Adversarial Perturbation Elimination with GAN Based Defense in Continuous-Variable Quantum Key Distribution Systems

Abstract

Machine learning is being applied to continuous-variable quantum key distribution (CVQKD) systems as defense countermeasures for attack classification. However, recent studies have demonstrated that most of these detection networks are not immune to adversarial attacks. In this paper, we propose to implement typical adversarial attack strategies against the CVQKD system and introduce a generalized defense scheme. Adversarial attacks essentially generate data points located near decision boundaries that are linearized based on iterations of the classifier to lead to misclassification. Using the DeepFool attack as an example, we test it on four different CVQKD detection networks and demonstrate that an adversarial attack can fool most CVQKD detection networks. To solve this problem, we propose an improved adversarial perturbation elimination with a generative adversarial network (APE-GAN) scheme to generate samples with similar distribution to the original samples to defend against adversarial attacks. The results show that the proposed scheme can effectively defend against adversarial attacks including DeepFool and other adversarial attacks and significantly improve the security of communication systems.

Generative AI in Medical Practice: In-Depth Exploration of Privacy and Security Challenges

“…(3) training phase-data poisoning is the main threat; defense strategies focus on techniques that can identify and remove poisoned data (eg, the certified defense technique proposed by Tang et al [58]) and provide robust and reliable AI models; (4) inference phase-this phase mainly faces adversarial example attacks such as white-box, gray-box, and black-box attacks depending on how much the attacker knows about the target model; a variety of defense strategies can be implemented to tackle such attacks, such as adopting strategies in phases 1 to 3 to modify data (eg, data reconstruction and randomization) or modify or enhance models with newer model construction methods resistant to adversarial example attacks (eg, using deep neural networks and GAN-based networks [58,59]); (5) integration phase-AI models face AI biases, confidentiality attacks (eg, model inversion, model extraction, and various privacy attacks), and code vulnerability exploitation; defense strategies in this phase should be comprehensive via integrating various solutions such as fuzz testing and blockchain-based privacy protection.…”

An Adaptive Enhancement Method of Malicious Traffic Samples Based on DCGAN-ResNet System

“…Furthermore, the DCGAN, based on generative adversarial network (GAN) (Ring et al, 2019;Kawai et al, 2019;Hu & Tan, 2017;Frid-Adar et al, 2018), introduces convolutional layers to enhance the quality of generated traffic and the training stability of the network (Jia et al, 2023). Through an automated parameter determination method, it further enhances the ability of various system models to detect unknown threats, strengthening the robustness of the model (Tang et al, 2023).…”