Mitigating Advanced Adversarial Attacks with More Advanced Gradient Obfuscation Techniques
This addresses a critical security issue in AI systems by providing a defense against powerful adversarial attacks, though it is incremental as it builds on existing defense efforts.
The paper tackles the problem of defending deep neural networks against advanced gradient-based adversarial attacks by proposing preprocessing functions that invalidate these attacks, achieving an attack success rate below 7% for the strongest attacks.
Deep Neural Networks (DNNs) are well-known to be vulnerable to Adversarial Examples (AEs). A large amount of efforts have been spent to launch and heat the arms race between the attackers and defenders. Recently, advanced gradient-based attack techniques were proposed (e.g., BPDA and EOT), which have defeated a considerable number of existing defense methods. Up to today, there are still no satisfactory solutions that can effectively and efficiently defend against those attacks. In this paper, we make a steady step towards mitigating those advanced gradient-based attacks with two major contributions. First, we perform an in-depth analysis about the root causes of those attacks, and propose four properties that can break the fundamental assumptions of those attacks. Second, we identify a set of operations that can meet those properties. By integrating these operations, we design two preprocessing functions that can invalidate these powerful attacks. Extensive evaluations indicate that our solutions can effectively mitigate all existing standard and advanced attack techniques, and beat 11 state-of-the-art defense solutions published in top-tier conferences over the past 2 years. The defender can employ our solutions to constrain the attack success rate below 7% for the strongest attacks even the adversary has spent dozens of GPU hours.