One-Pixel Attack for Continuous-Variable Quantum Key Distribution Systems

Deep neural networks (DNNs) have been employed in continuous-variable quantum key distribution (CV-QKD) systems as attacking detection portions of defense countermeasures. However, the vulnerability DNNs leaves security loopholes for hacking attacks, example, adversarial attacks. In this paper, we propose to implement one-pixel attack CV-QKD and accomplish misclassification on a minimum perturbation. This approach is based differential evolution, which makes our algorithm fool multiple with minimal inner information target networks. The simulation experimental results show that, four different networks, 52.8%, 26.4%, 21.2%, 23.8% input data can be perturbed another class by modifying just one feature, same pixel an image. We carry out success rate context original accuracy reaching up nearly 99% average. Further, enlarging number features, raised satisfactory higher level about 80%. According results, most deceived launching