Rectified Gradient: Layer-wise Thresholding for Sharp and Coherent Attribution Maps

Sep 27, 2018 ICLR 2019 Conference Blind Submission readers: everyone Show Bibtex
  • Abstract: Saliency map, or the gradient of the score function with respect to the input, is the most basic means of interpreting deep neural network decisions. However, saliency maps are often visually noisy. Although several hypotheses were proposed to account for this phenomenon, there is no work that provides a rigorous analysis of noisy saliency maps. This may be a problem as numerous advanced attribution methods were proposed under the assumption that the existing hypotheses are true. In this paper, we identify the cause of noisy saliency maps. Then, we propose Rectified Gradient, a simple method that significantly improves saliency maps by alleviating that cause. Experiments showed effectiveness of our method and its superiority to other attribution methods. Codes and examples for the experiments will be released in public.
  • Keywords: Interpretability, Attribution Method, Attribution Map
  • TL;DR: We propose a new attribution method that removes noise from saliency maps through layer-wise thresholding during backpropagation.
0 Replies