Saliency strikes back: How filtering out high frequencies improves white-box explanations

TL;DR

The paper introduces FORGrad, a filtering technique that removes high-frequency artifacts from white-box attribution methods, significantly improving their accuracy and competitiveness with black-box explanations.

Contribution

We propose FORGrad, a novel filtering approach that enhances white-box explainability methods by reducing high-frequency noise, making them more accurate and computationally efficient.

Findings

FORGrad improves white-box explanation accuracy.

Enhanced white-box methods rival black-box approaches.

Filtering high frequencies benefits model interpretability.

Abstract

Attribution methods correspond to a class of explainability methods (XAI) that aim to assess how individual inputs contribute to a model's decision-making process. We have identified a significant limitation in one type of attribution methods, known as ``white-box" methods. Although highly efficient, as we will show, these methods rely on a gradient signal that is often contaminated by high-frequency artifacts. To overcome this limitation, we introduce a new approach called "FORGrad". This simple method effectively filters out these high-frequency artifacts using optimal cut-off frequencies tailored to the unique characteristics of each model architecture. Our findings show that FORGrad consistently enhances the performance of already existing white-box methods, enabling them to compete effectively with more accurate yet computationally demanding "black-box" methods. We anticipate that our research will foster broader adoption of simpler and more efficient white-box methods for explainability, offering a better balance between faithfulness and computational efficiency.