BriefGPT.xyz
Sep, 2018
通过诱导ReLU稳定性进行更快的对抗鲁棒性验证训练
Training for Faster Adversarial Robustness Verification via Inducing ReLU Stability
HTML
PDF
Kai Y. Xiao, Vincent Tjeng, Nur Muhammad Shafiullah, Aleksander Madry
TL;DR
本篇论文研究神经网络验证中的协同设计概念，并通过改进权重稀疏性和ReLU稳定性的方法，将计算困难的验证问题转化为可处理的问题，并改善了验证的速度，该方法具有普适性。
Abstract
We explore the concept of
co-design
in the context of
neural network verification
. Specifically, we aim to train deep neural networks that not only are robust to adversarial perturbations but also whose
→