BriefGPT.xyz
Apr, 2022
随机平滑的组合架构 -- 鲁棒和精确
Robust and Accurate -- Compositional Architectures for Randomized Smoothing
HTML
PDF
Miklós Z. Horváth, Mark Niklas Müller, Marc Fischer, Martin Vechev
TL;DR
提出了一种基于ACAS的架构，可以在保证预测准确率的同时，检测每个样本是否存在较大威胁，并提供认证的保护措施。在ImageNet等数据集上，该方法可以同时提供高准确率和可证明的鲁棒性。
Abstract
randomized smoothing
(RS) is considered the state-of-the-art approach to obtain
certifiably robust models
for challenging tasks. However, current RS approaches drastically decrease standard accuracy on unperturbe
→