BriefGPT.xyz
May, 2021
通过对扰动样本进行对抗标记的鲁棒性正则化
Robust Regularization with Adversarial Labelling of Perturbed Samples
HTML
PDF
Xiaohui Guo, Richong Zhang, Yaowei Zheng, Yongyi Mao
TL;DR
本文提出了一种称为 ALPS 的正则化方案,通过向每个真实输入样本添加扰动样本并给予对抗性标签来训练神经网络,以提高其泛化能力和对抗鲁棒性,实验表明 ALPS 具有先进的正则化性能和有效的对抗性训练。
Abstract
Recent researches have suggested that the predictive accuracy of
neural network
may contend with its
adversarial robustness
. This presents challenges in designing effective regularization schemes that also provid
→