Certified Training: Small Boxes Are All You Need
Abstract
We propose the novel certified training method, SABR, which outperforms existing methods across perturbation magnitudes on MNIST, CIFAR-10, and TinyImageNet, in terms of both standard and certifiable accuracies. The key insight behind SABR is that propagating interval bounds for a small but carefully selected subset of the adversarial input region is sufficient to approximate the worst-case loss over the whole region while significantly reducing approximation errors. SABR does not only establish a new state-of-the-art in all commonly used benchmarks but more importantly, points to a new class of certified training methods promising to overcome the robustness-accuracy trade-off.
Cite
Text
Mueller et al. "Certified Training: Small Boxes Are All You Need." NeurIPS 2022 Workshops: TSRML, 2022.Markdown
[Mueller et al. "Certified Training: Small Boxes Are All You Need." NeurIPS 2022 Workshops: TSRML, 2022.](https://mlanthology.org/neuripsw/2022/mueller2022neuripsw-certified/)BibTeX
@inproceedings{mueller2022neuripsw-certified,
title = {{Certified Training: Small Boxes Are All You Need}},
author = {Mueller, Mark Niklas and Eckert, Franziska and Fischer, Marc and Vechev, Martin},
booktitle = {NeurIPS 2022 Workshops: TSRML},
year = {2022},
url = {https://mlanthology.org/neuripsw/2022/mueller2022neuripsw-certified/}
}