Enhancing Certified Robustness via Smoothed Weighted Ensembling

TL;DR

This paper introduces SWEEN, a weighted ensembling method that enhances the certified robustness of randomized smoothing classifiers against adversarial attacks, achieving better performance and efficiency.

Contribution

The paper proposes a novel SWEEN ensembling scheme that optimally combines base classifiers to improve certified robustness in randomized smoothing, with theoretical guarantees and practical algorithms.

Findings

SWEEN outperforms individual models significantly in certified robustness.

Small models combined via SWEEN match large model performance with less training time.

Adaptive prediction reduces certification costs effectively.

Abstract

Randomized smoothing has achieved state-of-the-art certified robustness against $l_2$-norm adversarial attacks. However, it is not wholly resolved on how to find the optimal base classifier for randomized smoothing. In this work, we employ a Smoothed WEighted ENsembling (SWEEN) scheme to improve the performance of randomized smoothed classifiers. We show the ensembling generality that SWEEN can help achieve optimal certified robustness. Furthermore, theoretical analysis proves that the optimal SWEEN model can be obtained from training under mild assumptions. We also develop an adaptive prediction algorithm to reduce the prediction and certification cost of SWEEN models. Extensive experiments show that SWEEN models outperform the upper envelope of their corresponding candidate models by a large margin. Moreover, SWEEN models constructed using a few small models can achieve comparable performance to a single large model with a notable reduction in training time.