In this section, we ablate the kernel size used to train DISCO on ImageNet. TableIshowsthats=3 achieves the best performance, which degrades fors = 5 by a significant margin (3.26%). This is consistent with the well known complexity of synthesizing images withglobalmodels, suchasGANs. For a single ImageNet image of size 224, STL requires 23.71 seconds while DISCO (K=1) only requires0.027. In this section, we list the url links that are used for training and evaluating DISCO.

The problem of adversarial defenses for image classification, where the goal is to robustify a classifier against adversarial examples, is considered.

As a crucial step for analyzing human actions, hand-object interaction understanding is researchworthy in a broad range of applications related to virtual or augmented reality. Current works largely focus on hand-object pose estimation (HOPE) [16, 19, 21], which aims to capture the pose configuration of the given hand-object image.

As with all data-driven approaches, BO is prone to cases where the given data shifts from the data of interest.

As with all data-driven approaches, BO is prone to cases where the given data shifts from the data of interest.