Optical Neural Network Based on Synthetic Nonlinear Photonic Lattices

Abstract

We reveal that a synthetic photonic lattice based on coupled optical loops can be utilized as a feed-forward neural network for processing of optical pulse sequences in time domain. As a proof of concept, we train the optical system to restore an initial shape of the pulse train from the signal distorted due to linear dispersion in a fiber-optic link. We also show efficient training of the optical network with an intrinsic Kerr-type nonlinearity for the realization of target nonlinear transmission functions and inference functionality for the discrimination of different pulse sequences. The theoretical modeling is performed under practical conditions and can guide future experimental realizations.