Timezone: »

Black-Box Certification with Randomized Smoothing: A Functional Optimization Based Framework
Dinghuai Zhang · Mao Ye · Chengyue Gong · Zhanxing Zhu · Qiang Liu

Tue Dec 08 09:00 PM -- 11:00 PM (PST) @ Poster Session 2 #705
Randomized classifiers have been shown to provide a promising approach for achieving certified robustness against adversarial attacks in deep learning. However, most existing methods only leverage Gaussian smoothing noise and only work for $\ell_2$ perturbation. We propose a general framework of adversarial certification with non-Gaussian noise and for more general types of attacks, from a unified \functional optimization perspective. Our new framework allows us to identify a key trade-off between accuracy and robustness via designing smoothing distributions, helping to design new families of non-Gaussian smoothing distributions that work more efficiently for different $\ell_p$ settings, including $\ell_1$, $\ell_2$ and $\ell_\infty$ attacks. Our proposed methods achieve better certification results than previous works and provide a new perspective on randomized smoothing certification.

Author Information

Dinghuai Zhang (Mila / Peking University)
Mao Ye (The University of Texas at Austin)
Chengyue Gong (Peking University)
Zhanxing Zhu (Peking University)
Qiang Liu (UT Austin)

More from the Same Authors