Parallel and deep reservoir computing using semiconductor lasers with optical feedback

Abstract Photonic reservoir computing has been intensively investigated to solve machine learning tasks effectively. A simple procedure of output weights is used for computing. However, the lack training input-node and inter-node connection limits performance The use multiple reservoirs can be a solution overcome this limitation In study, we investigate parallel deep configurations delay-based all-optical using semiconductor lasers with optical feedback by combining improve Furthermore, propose hybrid configuration maximize benefits reservoirs. We perform chaotic time-series prediction task, nonlinear channel equalization memory capacity measurement. Then, compare single, parallel, deep, configurations. find that are suitable whereas task. Hybrid outperform other all three tasks. further optimize number each configuration. Multiple show great potential improvement computing, which in turn applied high-performance edge