Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
OTAD: An Optimal Transport-Induced Robust Model for Agnostic Adversarial Attack
Kuo Gai, Sicong Wang, Shihua Zhang
Deep neural networks (DNNs) are vulnerable to small adversarial perturbations of the inputs, posing a significant challenge to their reliability and robustness. Empirical methods s…
cs.LG2024
Progressive Feedforward Collapse of ResNet Training
Sicong Wang, Kuo Gai, Shihua Zhang
Neural collapse (NC) is a simple and symmetric phenomenon for deep neural networks (DNNs) at the terminal phase of training, where the last-layer features collapse to their class m…