Why are Saliency Maps Noisy? Cause of and Solution to Noisy Saliency Maps

View Researcher's Other Codes

Disclaimer: The provided code links for this paper are external links. Science Nest has no responsibility for the accuracy, legality or content of these links. Also, by downloading this code(s), you agree to comply with the terms of use as set out by the author(s) of the code(s).

Please contact us in case of a broken link from here

Authors Beomsu Kim, Junghoon Seo, Taegyun Jeon, Jeongyeol Choe, Jamyoung Koo, SeungHyun Jeon
Journal/Conference Name Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019
Paper Category
Paper Abstract Saliency Map, the gradient of the score function with respect to the input, is the most basic technique for interpreting deep neural network decisions. However, saliency maps are often visually noisy. Although several hypotheses were proposed to account for this phenomenon, there are few works that provide rigorous analyses of noisy saliency maps. In this paper, we firstly propose a new hypothesis that noise may occur in saliency maps when irrelevant features pass through ReLU activation functions. Then, we propose Rectified Gradient, a method that alleviates this problem through layer-wise thresholding during backpropagation. Experiments with neural networks trained on CIFAR-10 and ImageNet showed effectiveness of our method and its superiority to other attribution methods.
Date of publication 2019
Code Programming Language Jupyter Notebook

Copyright Researcher 2022