SHAP values via sparse Fourier representation

TL;DR

This paper introduces a fast, Fourier-based method for computing SHAP values that works efficiently for both black-box and tree models, enabling scalable and approximate explanations in AI.

Contribution

It presents a novel two-stage Fourier-based approach for efficient SHAP value computation, including a closed-form formula for exact calculation using spectral representations.

Findings

Achieves significant speedups over existing methods.

Provides a tunable trade-off between efficiency and accuracy.

Enables amortized SHAP value computation for large models.

Abstract

SHAP (SHapley Additive exPlanations) values are a widely used method for local feature attribution in interpretable and explainable AI. We propose an efficient two-stage algorithm for computing SHAP values in both black-box setting and tree-based models. Motivated by spectral bias in real-world predictors, we first approximate models using compact Fourier representations, exactly for trees and approximately for black-box models. In the second stage, we introduce a closed-form formula for {\em exactly} computing SHAP values using the Fourier representation, that ``linearizes'' the computation into a simple summation and is amenable to parallelization. As the Fourier approximation is computed only once, our method enables amortized SHAP value computation, achieving significant speedups over existing methods and a tunable trade-off between efficiency and precision.