DeepAI AI Chat
Log In Sign Up

Intriguing Properties of Input-dependent Randomized Smoothing

by   Peter Súkeník, et al.

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 to overcome these flaws. However, we demonstrate that these methods lack formal guarantees and so the resulting certificates are not justified. We show that the input-dependent smoothing, in general, 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 and test it on CIFAR10 and MNIST. Our design solves some of the problems of classical smoothing and is formally underlined, yet further improvement of the design is still necessary.


page 4

page 15


Data Dependent Randomized Smoothing

Randomized smoothing is a recent technique that achieves state-of-art pe...

Center Smoothing for Certifiably Robust Vector-Valued Functions

Randomized smoothing has been successfully applied in high-dimensional i...

Riemannian data-dependent randomized smoothing for neural networks certification

Certification of neural networks is an important and challenging problem...

Improving the Accuracy-Robustness Trade-off of Classifiers via Adaptive Smoothing

While it is shown in the literature that simultaneously accurate and rob...

Higher-Order Certification for Randomized Smoothing

Randomized smoothing is a recently proposed defense against adversarial ...

Scalable Certified Segmentation via Randomized Smoothing

We present a new certification method for image and point cloud segmenta...

DeformRS: Certifying Input Deformations with Randomized Smoothing

Deep neural networks are vulnerable to input deformations in the form of...