Tangentially Aligned Integrated Gradients for User-Friendly Explanations
This work addresses the need for more reliable and user-friendly explanations in neural networks, particularly for image classification, though it is incremental as it builds on existing Integrated Gradients with a new base-point selection criterion.
The paper tackles the problem of inconsistent explanations from Integrated Gradients due to arbitrary base-point choices by proposing a method to select base-points that maximize tangential alignment, assuming data lies on a low-dimensional manifold. They demonstrate this approach on image datasets, showing improved explanations compared to common base-points and other gradient models.
Integrated gradients is prevalent within machine learning to address the black-box problem of neural networks. The explanations given by integrated gradients depend on a choice of base-point. The choice of base-point is not a priori obvious and can lead to drastically different explanations. There is a longstanding hypothesis that data lies on a low dimensional Riemannian manifold. The quality of explanations on a manifold can be measured by the extent to which an explanation for a point lies in its tangent space. In this work, we propose that the base-point should be chosen such that it maximises the tangential alignment of the explanation. We formalise the notion of tangential alignment and provide theoretical conditions under which a base-point choice will provide explanations lying in the tangent space. We demonstrate how to approximate the optimal base-point on several well-known image classification datasets. Furthermore, we compare the optimal base-point choice with common base-points and three gradient explainability models.