Derivative-based Shapley value for global sensitivity analysis and machine learning explainability
This work addresses the computational bottleneck in explainable AI for practitioners, though it is incremental as it builds on existing Shapley value frameworks.
The authors tackled the computational inefficiency of Shapley value methods for sensitivity analysis and ML explainability by introducing a derivative-based approach that reduces complexity from exponential to linear in dimension, demonstrating its effectiveness through numerical comparisons with existing methods like SHAP and KernelSHAP.
We introduce a new Shapley value approach for global sensitivity analysis and machine learning explainability. The method is based on the first-order partial derivatives of the underlying function. The computational complexity of the method is linear in dimension (number of features), as opposed to the exponential complexity of other Shapley value approaches in the literature. Examples from global sensitivity analysis and machine learning are used to compare the method numerically with activity scores, SHAP, and KernelSHAP.