BriefGPT.xyz
Jun, 2020
随机平滑用于稳健性保证的扩展和限制
Extensions and limitations of randomized smoothing for robustness guarantees
HTML
PDF
Jamie Hayes
TL;DR
通过随机平滑来证明分类器决策对于对抗性噪声不变，同时对噪声稳健性的保证受到多种因素的影响，例如平滑度量之间的差异和拟合威胁模型的选择。此外，该研究证明随着p的增加，随机平滑受到了维度诅咒的影响。
Abstract
randomized smoothing
, a method to certify a classifier's decision on an input is invariant under
adversarial noise
, offers attractive advantages over other
→