mlsanityData to 14 Oct 2026

SHAP-IQ: Unified Approximation of any-order Shapley Interactions

arXiv 2303.01179 ↗First posted 2 Mar 2023cs.LG cs.AIcited by 51

Not on the map. Why?

Abstract (arXiv)

Predominately in explainable artificial intelligence (XAI) research, the Shapley value (SV) is applied to determine feature attributions for any black box model. Shapley interaction indices extend the SV to define any-order feature interactions. Defining a unique Shapley interaction index is an open research question and, so far, three definitions have been proposed, which differ by their choice of axioms. Moreover, each definition requires a specific approximation technique. Here, we propose SHAPley Interaction Quantification (SHAP-IQ), an efficient sampling-based approximator to compute Shapley interactions for arbitrary cardinal interaction indices (CII), i.e. interaction indices that satisfy the linearity, symmetry and dummy axiom. SHAP-IQ is based on a novel representation and, in contrast to existing methods, we provide theoretical guarantees for its approximation quality, as well as estimates for the variance of the point estimates. For the special case of SV, our approach reveals a novel representation of the SV and corresponds to Unbiased KernelSHAP with a greatly simplified calculation. We illustrate the computational efficiency and effectiveness by explaining language, image classification and high-dimensional synthetic models.

Key facts. Cited by 51 papers in the map (data through 14 Oct 2026). Builds on 24 of its 54 references that are in the map. First posted on arXiv 2 Mar 2023.

Builds on

24 of 54 references.

References · 24

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  3. 2017A Unified Approach to Interpreting Model Predictionscited by 6,555
  4. 2019DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lightercited by 4,307
  5. 2019Data Shapley: Equitable Valuation of Data for Machine Learningcited by 629
  6. 2021Datasets: A Community Library for Natural Language Processingcited by 530
  7. 2019Towards Efficient Data Valuation Based on the Shapley Valuecited by 281
  8. 2020Explaining by Removing: A Unified Framework for Model Explanationcited by 218
  9. 2018Representer Point Selection for Explaining Deep Neural Networkscited by 187
  10. 2020Improving KernelSHAP: Practical Shapley Value Estimation via Linear Regressioncited by 154
  11. 2019The Shapley Taylor Interaction Indexcited by 143
  12. 2022Algorithms to estimate Shapley value feature attributionscited by 115
12 more
  1. 2020Explaining Explanations: Axiomatic Feature Interactions for Deep Networkscited by 103
  2. 2017Detecting Statistical Interactions from Neural Network Weightscited by 80
  3. 2020How does this interaction affect me? Interpretable attribution for feature interactionscited by 73
  4. 2022Faith-Shap: The Faithful Shapley Interaction Indexcited by 69
  5. 2018Beyond Word Importance: Contextual Decomposition to Extract Interactions from LSTMscited by 67
  6. 2018Hierarchical interpretations for neural network predictionscited by 66
  7. 2022TalkToModel: Explaining Machine Learning Models with Interactive Natural Language Conversationscited by 66
  8. 2020Efficient nonparametric statistical inference on population feature importance using Shapley valuescited by 61
  9. 2022From Shapley Values to Generalized Additive Models and backcited by 60
  10. 2013Bounding the Estimation Error of Sampling-based Shapley Value Approximationcited by 58
  11. 2020Feature Interaction Interpretability: A Case for Explaining Ad-Recommendation Systems via Neural Interaction Detectioncited by 32
  12. 2019Quantifying Model Complexity via Functional Decomposition for Better Post-Hoc Interpretabilitycited by 22

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 2; 2024: 17; 2025: 14; 2026: 18.

51 papers.

Sort:

By area:

