GradLIME: A CNN Local Interpretation Model Based on Feature Gradient Activation

GradLIME: A CNN Local Interpretation Model Based on Feature Gradient Activation | 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 GradLIME: A CNN Local Interpretation Model Based on Feature Gradient Activation Jinwei Zhao, Jiedong Liu, Zhenghao Shi, Yu Liu, Majid Habib Khan, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7004733/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 6 You are reading this latest preprint version Abstract As deep learning technologies advance rapidly, there is a growing demand for greater transparency and reliability in neural network decision-making. This demand has spurred progress in the explainability of Convolutional Neural Networks (CNNs) in recent years, though significant challenges persist. Current explanation methods typically fall into two categories: those that rely entirely on the internal feature information of neural networks to construct explanations and those that use model-agnostic approaches based on visual concepts. The first approach faces limitations due to the highly abstract nature of the embedded features within neural networks and their fundamental differences from human reasoning processes, leading to inevitable deviations from human cognition. On the other hand, while model-agnostic methods can explore CNNs’ computational logic from a human-centric perspective, their independence from specific models makes it challenging to provide explanations directly linked to the network's computational structure. In some cases, these explanations may even deviate from the true underlying mechanisms of the model. To address these issues, this paper proposes a local explanation model based on feature gradient activation for CNNs, called GradLIME, which is built upon the local interpretable model-agnostic explanations (LIME) method. In the construction of the local linear explanation model, GradLIME incorporates feature gradient activation data from multiple layers of the CNN, facilitating the generation of a comprehensible local linear explanation that also fully utilises the embedded feature pertaining to the network's computational structure. Finally, experiments were conducted on standard datasets to provide qualitative and quantitative evaluations of the local explanations generated by GradLIME. The results demonstrate that, in comparison to numerous state-of-the-art explanation methods that provide visual explanations, GradLIME is more effective at distinguishing between important and unimportant features, and at extracting accurate local explanations that are easier for humans to understand in the context of CNN reasoning. Physical sciences/Mathematics and computing Biological sciences/Neuroscience Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 21 May, 2026 Reviewers invited by journal 16 Sep, 2025 Editor assigned by journal 25 Aug, 2025 Editor invited by journal 22 Jul, 2025 Submission checks completed at journal 06 Jul, 2025 First submitted to journal 06 Jul, 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-7004733","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":516092517,"identity":"0a50d643-31b3-472d-bad0-be9d162d3254","order_by":0,"name":"Jinwei Zhao","email":"","orcid":"","institution":"Xi'an University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Jinwei","middleName":"","lastName":"Zhao","suffix":""},{"id":516092518,"identity":"00dc4003-53aa-48c4-8a67-74e931b7d351","order_by":1,"name":"Jiedong Liu","email":"","orcid":"","institution":"Xi'an University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Jiedong","middleName":"","lastName":"Liu","suffix":""},{"id":516092519,"identity":"5a0759b5-62d6-4781-bc40-0dccd81706d0","order_by":2,"name":"Zhenghao Shi","email":"","orcid":"","institution":"Xi'an University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Zhenghao","middleName":"","lastName":"Shi","suffix":""},{"id":516092520,"identity":"fcee3225-29b4-4418-b621-acaef4e34b16","order_by":3,"name":"Yu Liu","email":"","orcid":"","institution":"Xi'an University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Liu","suffix":""},{"id":516092521,"identity":"a984a0fe-385b-489f-9421-d003def2ba7e","order_by":4,"name":"Majid Habib Khan","email":"","orcid":"","institution":"Xi'an University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Majid","middleName":"Habib","lastName":"Khan","suffix":""},{"id":516092522,"identity":"bc47a05f-0961-46b4-a49c-c925409e3522","order_by":5,"name":"Wei Wang","email":"","orcid":"","institution":"Xi'an University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Wei","middleName":"","lastName":"Wang","suffix":""},{"id":516092523,"identity":"f2bd2c46-9144-4e99-b47a-25e3d69800de","order_by":6,"name":"Minhui Zhu","email":"","orcid":"","institution":"Xi'an Polytechnic University","correspondingAuthor":false,"prefix":"","firstName":"Minhui","middleName":"","lastName":"Zhu","suffix":""},{"id":516092524,"identity":"b072b7ee-2584-4fc1-bd53-5d4a50fa7ce5","order_by":7,"name":"Xinhong Hei","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7UlEQVRIiWNgGAWjYBACAwbGBoYEhgNAJvMBqFgC0VrYEsB8IrSAAUgLjwFxWszZm9skHu64I2fOv+abxM+2Pwz87DkGDD934NZi2XOw2SDxzDNjyxlvNxv2thkwSPa8MWDsPYPHYTcSGx8kth1O3HDj7MYHvEAtBjdyDJgZ2/Bouf+w4QBEy5kHB/8CtdgT1HKDEWrL+R7Gx2BbJAhpOZMI9EvbYWODG2zGxjLnjHkkzjwrONiLT8vx488kf7YdljM4f/iZ5JsyOTn+9uSND37i0YIAEglgigdEHCBGAwMDP5HqRsEoGAWjYOQBAHhlWlgtxahlAAAAAElFTkSuQmCC","orcid":"","institution":"Xi'an