2 code implementations • NeurIPS 2019 • Zhonghui You, Kun Yan, Jinmian Ye, Meng Ma, Ping Wang
When the scaling factor is set to zero, it is equivalent to removing the corresponding filter.
1 code implementation • 21 May 2018 • Zhonghui You, Jinmian Ye, Kunming Li, Zenglin Xu, Ping Wang
In this paper, we introduce a novel regularization method called Adversarial Noise Layer (ANL) and its efficient version called Class Adversarial Noise Layer (CANL), which are able to significantly improve CNN's generalization ability by adding carefully crafted noise into the intermediate layer activations.