  1. 2024shapiq: Shapley Interactions for Machine Learningcited by 43
  2. 2024Beyond TreeSHAP: Efficient Computation of Any-Order Shapley Interactions for Tree Ensemblescited by 31
  3. 2024KernelSHAP-IQ: Weighted Least-Square Optimization for Shapley Interactionscited by 22
  4. 2024Stochastic Amortization: A Unified Approach to Accelerate Feature and Data Attributioncited by 21
  5. 2025ProxySPEX: Inference-Efficient Interpretability via Sparse Feature Interactions in LLMscited by 16
  6. 2024Learning to Understand: Identifying Interactions via the M"obius Transformcited by 13
  7. 2025Exact Computation of Any-Order Shapley Interactions for Graph Neural Networkscited by 13
  8. 2025SPEX: Scaling Feature Interaction Explanations for LLMscited by 12
  9. 2024Unifying Feature-Based Explanations with Functional ANOVA and Cooperative Game Theorycited by 10
  10. 2024Explaining Text Similarity in Transformer Modelscited by 9
  11. 2024One Sample Fits All: Approximating All Probabilistic Values Simultaneously and Efficientlycited by 9
  12. 2024Feature Responsiveness Scores: Model-Agnostic Explanations for Recoursecited by 8
38 more
  1. 2025SHAP Meets Tensor Networks: Provably Tractable Explanations with Parallelismcited by 8
  2. 2026PolySHAP: Extending KernelSHAP with Interaction-Informed Polynomial Regressioncited by 7
  3. 2025Explaining Similarity in Vision-Language Encoders with Weighted Banzhaf Interactionscited by 6
  4. 2025Position: We Need An Algorithmic Understanding of Generative AIcited by 5
  5. 2026An Odd Estimator for Shapley Valuescited by 5
  6. 2024Efficient and Accurate Explanation Estimation with Distribution Compressioncited by 4
  7. 2024SHAP zero Explains Biological Sequence Models with Near-zero Marginal Cost for Future Queriescited by 4
  8. 2024Disentangling Interactions and Dependencies in Feature Attributioncited by 4
  9. 2023SurvBeX: An explanation method of the machine learning survival models based on the Beran estimatorcited by 3
  10. 2024Explaining Caption-Image Interactions in CLIP Models with Second-Order Attributionscited by 3
  11. 2026Exactly Computing do-Shapley Valuescited by 3
  12. 2025Shapley Value Approximation Based on k-Additive Gamescited by 2
  13. 2025Adaptive Prompting: Ad-hoc Prompt Composition for Social Bias Detectioncited by 2
  14. 2025EvalxNLP: A Framework for Benchmarking Post-Hoc Explainability Methods on NLP Modelscited by 2
  15. 2026Quadrature-TreeSHAP: Depth-Independent TreeSHAP and Shapley Interactionscited by 2
  16. 2024Using Shapley interactions to understand how models use structurecited by 1
  17. 2024Energy-Based Model for Accurate Estimation of Shapley Values in Feature Attributioncited by 1
  18. 2024Data Science Principles for Interpretable and Explainable AIcited by 1
  19. 2025SIM-Shapley: A Stable and Computationally Efficient Approach to Shapley Value Approximationcited by 1
  20. 2025Higher-Order Feature Attribution: Bridging Statistics, Explainable AI, and Topological Signal Processingcited by 1
  21. 2025Interaction Tensor SHAPcited by 1
  22. 2026Functional Decomposition and Shapley Interactions for Interpreting Survival Modelscited by 1
  23. 2026Adalina: Adaptive Linear Approximation for the Shapley Value and Beyondcited by 1
  24. 2026More Is Not Always Better: Cross-Component Interference in LLM Agent Scaffoldingcited by 1
  25. 2023SurvBeNIM: The Beran-Based Neural Importance Model for Explaining the Survival Modelscited by 0
  26. 2024Can I trust my anomaly detection system? A case study based on explainable AIcited by 0
  27. 2025OrdShap: Feature Position Importance for Sequential Black-Box Modelscited by 0
  28. 2025CUBE: Contrastive Understanding by Balanced Experimentscited by 0
  29. 2026ConfoundingSHAP: Quantifying confounding strength in causal inferencecited by 0
  30. 2026Verified SHAP: Provable Bounds for Exact Shapley Values of Neural Networkscited by 0
  31. 2026TN-SHAP-G: Graph-Structured Tensor Network Surrogates for Shapley Values and Interactionscited by 0
  32. 2026ShaplEIG: Bayesian Experimental Design for Shapley Value Estimationcited by 0
  33. 2026SAILS: Surrogate-based Analysis of Interactions via Local Effect Smoothscited by 0
  34. 2026We Need Explanation Cards to Connect Explanation Algorithms to the Real Worldcited by 0
  35. 2026OperatorSHAP: Fast and Accurate Shapley Value Estimation for Neural Operatorscited by 0
  36. 2026Explaining BiomedCLIP with Weighted Banzhaf Interactions Supported by Tree-Gram Parsingcited by 0
  37. 2026A Hilbert-Valued Functional Decomposition Framework for Explaining Time-Dependent Outputscited by 0
  38. 2026DRSR: Learning Set-Level Deletion Risk for Efficient Long-Horizon Agentscited by 0

50 most cited of 51 All 51 (JSON)