LGAICVJun 13, 2021

Boosting Randomized Smoothing with Variance Reduced Classifiers

arXiv:2106.06946v358 citationsHas Code
Originality Incremental advance
AI Analysis

This work addresses the computational efficiency and robustness certification problem in machine learning security, offering incremental improvements with practical optimizations.

The paper tackled improving Randomized Smoothing for robustness certificates by using ensembles as base models, resulting in state-of-the-art average certified radii of 0.86 on CIFAR10 and 1.11 on ImageNet, with up to 55-fold reduction in sample complexity.

Randomized Smoothing (RS) is a promising method for obtaining robustness certificates by evaluating a base model under noise. In this work, we: (i) theoretically motivate why ensembles are a particularly suitable choice as base models for RS, and (ii) empirically confirm this choice, obtaining state-of-the-art results in multiple settings. The key insight of our work is that the reduced variance of ensembles over the perturbations introduced in RS leads to significantly more consistent classifications for a given input. This, in turn, leads to substantially increased certifiable radii for samples close to the decision boundary. Additionally, we introduce key optimizations which enable an up to 55-fold decrease in sample complexity of RS for predetermined radii, thus drastically reducing its computational overhead. Experimentally, we show that ensembles of only 3 to 10 classifiers consistently improve on their strongest constituting model with respect to their average certified radius (ACR) by 5% to 21% on both CIFAR10 and ImageNet, achieving a new state-of-the-art ACR of 0.86 and 1.11, respectively. We release all code and models required to reproduce our results at https://github.com/eth-sri/smoothing-ensembles.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes