A neural network architecture is proposed to determine the number of solitons generated by random processes in optical wavelength-division multiplexed telecommunication systems with QPSK, 16-QAM, 64-QAM, and 1024-QAM modulation. The dependence of the prediction quality of a neural network with a special architecture on the number of soliton modes in the signal and the parameters of this signal is studied.
Bibliographical noteFunding Information:
The work was supported by the Fund of the President of the Russian Federation for State Support of Young Russian Scientists (Grant No. MK-677.2020.9). The work of I.S. Chekhovskoy was supported by the state assignment for fundamental research (FSUS-2020-0034), and the work of J.E. Prilepsky was supported by the Leverhulme Trust (Project RPG-2018-063).
- inverse scattering problem method
- machine learning
- neural networks
- nonlinear Fourier transform
- nonlinear Schrödinger equation
- optical telecommunication systems
- wavelength-division multiplexing
- Zakharov - Shabat problem