University of Technology","correspondingAuthor":true,"prefix":"","firstName":"Xinhong","middleName":"","lastName":"Hei","suffix":""}],"badges":[],"createdAt":"2025-06-29 20:38:06","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7004733/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7004733/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":92129539,"identity":"5b724957-155b-4a6b-8151-7f27d4c85519","added_by":"auto","created_at":"2025-09-25 02:38:16","extension":"json","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9738,"visible":true,"origin":"","legend":"","description":"","filename":"47fb4754d8df4e17bf1a7e1403cfed27.json","url":"https://assets-eu.researchsquare.com/files/rs-7004733/v1/6211005edccb06d4beeec812.json"},{"id":92129669,"identity":"9f51c308-e993-49d0-b66c-553175765d3d","added_by":"auto","created_at":"2025-09-25 02:38:27","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":725929,"visible":true,"origin":"","legend":"","description":"","filename":"ScientificReportsrevision.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7004733/v1_covered_a3063754-f6ae-4fc5-a7be-1c3726339367.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"GradLIME: A CNN Local Interpretation Model Based on Feature Gradient Activation","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"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":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7004733/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7004733/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"As deep learning technologies advance rapidly, there is a growing demand for greater transparency and reliability in neural network decision-making. This demand has spurred progress in the explainability of Convolutional Neural Networks (CNNs) in recent years, though significant challenges persist. Current explanation methods typically fall into two categories: those that rely entirely on the internal feature information of neural networks to construct explanations and those that use model-agnostic approaches based on visual concepts. The first approach faces limitations due to the highly abstract nature of the embedded features within neural networks and their fundamental differences from human reasoning processes, leading to inevitable deviations from human cognition. On the other hand, while model-agnostic methods can explore CNNs’ computational logic from a human-centric perspective, their independence from specific models makes it challenging to provide explanations directly linked to the network's computational structure. In some cases, these explanations may even deviate from the true underlying mechanisms of the model. To address these issues, this paper proposes a local explanation model based on feature gradient activation for CNNs, called GradLIME, which is built upon the local interpretable model-agnostic explanations (LIME) method. In the construction of the local linear explanation model, GradLIME incorporates feature gradient activation data from multiple layers of the CNN, facilitating the generation of a comprehensible local linear explanation that also fully utilises the embedded feature pertaining to the network's computational structure. Finally, experiments were conducted on standard datasets to provide qualitative and quantitative evaluations of the local explanations generated by GradLIME. The results demonstrate that, in comparison to numerous state-of-the-art explanation methods that provide visual explanations, GradLIME is more effective at distinguishing between important and unimportant features, and at extracting accurate local explanations that are easier for humans to understand in the context of CNN reasoning.","manuscriptTitle":"GradLIME: A CNN Local Interpretation Model Based on Feature Gradient Activation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-25 02:38:10","doi":"10.21203/rs.3.rs-7004733/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"294984354397217207475126530946821691069","date":"2026-05-21T16:06:39+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-16T15:21:42+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-08-25T06:05:06+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-07-23T03:16:51+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-07-06T15:46:09+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-07-06T15:44:39+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0d0c2b0e-1be4-45c4-b270-027f1815821d","owner":[],"postedDate":"September 25th, 2025","published":true,"recentEditorialEvents":[{"type":"reviewerAgreed","content":"294984354397217207475126530946821691069","date":"2026-05-21T16:06:39+00:00","index":92,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":54830292,"name":"Physical sciences/Mathematics and computing"},{"id":54830293,"name":"Biological sciences/Neuroscience"}],"tags":[],"updatedAt":"2025-09-25T02:38:10+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-25 02:38:10","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7004733","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7004733","identity":"rs-7004733","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}