shapley values
A concept from cooperative game theory applied to ML that assigns a value to each feature based on its contribution to a model's output, helping interpret model predictions and the importance of inputs.
- A Unified Framework for Provably Efficient Algorithms to Estimate Shapley Values
- Approximating Shapley Explanations in Reinforcement Learning
- Proxy-SPEX: Sample-Efficient Interpretability via Sparse Feature Interactions in LLMs
- Regression-adjusted Monte Carlo Estimators for Shapley Values and Probabilistic Values
- SHAP zero Explains Biological Sequence Models with Near-zero Marginal Cost for Future Queries
- ShapeX: Shapelet-Driven Post Hoc Explanations for Time Series Classification Models
- Shapley-Based Data Valuation for Weighted $k$-Nearest Neighbors
- Tree Ensemble Explainability through the Hoeffding Functional Decomposition and TreeHFD Algorithm