Intriguing Properties of Input-dependent Randomized Smoothing

Abstract

Randomized smoothing is currently considered the state-of-the-art method to obtain certifiably robust classifiers. Despite its remarkable performance, the method is associated with various serious problems such as "certified accuracy waterfalls", certification vs.\ accuracy trade-off, or even fairness issues. Input-dependent smoothing approaches have been proposed with intention of overcoming these flaws. However, we demonstrate that these methods lack formal guarantees and so the resulting certificates are not justified. We show that in general, the input-dependent smoothing suffers from the curse of dimensionality, forcing the variance function to have low semi-elasticity. On the other hand, we provide a theoretical and practical framework that enables the usage of input-dependent smoothing even in the presence of the curse of dimensionality, under strict restrictions. We present one concrete design of the smoothing variance function and test it on CIFAR10 and MNIST. Our design mitigates some of the problems of classical smoothing and is formally underlined, yet further improvement of the design is still necessary.

References in corpus (15)

• Intriguing properties of neural networks
• Rényi Divergence and Kullback-Leibler Divergence
• Towards Deep Learning Models Resistant to Adversarial Attacks
• Provable defenses against adversarial examples via the convex outer adversarial polytope
• Certified Adversarial Robustness via Randomized Smoothing
• Certified Defenses against Adversarial Examples
• Formal Guarantees on the Robustness of a Classifier against Adversarial Manipulation
• Provably Robust Deep Learning via Adversarially Trained Smoothed Classifiers
• Certified Adversarial Robustness with Additive Noise
• On the Robustness of the CVPR 2018 White-Box Adversarial Example Defenses
• MACER: Attack-free and Scalable Robust Training via Maximizing Certified Radius
• Data-Dependent Randomized Smoothing
• Analyzing Accuracy Loss in Randomized Smoothing Defenses
• Insta-RS: Instance-wise Randomized Smoothing for Improved Robustness and Accuracy
• ANCER: Anisotropic Certification via Sample-wise Volume Maximization