Timezone: »

 
Poster
Curse of Dimensionality on Randomized Smoothing for Certifiable Robustness
Aounon Kumar · Alexander Levine · Tom Goldstein · Soheil Feizi

Thu Jul 16 03:00 PM -- 03:45 PM & Fri Jul 17 04:00 AM -- 04:45 AM (PDT) @
in Poster Session 48 »
Randomized smoothing, using just a simple isotropic Gaussian distribution, has been shown to produce good robustness guarantees against $\ell_2$-norm bounded adversaries. In this work, we show that extending the smoothing technique to defend against other attack models can be challenging, especially in the high-dimensional regime. In particular, for a vast class of i.i.d.~smoothing distributions, we prove that the largest $\ell_p$-radius that can be certified decreases as $O(1/d^{\frac{1}{2} - \frac{1}{p}})$ with dimension $d$ for $p > 2$. Notably, for $p \geq 2$, this dependence on $d$ is no better than that of the $\ell_p$-radius that can be certified using isotropic Gaussian smoothing, essentially putting a matching lower bound on the robustness radius. When restricted to {\it generalized} Gaussian smoothing, these two bounds can be shown to be within a constant factor of each other in an asymptotic sense, establishing that Gaussian smoothing provides the best possible results, up to a constant factor, when $p \geq 2$. We present experimental results on CIFAR to validate our theory. For other smoothing distributions, such as, a uniform distribution within an $\ell_1$ or an $\ell_\infty$-norm ball, we show upper bounds of the form $O(1 / d)$ and $O(1 / d^{1 - \frac{1}{p}})$ respectively, which have an even worse dependence on $d$.
Keywords: Adversarial Examples · Deep Learning Theory · Robust Statistics and Machine Learning

Author Information

Aounon Kumar (University of Maryland, College Park)
Alexander Levine (University of Maryland)
Tom Goldstein (University of Maryland)
Soheil Feizi (University of Maryland)

More from the Same Authors