demonstrate that AWHIAT significantly improves mean episode reward and sample efficiency compared to four representative HIRL methods. Reinforcement learning Human-in-the-loop Deep learning Auxiliary task Full Text Additional Declarations No competing interests reported. Supplementary Files HIAT20241230.zip Cite Share Download PDF Status: Published Journal Publication published 09 Jun, 2025 Read the published version in Artificial Intelligence Review → Version 1 posted Editorial decision: Accepted 11 May, 2025 Reviews received at journal 17 Apr, 2025 Reviews received at journal 15 Apr, 2025 Reviews received at journal 30 Mar, 2025 Reviewers agreed at journal 29 Mar, 2025 Reviewers agreed at journal 27 Mar, 2025 Reviewers agreed at journal 27 Mar, 2025 Reviewers invited by journal 27 Mar, 2025 Submission checks completed at journal 26 Mar, 2025 First submitted to journal 22 Mar, 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-5823719","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":435054859,"identity":"a0f9e7c9-50ba-4a4f-8b7d-3ce17c6bc331","order_by":0,"name":"Bo Niu","email":"","orcid":"","institution":"Central South University","correspondingAuthor":false,"prefix":"","firstName":"Bo","middleName":"","lastName":"Niu","suffix":""},{"id":435054860,"identity":"0987f556-4cf0-430c-81e2-8bc179b19f2b","order_by":1,"name":"Biao Luo","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1UlEQVRIiWNgGAWjYBACA2YQWcDAwC8B5kvIEKnFgIFBcgYDYwNQCw9hLTDS4AZYCwNhLebszA8fMBgcljO+3Xz80Y0aCx4G9sNHN+DTYtnMZmwA1GJsdudYYnPOMaDDeNLSbuB12GEGMwkgmbjtRo5hcw4bUIsEjxkBLezfwFo2zwBp+UeUFh6ILRskgFpy24jQYtnMU2yQYJBuLHEjLXF2bp8EDxshv5jzH9/44EOFtRz/jOQDn3O+1cnxsx8+hlcLGCQgc9gIKh8Fo2AUjIJRQBAAAGr+QPUE4vCEAAAAAElFTkSuQmCC","orcid":"","institution":"Central South University","correspondingAuthor":true,"prefix":"","firstName":"Biao","middleName":"","lastName":"Luo","suffix":""},{"id":435054861,"identity":"25b01930-bddb-4eb6-836a-a54210358f17","order_by":2,"name":"Yongzheng Cui","email":"","orcid":"","institution":"Central South University","correspondingAuthor":false,"prefix":"","firstName":"Yongzheng","middleName":"","lastName":"Cui","suffix":""},{"id":435054862,"identity":"0d000728-481d-4fb4-a7ef-52b6701f86af","order_by":3,"name":"Xiaodong Xu","email":"","orcid":"","institution":"Central South University","correspondingAuthor":false,"prefix":"","firstName":"Xiaodong","middleName":"","lastName":"Xu","suffix":""},{"id":435054863,"identity":"20e49430-8338-499f-bc41-14f1052bb2d2","order_by":4,"name":"Yuqian Zhao","email":"","orcid":"","institution":"Central South University","correspondingAuthor":false,"prefix":"","firstName":"Yuqian","middleName":"","lastName":"Zhao","suffix":""},{"id":435054864,"identity":"99752d0a-c712-49d3-8704-fa50518a4778","order_by":5,"name":"Yu Feng","email":"","orcid":"","institution":"Zhejiang University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Feng","suffix":""}],"badges":[],"createdAt":"2025-01-14 03:38:06","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5823719/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5823719/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s10462-025-11260-4","type":"published","date":"2025-06-09T15:57:23+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":84726501,"identity":"97409d0b-e30f-4d96-94cb-163ee8a89b51","added_by":"auto","created_at":"2025-06-16 16:06:03","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1662875,"visible":true,"origin":"","legend":"","description":"","filename":"AIRE0322.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5823719/v1_covered_43e556ea-b03e-4c6a-9ef6-8870e0e79c0b.pdf"},{"id":79406860,"identity":"09d654db-03b8-4630-8aeb-57727b874df5","added_by":"auto","created_at":"2025-03-28 04:28:14","extension":"zip","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":3211077,"visible":true,"origin":"","legend":"","description":"","filename":"HIAT20241230.zip","url":"https://assets-eu.researchsquare.com/files/rs-5823719/v1/b6a03bc9e85a0176809dda6f.zip"}],"financialInterests":"No competing interests reported.","formattedTitle":"HIAT: Human-in-the-Loop Reinforcement Learning with Auxiliary Task","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":"artificial-intelligence-review","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"aire","sideBox":"Learn more about [Artificial Intelligence