Assessing Fidelity in XAI post-hoc techniques: A Comparative Study with Ground Truth Explanations Datasets

TL;DR

This study compares state-of-the-art XAI methods using new datasets with ground truth explanations, revealing that backpropagation-based methods are more accurate but noisier, guiding future development of trustworthy XAI techniques.

Contribution

Introduces three novel datasets with ground truth explanations for fair evaluation of XAI methods and identifies the superior fidelity of backpropagation-based techniques.

Findings

Backpropagation-based XAI methods outperform sensitivity analysis and CAM in fidelity.

Backpropagation methods produce more accurate explanations but are noisier.

Eliminating low-fidelity methods can improve XAI reliability.

Abstract

The evaluation of the fidelity of eXplainable Artificial Intelligence (XAI) methods to their underlying models is a challenging task, primarily due to the absence of a ground truth for explanations. However, assessing fidelity is a necessary step for ensuring a correct XAI methodology. In this study, we conduct a fair and objective comparison of the current state-of-the-art XAI methods by introducing three novel image datasets with reliable ground truth for explanations. The primary objective of this comparison is to identify methods with low fidelity and eliminate them from further research, thereby promoting the development of more trustworthy and effective XAI techniques. Our results demonstrate that XAI methods based on the backpropagation of output information to input yield higher accuracy and reliability compared to methods relying on sensitivity analysis or Class Activation Maps (CAM). However, the backpropagation method tends to generate more noisy saliency maps. These findings have significant implications for the advancement of XAI methods, enabling the elimination of erroneous explanations and fostering the development of more robust and reliable XAI.