On ℓ_p-norm Robustness of Ensemble Stumps and Trees

08/20/2020 ∙ by Yihan Wang, et al. ∙ 0

Recent papers have demonstrated that ensemble stumps and trees could be vulnerable to small input perturbations, so robustness verification and defense for those models have become an important research problem. However, due to the structure of decision trees, where each node makes decision purely based on one feature value, all the previous works only consider the ℓ_∞ norm perturbation. To study robustness with respect to a general ℓ_p norm perturbation, one has to consider the correlation between perturbations on different features, which has not been handled by previous algorithms. In this paper, we study the problem of robustness verification and certified defense with respect to general ℓ_p norm perturbations for ensemble decision stumps and trees. For robustness verification of ensemble stumps, we prove that complete verification is NP-complete for p∈(0, ∞) while polynomial time algorithms exist for p=0 or ∞. For p∈(0, ∞) we develop an efficient dynamic programming based algorithm for sound verification of ensemble stumps. For ensemble trees, we generalize the previous multi-level robustness verification algorithm to ℓ_p norm. We demonstrate the first certified defense method for training ensemble stumps and trees with respect to ℓ_p norm perturbations, and verify its effectiveness empirically on real datasets.



There are no comments yet.


page 1

page 2

page 3

page 4

This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.