Prediction via Shapley Value Regression

TL;DR

This paper introduces ViaSHAP, a novel method that learns to compute Shapley values directly, enabling faster and more accurate explanations of black-box model predictions for tabular data and images.

Contribution

ViaSHAP is a new approach that learns to compute Shapley values using neural networks, reducing computational costs and improving explanation accuracy.

Findings

ViaSHAP with Kolmogorov-Arnold Networks matches state-of-the-art performance.

ViaSHAP explanations are more accurate than FastSHAP.

The method is effective for both tabular data and images.

Abstract

Shapley values have several desirable, theoretically well-supported, properties for explaining black-box model predictions. Traditionally, Shapley values are computed post-hoc, leading to additional computational cost at inference time. To overcome this, a novel method, called ViaSHAP, is proposed, that learns a function to compute Shapley values, from which the predictions can be derived directly by summation. Two approaches to implement the proposed method are explored; one based on the universal approximation theorem and the other on the Kolmogorov-Arnold representation theorem. Results from a large-scale empirical investigation are presented, showing that ViaSHAP using Kolmogorov-Arnold Networks performs on par with state-of-the-art algorithms for tabular data. It is also shown that the explanations of ViaSHAP are significantly more accurate than the popular approximator FastSHAP on both tabular data and images.