Scalable Delay Line-Free All-Optical Reservoir
Ivan Boikov, Daniel Brunner, Alfredo De Rossi · 2023
All-optical neural networks (NN) are an exciting opportunity to increase computing speed and efficiency [1]. For this purpose Reservoir Computing (RC) [2] (see Fig. 1a) has attracted a lot of interest as it provides performance competitive with conventional NNs, but with a simpler implementation. Particular interest is on RC based on integrated photonics, as it is compact, scalable and provides intrinsic nonlinear effects [3] necessary for a NN.