Calibrated Adversarial Training

Abstract

Adversarial training is an approach of increasing the robustness of models to adversarial attacks by including adversarial examples in the training set. One major challenge of producing adversarial examples is to contain sufficient perturbation in the example to flip the model's output while not making severe changes in the example's semantical content. Exuberant change in the semantical content could also change the true label of the example. Adding such examples to the training set results in adverse effects. In this paper, we present the Calibrated Adversarial Training, a method that reduces the adverse effects of semantic perturbations in adversarial training. The method produces pixel-level adaptations to the perturbations based on novel calibrated robust error. We provide theoretical analysis on the calibrated robust error and derive an upper bound for it. Our empirical results show a superior performance of the Calibrated Adversarial Training over a number of public datasets.

arXiv commentACML 2021 accepted,24 pages

References in corpus (10)

• Explaining and Harnessing Adversarial Examples
• Intriguing properties of neural networks
• Wide Residual Networks
• Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples
• Provable defenses against adversarial examples via the convex outer adversarial polytope
• Certified Adversarial Robustness via Randomized Smoothing
• On the Convergence and Robustness of Adversarial Training
• MMA Training: Direct Input Space Margin Maximization through Adversarial Training
• Instance adaptive adversarial training: Improved accuracy tradeoffs in neural nets
• Bridging the Performance Gap between FGSM and PGD Adversarial Training