BriefGPT.xyz
Apr, 2021
稀疏编码前端用于鲁棒神经网络
Sparse Coding Frontend for Robust Neural Networks
HTML
PDF
Can Bakiskan, Metehan Cekic, Ahmet Dundar Sezer, Upamanyu Madhow
TL;DR
该研究提出了一种基于稀疏编码的前端防御方法，该方法在CIFAR-10数据集上的实验结果表明，相较于其它对抗训练方法，该方法对Linf、L2和L1攻击的抵御效果更好。
Abstract
deep neural networks
are known to be vulnerable to small, adversarially crafted, perturbations. The current most effective
defense methods
against these
→