Artificial Neural Network Model to Investigate Supercontinuum Generation in Chalcogenide Planar Waveguide

This study introduces a Deep Learning (DL) approach to predict supercontinuum generation in optical waveguides, addressing computational and time challenges of traditional numerical methods. The novel Artificial Neural Network (ANN) model, trained from scratch, demonstrates impressive accuracy in forecasting supercontinuum spectral bandwidth with a mean squared error of 0.017. Notably, the ANN model dramatically accelerates computations, taking only half the time of simulations in COMSOL Multiphysics and MATLAB. This DL-powered model presents a promising alternative for predicting supercontinuum generation, effectively overcoming resource constraints and complexity associated with conventional numerical methods.