BriefGPT.xyz
Sep, 2022
对抗鲁棒性与架构组件的相互作用：补丁、卷积和注意力
On the interplay of adversarial robustness and architecture components: patches, convolution and attention
HTML
PDF
Francesco Croce, Matthias Hein
TL;DR
该研究比较了多种不同架构的分类器，通过对其对抗训练的表现及可解释性下降等性质进行研究，揭示了ResNet与ConvNeXt等构架变化与模型对抗鲁棒性之间的关系。
Abstract
In recent years novel
architecture components
for
image classification
have been developed, starting with attention and patches used in transformers. While prior works have analyzed the influence of some aspects
→