Capsule networks (CapsNets) are new neural networks that classify images
based on the spatial relationships of features. By analyzing the pose of
features and their relative positions, it is more capable to recognize images
after affine transformation. The stacked capsule autoencoder (SCAE) is a
state-of-the-art CapsNet, and achieved unsupervised classification of CapsNets
for the first time. However, the security vulnerabilities and the robustness of
the SCAE has rarely been explored. In this paper, we propose an evasion attack
against SCAE, where the attacker can generate adversarial perturbations based
on reducing the contribution of the object capsules in SCAE related to the
original category of the image. The adversarial perturbations are then applied
to the original images, and the perturbed images will be misclassified.
Furthermore, we propose a defense method called Hybrid Adversarial Training
(HAT) against such evasion attacks. HAT makes use of adversarial training and
adversarial distillation to achieve better robustness and stability. We
evaluate the defense method and the experimental results show that the refined
SCAE model can achieve 82.14% classification accuracy under evasion attack. The
source code is available at this https URL.

该研究提出了一种针对 SCAE 的逃避攻击和名为 Hybrid Adversarial Training 的防御方法，该方法结合了对抗训练和对抗蒸馏以实现更好的稳健性和稳定性。实验结果表明，该防御方法的确能够提高对逃避攻击的识别能力。