IDGI: A Framework to Eliminate Explanation Noise from Integrated Gradients

2023/03/24 by Yang, Ruo, Wang, Binghui, Bilgic, Mustafa · 1 citation
#Artificial Intelligence (cs.AI) #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Machine Learning (cs.LG)

paper · doi:10.48550/arxiv.2303.14242

Abstract

Integrated Gradients (IG) as well as its variants are well-known techniques for interpreting the decisions of deep neural networks. While IG-based approaches attain state-of-the-art performance, they often integrate noise into their explanation saliency maps, which reduce their interpretability. To minimize the noise, we examine the source of the noise analytically and propose a new approach to reduce the explanation noise based on our analytical findings. We propose the Important Direction Gradient Integration (IDGI) framework, which can be easily incorporated into any IG-based method that uses the Reimann Integration for integrated gradient computation. Extensive experiments with three IG-based methods show that IDGI improves them drastically on numerous interpretability metrics.

Cited by

• Enhancing Interpretability for Vision Models via Shapley Value Optimization

Related

• Integrated Decision Gradients: Compute Your Attributions Where the Model Makes Its Decision
• Strengthening Interpretability: An Investigative Study of Integrated Gradient Methods
• Geometrically Guided Integrated Gradients
• Pattern-Guided Integrated Gradients
• A Rigorous Study of Integrated Gradients Method and Extensions to Internal Neuron Attributions