Summary
International Symposium on Nonlinear Theory and its Applications
2017
Session Number:A1L-C
Session:
Number:A1L-C-4
A Fast Test Method for Noise Robustness of Deep Neural Networks
Muneki Yasuda, Hironori Sakata, Seungil Cho, Tomochika Harada, Atushi Tanaka, Michio Yokoyama,
pp.46-49
Publication Date:2017/12/4
Online ISSN:2188-5079
Summary:
In this paper, we propose a fast test method for noise robustness of pattern recognition systems based on deep neural networks. The proposed method enable us to compare the noise robustnesses of different models and can be applicable to any type of deep neural networks. We demonstrate the validity of our method in the numerical experiments for MNIST data set and our sleep data set.