Reviews: Robustness Verification of Tree-based Models

Originality: The robustness verification methods presented in the paper is new and interesting. The authors provided a fair list of related work and compared the existing methods with their method in the experiment section. Quality: The paper provides a complete presentation of three verification methods, 1) verifying the robustness of a single decision tree, 2) verifying the robustness of a tree ensemble using existing algorithms for finding k-cliques, and 3) a fast and approximate method for estimating a lower bound on the robustness. The theoretical claims and their proofs make sense to me. Overall the empirical evaluation is well designed and convincing.