BriefGPT.xyz
Feb, 2021
对抗训练在对抗鲁棒性方面的最新进展
Recent Advances in Adversarial Training for Adversarial Robustness
HTML
PDF
Tao Bai, Jinqi Luo, Jun Zhao, Bihan Wen
TL;DR
该研究论文系统地回顾了针对深度学习模型的对抗训练在对抗鲁棒性方面的最新进展，并从三个视角讨论了对抗训练中的泛化问题，同时指出了尚未完全解决的挑战并提出潜在的未来研究方向。
Abstract
adversarial examples
for fooling
deep learning
models have been studied for several years and are still a hot topic.
adversarial training
→