Geometric Remove-and-Retrain (GOAR): Coordinate-Invariant eXplainable AI Assessment

TL;DR

This paper introduces GOAR, a geometric, coordinate-invariant method for evaluating feature importance in explainable AI, overcoming limitations of traditional pixel-perturbation approaches.

Contribution

The paper proposes GOAR, a novel geometric approach for assessing feature importance that is coordinate-invariant and more reliable than existing pixel-based methods.

Findings

GOAR outperforms pixel-perturbation metrics in synthetic and real datasets.

GOAR provides more reliable and geometric-aware feature importance assessments.

Traditional pixel-based metrics fail to discriminate between feature attribution methods geometrically.

Abstract

Identifying the relevant input features that have a critical influence on the output results is indispensable for the development of explainable artificial intelligence (XAI). Remove-and-Retrain (ROAR) is a widely accepted approach for assessing the importance of individual pixels by measuring changes in accuracy following their removal and subsequent retraining of the modified dataset. However, we uncover notable limitations in pixel-perturbation strategies. When viewed from a geometric perspective, we discover that these metrics fail to discriminate between differences among feature attribution methods, thereby compromising the reliability of the evaluation. To address this challenge, we introduce an alternative feature-perturbation approach named Geometric Remove-and-Retrain (GOAR). Through a series of experiments with both synthetic and real datasets, we substantiate that GOAR transcends the limitations of pixel-centric metrics.