← Search

Jeremy D. Bernstein

2 accepted papers

2022

Investigating Generalization by Controlling Normalized Margin

ICML 2022spotlight

Weight norm $\|w\|$ and margin $\gamma$ participate in learning theory via the normalized margin $\gamma/\|w\|$. Since standard neural net optimizers do not control normalized margin, it is hard to test whether this quantity causally relates to generalization. This paper designs a series of experime…

2018

Stochastic Activation Pruning for Robust Adversarial Defense

ICLR 2018poster

Neural networks are known to be vulnerable to adversarial examples. Carefully chosen perturbations to real images, while imperceptible to humans, induce misclassification and threaten the reliability of deep learning systems in the wild. To guard against adversarial examples, we take inspiration fro…