Review](http://link.springer.com/journal/10462)","snPcode":"10462","submissionUrl":"https://submission.nature.com/new-submission/10462/3","title":"Artificial Intelligence Review","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Reinforcement learning, Human-in-the-loop, Deep learning, Auxiliary task","lastPublishedDoi":"10.21203/rs.3.rs-5823719/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5823719/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Human-in-the-loop reinforcement learning (HIRL) tackles the challenges of inefficient exploration and low sample efficiency by incorporating human feedback into the learning process to guide exploration and speed up learning. Nevertheless, most recent approaches overlook the problem of sample imbalance caused by varying behavioral policy distributions in the early stages of training, which results in policy performance degradation due to biased learning toward dominant low-quality data. In this paper, we propose a human-in-the-loop reinforcement learning method by constructing an auxiliary task, namely Human-in-the-Loop Reinforcement Learning with Auxiliary Task (HIAT). Specifically, state-action pairs are utilized to create an auxiliary task that assesses the similarity between different policies, forcing the model to focus on the differences in data sources. The HIAT method guides the learning process of new policies during the policy evaluation phase by implicitly influencing the assessment of different actions, thereby compensating for sample imbalance. Subsequently, we introduce an adaptive weighting version, namely adaptive weighting based HIAT (AWHIAT), which adjusts the impact of the auxiliary task on the primary task based on their task-relevance and mitigates conflicts between different tasks caused by cognitive load. Through empirical evaluations on six continuous control benchmarks, the results demonstrate that AWHIAT significantly improves mean episode reward and sample efficiency compared to four representative HIRL methods.","manuscriptTitle":"HIAT: Human-in-the-Loop Reinforcement Learning with Auxiliary Task","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-28 04:28:10","doi":"10.21203/rs.3.rs-5823719/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Accepted","date":"2025-05-11T13:18:44+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-04-17T13:21:57+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-04-16T00:37:24+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-03-30T09:54:55+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"25320431308419251699312721458437556091","date":"2025-03-29T13:54:07+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"233621644749240290030722276369957612956","date":"2025-03-27T23:11:08+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"120947683489159721340196493627537189270","date":"2025-03-27T12:26:19+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-03-27T12:11:59+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-03-26T11:04:30+00:00","index":"","fulltext":""},{"type":"submitted","content":"Artificial Intelligence Review","date":"2025-03-22T04:44:16+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"artificial-intelligence-review","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"aire","sideBox":"Learn more about [Artificial Intelligence Review](http://link.springer.com/journal/10462)","snPcode":"10462","submissionUrl":"https://submission.nature.com/new-submission/10462/3","title":"Artificial Intelligence Review","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"79c1e7dd-5dec-4567-a51c-25ddbb24843c","owner":[],"postedDate":"March 28th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-06-16T16:00:12+00:00","versionOfRecord":{"articleIdentity":"rs-5823719","link":"https://doi.org/10.1007/s10462-025-11260-4","journal":{"identity":"artificial-intelligence-review","isVorOnly":false,"title":"Artificial Intelligence Review"},"publishedOn":"2025-06-09 15:57:23","publishedOnDateReadable":"June 9th, 2025"},"versionCreatedAt":"2025-03-28 04:28:10","video":"","vorDoi":"10.1007/s10462-025-11260-4","vorDoiUrl":"https://doi.org/10.1007/s10462-025-11260-4","workflowStages":[]},"version":"v1","identity":"rs-5823719","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5823719","identity":"rs-5823719","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}