Towards Understanding Clean Generalization and Robust Overfitting in Adversarial Training

원문 초록

Similar to surprising performance in the standard deep learning, deep nets trained by adversarial training also generalize well for $\textit{unseen clean data (natural data)}$. However, despite adversarial training can achieve low robust training error, there exists a significant $\textit{robust generalization gap}$. We call this phenomenon the $\textit{Clean Generalization and Robust Overfitting (CGRO)}$. In this work, we study the CGRO phenomenon in adversarial training from two views: $\textit{representation complexity}$ and $\textit{training dynamics}$. Specifically, we consider a binary classification setting with $N$ separated training data points. $\textit{First}$, we prove that, based on the assumption that we assume there is $\operatorname{poly}(D)$-size clean classifier (where $D$ is the data dimension), ReLU net with only $O(N D)$ extra parameters is able to leverages robust memorization to achieve the CGRO, while robust classifier still requires exponential representation complexity in worst case. $\textit{Next}$, we focus on a structured-data case to analyze training dynamics, where we train a two-layer convolutional network with $O(N D)$ width against adversarial perturbation. We then show that a three-stage phase transition occurs during learning process and the network provably converges to robust memorization regime, which thereby results in the CGRO. $\textit{Besides}$, we also empirically verify our theoretical analysis by experiments in real-image recognition datasets.