Predicting nonlinear reshaping of periodic signals in optical fibre with a neural network

We deploy a supervised machine-learning model based on neural network to predict the temporal and spectral reshaping of simple sinusoidal modulation into pulse train having comb structure in frequency domain, which occurs upon nonlinear propagation an optical fibre. Both normal anomalous second-order dispersion regimes fibre are studied, speed is leveraged probe space input parameters for generation custom combs or occurrence significant focusing.