Optical communication, optical information science, and optical computations

Application of machine learning algorithms for demodulation in QKD system with quadrature modualtion

B. E. Pervushin^a, A. V. Zinovev^b, D. N. Kirichenko^b, I. M. Filipov^b, R. K. Goncharov^b, E. O. Samsonov<sup>b

a</sup> Yandex company, Moscow
^b St. Petersburg National Research University of Information Technologies, Mechanics and Optics

Abstract: The paper shows the application of machine learning algorithms for state demodulation in a continuous-variable quantum key distribution (CV-QKD) system with quadrature-amplitude modulation (QAM) with 16 states. When using classification algorithms, the average BER was 0.019, and 0.022 for using clustering, which is 1.75 and 1.5 times less than when using the classical demodulation method using LLR.

Keywords: quantum key distribution, machine learning.

Received: 28.04.2025
Revised: 27.06.2025
Accepted: 28.06.2025

DOI: 10.61011/OS.2025.07.61109.7725-25

Bibliographic databases:

•