mlsanityData to 14 Oct 2026

Improving KernelSHAP: Practical Shapley Value Estimation via Linear Regression

arXiv 2012.01536 ↗First posted 2 Dec 2020cs.LG stat.MLcited by 154

Not on the map. Why?

Abstract (arXiv)

The Shapley value concept from cooperative game theory has become a popular technique for interpreting ML models, but efficiently estimating these values remains challenging, particularly in the model-agnostic setting. Here, we revisit the idea of estimating Shapley values via linear regression to understand and improve upon this approach. By analyzing the original KernelSHAP alongside a newly proposed unbiased version, we develop techniques to detect its convergence and calculate uncertainty estimates. We also find that the original version incurs a negligible increase in bias in exchange for significantly lower variance, and we propose a variance reduction technique that further accelerates the convergence of both estimators. Finally, we develop a version of KernelSHAP for stochastic cooperative games that yields fast new estimators for two global explanation methods.

Key facts. Cited by 154 papers in the map (data through 14 Oct 2026). First posted on arXiv 2 Dec 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 15; 2022: 18; 2023: 23; 2024: 39; 2025: 30; 2026: 29.

154 papers.

Sort:

By area:

  1. 2021Beta Shapley: a Unified and Noise-reduced Data Valuation Framework for Machine Learningcited by 165
  2. 2021FastSHAP: Real-Time Shapley Value Estimationcited by 125
  3. 2022Algorithms to estimate Shapley value feature attributionscited by 115
  4. 2022The Shapley Value in Machine Learningcited by 98
  5. 2022Faith-Shap: The Faithful Shapley Interaction Indexcited by 69
  6. 2022Rethinking Explainability as a Dialogue: A Practitioner's Perspectivecited by 52
  7. 2023SHAP-IQ: Unified Approximation of any-order Shapley Interactionscited by 51
  8. 2022Measuring the Effect of Training Data on Deep Learning Predictions via Randomized Experimentscited by 50
  9. 2022Learning to Estimate Shapley Values with Vision Transformerscited by 45
  10. 2024Everybody Prune Now: Structured Pruning of LLMs with only Forward Passescited by 43
  11. 2024shapiq: Shapley Interactions for Machine Learningcited by 43
  12. 2021Explaining a Series of Models by Propagating Shapley Valuescited by 37
38 more
  1. 2023Explain Any Concept: Segment Anything Meets Concept-Based Explanationcited by 37
  2. 2021Defining and Quantifying the Emergence of Sparse Concepts in DNNscited by 36
  3. 2022WeightedSHAP: analyzing and improving Shapley based feature attributionscited by 36
  4. 2023Explaining Predictive Uncertainty with Information Theoretic Shapley Valuescited by 36
  5. 2023Approximating the Shapley Value without Marginal Contributionscited by 34
  6. 2021On the Complexity of SHAP-Score-Based Explanations: Tractability via Knowledge Compilation and Non-Approximability Resultscited by 32
  7. 2023The Inadequacy of Shapley Values for Explainabilitycited by 31
  8. 2024Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensemblescited by 31
  9. 2021RKHS-SHAP: Shapley Values for Kernel Methodscited by 30
  10. 2021On Locality of Local Explanation Modelscited by 25
  11. 2021The Shapley Value of Classifiers in Ensemble Gamescited by 23
  12. 2024Provably Accurate Shapley Value Estimation via Leverage Score Samplingcited by 23
  13. 2021SHAFF: Fast and consistent SHApley eFfect estimates via random Forestscited by 22
  14. 2024KernelSHAP-IQ: Weighted Least-Square Optimization for Shapley Interactionscited by 22
  15. 2021groupShapley: Efficient prediction explanation with Shapley values for feature groupscited by 21
  16. 2024Stochastic Amortization: A Unified Approach to Accelerate Feature and Data Attributioncited by 21
  17. 2023CoRTX: Contrastive Framework for Real-time Explanationcited by 20
  18. 2023Explaining the Uncertain: Stochastic Shapley Values for Gaussian Process Modelscited by 19
  19. 2025Regression-adjusted Monte Carlo Estimators for Shapley Values and Probabilistic Valuescited by 19
  20. 2023HarsanyiNet: Computing Accurate Shapley Values in a Single Forward Propagationcited by 17
  21. 2023Efficient XAI Techniques: A Taxonomic Surveycited by 14
  22. 2021Decorrelated Variable Importancecited by 13
  23. 2022Accelerating Shapley Explanation via Contributive Cooperator Selectioncited by 13
  24. 2025Exact Computation of Any-Order Shapley Interactions for Graph Neural Networkscited by 13
  25. 2022Statistical Aspects of SHAP: Functional ANOVA for Model Interpretationcited by 12
  26. 2022Model Agnostic Interpretability for Multiple Instance Learningcited by 10
  27. 2021Efficient Explanations from Empirical Explainerscited by 9
  28. 2021Accurate Shapley Values for explaining tree-based modelscited by 9
  29. 2022Incremental Permutation Feature Importance (iPFI): Towards Online Explanations on Data Streamscited by 9
  30. 2023Stabilizing Estimates of Shapley Values with Control Variatescited by 9
  31. 2024Feature Inference Attack on Shapley Valuescited by 9
  32. 2024One Sample Fits All: Approximating All Probabilistic Values Simultaneously and Efficientlycited by 9
  33. 2022Differentially Private Shapley Values for Data Evaluationcited by 8
  34. 2024SHAP values via sparse Fourier representationcited by 8
  35. 2025SHAP Meets Tensor Networks: Provably Tractable Explanations with Parallelismcited by 8
  36. 2021Shapley variable importance clouds for interpretable machine learningcited by 7
  37. 2024Statistical Significance of Feature Importance Rankingscited by 7
  38. 2026PolySHAP: Extending KernelSHAP with Interaction-Informed Polynomial Regressioncited by 7

50 most cited of 154 All 154 (JSON)