Predicting Dielectric Waveguides Characteristics Using Deep Learning

We propose an unsupervised deep learning model based on physics-informed neural network (PINNS) to find the effective refractive index of a slab waveguide. The model accuracy could reach 99% within a time range from 60 to 120 seconds for symmetric and anti-symmetric waveguide. The results show the success of the introduced method in solving fail cases of the compared methods.