Multi-Attack Detection: General Defense Strategy Based on Neural Networks for CV-QKD

Abstract

The security of the continuous-variable quantum key distribution (CVQKD) system is subject to various attacks by hackers. The traditional detection method of parameter estimation requires professionals to judge known attacks individually, so the general detection model emerges to improve the universality of detection; however, current universal detection methods only consider the independent existence of attacks but ignore the possible coexistence of multiple attacks in reality. Here, we propose two multi-attack neural network detection models to handle the coexistence of multiple attacks. The models adopt two methods in multi-label learning: binary relevance (BR) and label power (LP) to deal with the coexistence of multiple attacks and can identify attacks in real-time by autonomously learning the features of known attacks in a deep neural network. Further, we improve the model to detect unknown attacks simultaneously. The experimental results show that the proposed scheme can achieve high-precision detection for most known and unknown attacks without reducing the key rate and maximum transmission distance.

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

“…At last, in order to accommodate the universal ANN models in the image field and strengthen the stability of the input data as well, we combine 25 continuous feature vectors as an input matrix, which can be seen as a 25 × 4 image with one channel. The choice of this number refers the experiments of Luo et al [26] and Du et al [27]. The group generated here is the basic unit for our network to classify.…”

“…In recent years, with the swift development of artificial intelligence (AI) [23], many innovations based on the artificial neural network (ANN) has been proven to be effective. For example [24], Mao et al [25] proposed an ANN model to classify their attack strategy, Luo et al [26] proposed a semi-supervised deep learning method to detect known attacks and potential unknown attacks, and Du et al [27] proposed an ANN model for multi-attacks detection. The main idea of these methods is to implement specific defense countermeasures based on the classification result from the ANN model.…”

General multi-attack detection for continuous-variable quantum key distribution with local local oscillator

“…In this paper, to improve the practical security of LLO CVQKD, we propose a general defense strategy via deep learning that can be monitored in real-time without adding additional optics considering the development of machine learning opens a door for QKD. [27][28][29][30][31][32][33][34][35] We investigate the typical characteristics of pulses affected by attacks and the deviations in the changes in these characteristics after being subjected to different attacks. After analysis, several features are selected to construct feature vectors for training generative adversarial networks (GANs).…”

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

“…Fortunately, the boom in machine learning has brought benefits to CVQKD defense, and numerous developments based on artificial neural networks (ANN) have been demonstrated to be successful. An ANN model for attack detection and classification was proposed by Mao et al [14], which functions by extracting the feature vectors of most known attacks as input; Du et al [15] proposed a multi-attack ANN detection model to handle the coexistence of multiple attacks; and a semi-supervised deep learning strategy was put forth by Luo et al [16] to identify known assaults or potential threats. The key to the above approaches is to implement specific defensive measures according to the classification results of the ANN model.…”