Parameter Optimization Based BPNN of Atmosphere Continuous-Variable Quantum Key Distribution

Abstract

The goal of continuous variable quantum key distribution (CVQKD) is to be diffusely used and adopted in diverse scenarios, so the adhibition of atmospheric channel will play a crucial part in constituting global secure quantum communications. Atmospheric channel transmittance is affected by many factors and does not vary linearly, leading to great changes in signal-to-noise ratio. It is crucial to choose the appropriate modulation variance under different turbulence intensities to acquire the optimal secret key rate. In this paper, the four-state protocol, back-propagation neural network (BPNN) algorithm was discussed in the proposed scheme. We employ BPNN to CVQKD, which could adjust the modulation variance to an optimum value for ensuring the system security and making the system performance optimal. The numerical results show that the proposed scheme is equipped to improve the secret key rate efficiently.

A Survey of Machine Learning Assisted Continuous-Variable Quantum Key Distribution

“…[28] was applied to key sifting, ref. [25,26,55] were applied to parameter estimation and optimization, ref. [29] was applied to information reconciliation, and [30,31] were applied to key rate estimation.…”

“…A linear activation function is commonly used in the output layer of NNs designed for regression, such that the output can take any continuous value, while sigmoid activation functions are commonly used for classification NNs to return discrete outputs. In the surveyed works, MLPs were applied to excess noise filtering in [12], parameter optimization in [26], reconciliation in [29], and key rate estimation in [30]. Figure 5a gives an example of a generic fully connected MLP, where the connections between the input layer, hidden layer, and output layer are shown, as well as an example perceptron in Figure 5b, indicating how the inputs and weights from the previous layer are transferred into an output via the activation function (following the path outlined in red).…”

“…While many traditional approaches have been suggested to overcome the limitations on CV-QKD, machine learning (ML) has recently been shown to have advantages in terms of phase error estimation and excess noise filtering [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20], state discrimination [21][22][23][24], parameter estimation and optimization [25][26][27], key sifting [28], reconciliation [29], and key rate estimation [30,31]. ML-based phase error estimation and noise filtering algorithms offer the potential of improved filtering capabilities due to their ability to map complex relationship between inputs and outputs based on the data alone, without being based on idealistic models that may not represent reality.…”

Short-wave infrared continuous-variable quantum key distribution over satellite-to-submarine channels

“…[11][12][13] In previous experimental implementations, the 1500-to-1600-nm light is usually regarded as an ideal quantum pulse carrier in the atmospheric channel due to compatibility with current telecom technology, the wavelength-dependent nature of scattering, and lower solar radiance. [14][15][16][17] However, the underwater transmission of 1500-to-1600-nm light tends to suffer from severe attenuation, which inevitably deteriorates the security of the satellite-to-submarine system. Therefore, we need to study the parameters of the atmosphere, sea surface, and sea water channel, and establish a complete satellite-submarine quantum channel model.…”

Secure Continuous-Variable Quantum Key Distribution with Machine Learning

“…Besides, in [29], a distance-weighted k-nearest-neighbors-based machine-learning detector was proposed to directly deal with the raw secret key, but not the system parameters, which is a new idea for improving the performance of CV systems. Reference [30] employed a backpropagation neural network to adjust the modulation variance to an optimal value and to furnish a higher achievable key rate and a more efficient parameter optimization than the local search algorithm in the practical four-state CVQKD system.…”