Adversarial Weight Perturbation Helps Robust Generalization
This work addresses the challenge of robust generalization against adversarial examples for deep learning practitioners, offering an incremental improvement by building on existing adversarial training techniques.
The paper tackles the problem of improving adversarial robustness in deep neural networks by identifying a correlation between the flatness of the weight loss landscape and robust generalization gap, and proposes Adversarial Weight Perturbation (AWP) to explicitly regularize this flatness, which boosts adversarial robustness when incorporated into existing methods.
The study on improving the robustness of deep neural networks against adversarial examples grows rapidly in recent years. Among them, adversarial training is the most promising one, which flattens the input loss landscape (loss change with respect to input) via training on adversarially perturbed examples. However, how the widely used weight loss landscape (loss change with respect to weight) performs in adversarial training is rarely explored. In this paper, we investigate the weight loss landscape from a new perspective, and identify a clear correlation between the flatness of weight loss landscape and robust generalization gap. Several well-recognized adversarial training improvements, such as early stopping, designing new objective functions, or leveraging unlabeled data, all implicitly flatten the weight loss landscape. Based on these observations, we propose a simple yet effective Adversarial Weight Perturbation (AWP) to explicitly regularize the flatness of weight loss landscape, forming a double-perturbation mechanism in the adversarial training framework that adversarially perturbs both inputs and weights. Extensive experiments demonstrate that AWP indeed brings flatter weight loss landscape and can be easily incorporated into various existing adversarial training methods to further boost their adversarial robustness.