BriefGPT.xyz
Dec, 2014
面向对抗样本具鲁棒性的深度神经网络架构
Towards Deep Neural Network Architectures Robust to Adversarial Examples
HTML
PDF
Shixiang Gu, Luca Rigazio
TL;DR
该研究论文研究了深度神经网络的鲁棒性问题,特别是针对对抗样本的攻击。通过探索神经网络的结构,拓扑结构,预处理和训练策略等方面来提高深度神经网络的抗干扰能力,并且通过引入平滑性惩罚来提高其稳健性。
Abstract
Recent work has shown
deep neural networks
(DNNs) to be highly susceptible to well-designed, small perturbations at the input layer, or so-called
adversarial examples
. Taking images as an example, such distortion
→