One Wave To Explain Them All: A Unifying Perspective On Feature Attribution

TL;DR

This paper introduces the Wavelet Attribution Method (WAM), a unified approach for feature attribution that uses wavelet transforms to provide meaningful explanations across various data modalities, improving transparency and robustness.

Contribution

The paper proposes WAM, a novel wavelet-based attribution technique that captures both spatial and scale information, unifying feature attribution across multiple data types.

Findings

WAM outperforms existing gradient-based methods on multiple modalities.

WAM provides more meaningful and interpretable attributions.

WAM links feature attribution with model robustness and transparency.

Abstract

Feature attribution methods aim to improve the transparency of deep neural networks by identifying the input features that influence a model's decision. Pixel-based heatmaps have become the standard for attributing features to high-dimensional inputs, such as images, audio representations, and volumes. While intuitive and convenient, these pixel-based attributions fail to capture the underlying structure of the data. Moreover, the choice of domain for computing attributions has often been overlooked. This work demonstrates that the wavelet domain allows for informative and meaningful attributions. It handles any input dimension and offers a unified approach to feature attribution. Our method, the Wavelet Attribution Method (WAM), leverages the spatial and scale-localized properties of wavelet coefficients to provide explanations that capture both the where and what of a model's decision-making process. We show that WAM quantitatively matches or outperforms existing gradient-based methods across multiple modalities, including audio, images, and volumes. Additionally, we discuss how WAM bridges attribution with broader aspects of model robustness and transparency. Project page: https://gabrielkasmi.github.io/wam/