Perturbation-Based Feature Attribution

Ask about this topic

Overview

This field studies which input features influence a model’s prediction by changing or removing parts of an input and observing the effect. Methods include masking or occlusion, feature permutation, LIME- and SHAP-style sampling, counterfactual changes, and learned or adaptive perturbations across images, text, tabular data, graphs, and audio. Researchers assess explanations using deletion and insertion tests, stability under input changes, agreement between methods, and uncertainty estimates. A central challenge is that the choice of what to perturb and how to replace it can create unrealistic inputs and alter the explanation, motivating more informative, distribution-aware interventions and standardized evaluation.

Latest papers

CardsList

41 papers6 papers in the last four weeks, up 50% on the four weeks before. 0.0% of all new papers.

  1. Kinematic Causal Attribution: Black-Box Explanations of Motion versus Geometry for 4D Point-Cloud Video Models

    Oct 29, 2026Zerui WangFeature AttributionPerturbation-Based Feature Attribution

  2. LUCID: Local Understanding of Covariate Importance with Conditional Density Estimation

    Oct 26, 2026Yixiao Wang, Srikar Katta, Jon Donnelly +2Perturbation-Based Feature Attribution

    Duke University, Durham, USA · Johns Hopkins University, Baltimore, USA · Bocconi University, Milan, Italy

  3. TempShap: When and How Much a Feature Matters in Time-Series Forecasting

    Oct 23, 2026Chi Zhao, Anastasiia Chelnokova, Jing Liu +1Explainable Time Series ForecastingFeature Attribution

    Saint Petersburg State University, 7/9 Universitetskaya nab., Saint Petersburg 199034, Russia

  4. GRAIL: From Saliency to Semantics in Time-Series Explanation

    Oct 21, 2026Enver Menadjiev, Yiqiu Shen, Jaesik Choi +1Perturbation-Based Feature AttributionCounterfactual Explanations

    KAIST · NYU

  5. Towards Explainable Benchmarking for Data-driven Post-Wildfire Debris Flow Prediction

    Oct 5, 2026Zhisheng Qi, Li Zhu, Utkarsh Sahu +3Perturbation-Based Feature AttributionForecasting Benchmarks

    University of Georgia Athens, Georgia, USA · University of Oregon Eugene, Oregon, USA

  6. A Comparative Explainability Framework for DeBERTa-v3 in Zero-Shot Medical Abstract Classification

    Oct 1, 2026Javier Diaz Esteban-Herreros, David Muñoz-Valero, Raquel Martínez-España +2Transformer InterpretabilityPerturbation-Based Feature Attribution

    Department of Technologies and Information Systems, Universidad de Castilla-La Mancha, Avenida Carlos III, s/n, Toledo, 45071, Spain · University of Murcia, Avenida Teniente Flomesta, 5, Murcia, 3003, Spain · Murcian Bio-Health Institute (IMIB-Arrixaca), Pabellón Docente del Hospital Clínico Universitario Virgen de la Arrixaca, Murcia, 3120, Spain

  7. LOCO-AdaMP: Built-in LOCO Inference for Adaptive Minipatch Ensembles with Enhanced Prediction

    Sep 28, 2026Yinan Cheng, Lili ZhengEnsemble LearningAdaptive Sampling

    Department of Statistics University of Illinois Urbana-Champaign

  8. When Is an SAE Feature Interpretable? A Validation Ladder for EEG Foundation Models

    Sep 28, 2026Yucong Cao, Chenqi Li, Tingting ZhuPerturbation-Based Feature AttributionSparse Autoencoders

    Department of Engineering Science University of Oxford

  9. Mask-Induced Displacement in Audio XAI via Logit Trajectory Decomposition

    Sep 27, 2026Nico García-Peguinho, David Kelly, Fabrizio Smeraldi +1Explainable Artificial IntelligencePerturbation-Based Feature Attribution

    School of Electronic Engineering and Computer Science, Queen Mary University of London · Department of Informatics, King's College London

  10. Faithful Faithfulness Evaluations: Challenges & Pitfalls Learned from a Breast MRI Case Study

    Sep 22, 2026Peachapong Poolpol, Henrik H. J. Detjen, Eike PetersenPerturbation-Based Feature AttributionSaliency Map Evaluation

    Fraunhofer Institute for Digital Medicine MEVIS, Bremen, Germany · Deggendorf Institute of Technology, Deggendorf, Germany

  11. EEG-Xplain: Decoding Neural Black-Boxes of EEG Foundation Models

    Sep 14, 2026Hansong Ma, Junxiao WangGradient-Based AttributionPerturbation-Based Feature Attribution

    Guangzhou University, China

  12. How Faithful Is Attribution for Sales Forecasting? A Counterfactual Study

    Sep 4, 2026Glib KechynExplainable Time Series ForecastingFeature Attribution

  13. VINCENT: Validated Interaction Network for Cross-drug Explanation of Therapeutics

    Aug 26, 2026Fan-Sheng Chuang, Xuchen Li, Yujing Bian +1Drug DiscoveryDrug-Drug Interaction Prediction

    North Carolina State University Computer Science Raleigh, USA · North Carolina State University Electrical and Computer Engineering Raleigh, USA

  14. NICE: Scale-Stable Perturbations for Graph Neural Network Explanations via Noise Corruption

    Aug 17, 2026Ziluowen Luo, Jun Yin, Ruochen Liu +4GNN ExplainabilityPerturbation-Based Feature Attribution

    Central South University · Hong Kong Polytechnic University · Griffith University

