Breaking Certifiable Defenses

This repository contains all the code for experiments in the Breaking Certifiable Defenses paper. To reproduce all the resutls, please check out each of the folders corresponding to each of the experimetns. Also to cite this work, please use:

@inproceedings{
Ghiasi2020BREAKING,
title={BREAKING  CERTIFIED  DEFENSES:  SEMANTIC  ADVERSARIAL  EXAMPLES  WITH  SPOOFED  ROBUSTNESS  CERTIFICATES},
author={Amin Ghiasi and Ali Shafahi and Tom Goldstein},
booktitle={International Conference on Learning Representations},
year={2020},
url={https://openreview.net/forum?id=HJxdTxHYvB}
}