mlsanityData to 14 Oct 2026

The many Shapley values for model explanation

arXiv 1908.08474 ↗First posted 22 Aug 2019cs.AI cs.LG econ.THcited by 344

Not on the map. Why?

Abstract (arXiv)

The Shapley value has become a popular method to attribute the prediction of a machine-learning model on an input to its base features. The use of the Shapley value is justified by citing [16] showing that it is the unique method that satisfies certain good properties (axioms). There are, however, a multiplicity of ways in which the Shapley value is operationalized in the attribution problem. These differ in how they reference the model, the training data, and the explanation context. These give very different results, rendering the uniqueness result meaningless. Furthermore, we find that previously proposed approaches can produce counterintuitive attributions in theory and in practice---for instance, they can assign non-zero attributions to features that are not even referenced by the model. In this paper, we use the axiomatic approach to study the differences between some of the many operationalizations of the Shapley value for attribution, and propose a technique called Baseline Shapley (BShap) that is backed by a proper uniqueness result. We also contrast BShap with Integrated Gradients, another extension of Shapley value to the continuous setting.

Key facts. Cited by 344 papers in the map (data through 14 Oct 2026). First posted on arXiv 22 Aug 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 53; 2022: 64; 2023: 67; 2024: 56; 2025: 59; 2026: 45.

344 papers.

Sort:

By area:

  1. 2022From Anecdotal Evidence to Quantitative Evaluation Methods: A Systematic Review on Evaluating Explainable AIcited by 302
  2. 2021Beta Shapley: a Unified and Noise-reduced Data Valuation Framework for Machine Learningcited by 165
  3. 2022Training Data Influence Analysis and Estimation: A Surveycited by 135
  4. 2023Foundation Models and Fair Usecited by 116
  5. 2022Algorithms to estimate Shapley value feature attributionscited by 115
  6. 2021Guided Integrated Gradients: An Adaptive Path Method for Removing Noisecited by 102
  7. 2022The Shapley Value in Machine Learningcited by 98
  8. 2021The Out-of-Distribution Problem in Explainability and Search Methods for Feature Importance Explanationscited by 87
  9. 2022Impossibility Theorems for Feature Attributioncited by 78
  10. 2022Faith-Shap: The Faithful Shapley Interaction Indexcited by 69
  11. 2022A Rigorous Study of Integrated Gradients Method and Extensions to Internal Neuron Attributionscited by 64
  12. 2022From Shapley Values to Generalized Additive Models and backcited by 60
38 more
  1. 2021Synthetic Benchmarks for Scientific Research in Explainable Machine Learningcited by 59
  2. 2022Additive MIL: Intrinsically Interpretable Multiple Instance Learning for Pathologycited by 58
  3. 2024Don't Say No: Jailbreaking LLM by Suppressing Refusalcited by 56
  4. 2021Shapley values for feature selection: The good, the bad, and the axiomscited by 54
  5. 2022Post-Hoc Explanations Fail to Achieve their Purpose in Adversarial Contextscited by 53
  6. 2021Local Explanations via Necessity and Sufficiency: Unifying Theory and Practicecited by 45
  7. 2024shapiq: Shapley Interactions for Machine Learningcited by 43
  8. 2021Shapley Explanation Networkscited by 41
  9. 2022A Survey on Explainable Anomaly Detectioncited by 39
  10. 2021Shapley Counterfactual Credits for Multi-Agent Reinforcement Learningcited by 38
  11. 2021Explaining a Series of Models by Propagating Shapley Valuescited by 37
  12. 2022MultiViz: Towards Visualizing and Understanding Multimodal Modelscited by 36
  13. 2022WeightedSHAP: analyzing and improving Shapley based feature attributionscited by 36
  14. 2023Explaining Predictive Uncertainty with Information Theoretic Shapley Valuescited by 36
  15. 2023CausalTime: Realistically Generated Time-series for Benchmarking of Causal Discoverycited by 36
  16. 2023LINe: Out-of-Distribution Detection by Leveraging Important Neuronscited by 34
  17. 2021Neural Response Interpretation through the Lens of Critical Pathwayscited by 32
  18. 2023The Inadequacy of Shapley Values for Explainabilitycited by 31
  19. 2024Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensemblescited by 31
  20. 2021RKHS-SHAP: Shapley Values for Kernel Methodscited by 30
  21. 2021On Baselines for Local Feature Attributionscited by 29
  22. 2021Fast TreeSHAP: Accelerating SHAP Value Computation for Treescited by 29
  23. 2024Decoupling Pixel Flipping and Occlusion Strategy for Consistent XAI Benchmarkscited by 28
  24. 2022Understanding Instance-Level Impact of Fairness Constraintscited by 27
  25. 2021On Locality of Local Explanation Modelscited by 25
  26. 2021Collective eXplainable AI: Explaining Cooperative Strategies and Agent Contribution in Multiagent Reinforcement Learning with Shapley Valuescited by 25
  27. 2021The Shapley Value of Classifiers in Ensemble Gamescited by 23
  28. 2021You Mostly Walk Alone: Analyzing Feature Attribution in Trajectory Predictioncited by 23
  29. 2021SHAFF: Fast and consistent SHApley eFfect estimates via random Forestscited by 22
  30. 2021Counterfactual Shapley Additive Explanationscited by 22
  31. 2021groupShapley: Efficient prediction explanation with Shapley values for feature groupscited by 21
  32. 2022Do Explanations Explain? Model Knows Bestcited by 20
  33. 2022Evaluating Feature Attribution Methods in the Image Domaincited by 19
  34. 2022Acceptability Judgements via Examining the Topology of Attention Mapscited by 19
  35. 2023On the Robustness of Removal-Based Feature Attributionscited by 19
  36. 2021Rational Shapley Valuescited by 18
  37. 2022Unifying local and global model explanations by functional decomposition of low dimensional structurescited by 18
  38. 2022"Why did the Model Fail?": Attributing Model Performance Changes to Distribution Shiftscited by 18

50 most cited of 344 All 344 (JSON)