  15. Decomposition of Evidence, Contradiction, and Fragility in Perturbation Responses

    Aug 13, 2026Lei YouFeature AttributionPerturbation-Based Feature Attribution

    Technical University of Denmark

  16. Spatial Attention Noise Masking for Causally Sufficient Interpretability

    Aug 12, 2026Benjamin Formby, Kuang-Ching Wang, D Hudson SmithExplainable Image ClassificationCausal Interpretability

    Electrical and Computer Engineering department at Clemson University, Clemson, SC · School of Computing at Binghamton University, Binghamton, NY · Department of Mathematical and Statistical Sciences at Clemson University, Clemson, SC

  17. Inpainting Insights: Elevating Visual XAI with Photorealistic Perturbations

    Jul 16, 2026Josef Lindl, Mariana Chaves, Damien GarreauPerturbation-Based Feature AttributionExplainable Image Classification

    Julius-Maximilians-Universität Würzburg, Sanderring 2, 97070 Würzburg, Germany · University of Groningen: Broerstraat 5, 9712 CP Groningen, The Netherlands

  18. On the Disagreement in Perturbation-based xAI -- Benchmarking Perturbation Choices for Flood Detection from SAR Images

    Jul 16, 2026Anastasia Schlegel, Ronny HänschFeature AttributionPerturbation-Based Feature Attribution

    Microwaves and Radar Institute, German Aerospace Center (DLR), Weßling, 82234, Germany · Chair of Remote Sensing Technology, Department of Aerospace and Geodesy, Technical University of Munich, Munich, 80333, Germany

  19. Local Additive Feature Attribution: A Mathematical Taxonomy and Reporting Checklist

    Jul 15, 2026Rebecca Afriyie Sarpong, Daniel CommeyGradient-Based AttributionFeature Attribution

  20. A Good Initialization is All You Need for Faithful Visual Attribution

    Jul 7, 2026Zihan Gu, Jiayu Wang, Hua Zhang +1Feature AttributionPerturbation-Based Feature Attribution

    Institute of Information Engineering, Chinese Academy of Sciences · School of Cyber Security, University of Chinese Academy of Sciences

  21. Re-Ranking Through an Attribution Lens for Citation Quality in Legal QA

    Jun 2, 2026Mohamed Hesham Elganayni, Selim SalehCross-Encoder RerankingFeature Attribution

    Technical University of Munich, Munich, Germany

  22. Learning Quantifiable Visual Explanations Without Ground-Truth

    May 18, 2026Amritpal Singh, Andrey Barsky, Mohamed Ali Souibgui +2Perturbation-Based Feature AttributionCausal Interpretability

    Computer Vision Center, Barcelona, Spain · Autonomous University of Barcelona, Spain

  23. From Weight Perturbation to Feature Attribution for Explaining Fully Connected Neural Networks

    May 14, 2026Thodoris Lymperopoulos, Denia KanellopoulouFeature AttributionPerturbation-Based Feature Attribution

    NCSR Demokritos Athens, Greece

  24. FAME: Feature Activation Map Explanation on Image Classification and Face Recognition

    May 12, 2026Xinyi Zhang, Manuel GüntherGradient-Based AttributionFeature Attribution

    Department of Informatics, University of Zurich

  25. BAMI: Training-Free Bias Mitigation in GUI Grounding

    May 7, 2026Borui Zhang, Bo Zhang, Bo Wang +6Perturbation-Based Feature AttributionGUI Grounding

    Tsinghua University, China · Lenovo Research, China

  26. Efficient KernelSHAP Explanations for Patch-based 3D Medical Image Segmentation

    Apr 13, 2026Ricardo Coimbra Brioso, Giulio Sichili, Damiano Dei +4Explainable Medical Image AnalysisSHAP Feature Attribution

    Dipartimento di Elettronica, Informazione e Bioingegneria, Politecnico di Milano, Milan, Italy · Radiotherapy and Radiosurgery Department, IRCCS Humanitas Research Hospital, Rozzano, Milan, Italy · Scuola di Specializzazione di Fisica Medica, Università Degli Studi di Milano, Milan, Italy +1

  27. Trustworthy Feature Importance Avoids Unrestricted Permutations

    Apr 13, 2026Emanuele Borgonovo, Francesco Cappelli, Xuefei Lu +2Feature AttributionPerturbation-Based Feature Attribution

    Department of Decision Sciences and Bocconi Institute for Data Science and Analytics, Bocconi University, Milan, Italy · Department of Decision Science, Bocconi University, Milan, Italy · SKEMA Business School, Université Côte d’Azur, France +2

  28. Non-identifiability of Explanations from Model Behavior in Deep Networks of Image Authenticity Judgments

    Apr 8, 2026Icaro Re Depaolini, Uri HassonExplainability EvaluationFeature Attribution

    Center for Mind/Brain Sciences, The University of Trento, Trento, Italy

  29. Feature Attribution Stability Suite: How Stable Are Post-Hoc Attributions?

    Apr 2, 2026Kamalasankari Subramaniakuppusamy, Jugal GajjarExplainability EvaluationPerturbation-Based Feature Attribution

    The George Washington University Washington D.C.

  30. VISTA: Visualization of Token Attribution via Efficient Analysis

    Apr 2, 2026Syed Ahmed, Bharathi Vokkaliga Ganesh, Jagadish Babu P +8LLM InterpretabilityPerturbation-Based Feature Attribution

    Responsible AI Office Infosys Limited, Bangalore, India