Time and wavelength multiplexing in photonic neural networks

In this master’s thesis, we explore the Reservoir Computing paradigm, specifically focusing
on Delay-based Reservoir Computing. This concept posits that a vast physical network of
nodes can be substituted with a single nonlinear node with delayed feedback. In this setup,
actual nodes in the physical network become virtual nodes in the delay system via temporal
multiplexing. We test the efficiency of a photonic approach in a benchmark task: predicting
chaotic time series. We model the system using a semiconductor laser-based photonic reservoir
computer and time-delayed optical feedback rooted in the Lang-Kobayashi equations.
Additionally, we enhance this concept by optimizing resources, utilizing various frequency
filters at the output of the laser, and demonstrating the improvements these filters bring
to the prediction task. Our numerical results show that the prediction performance is
improved when frequency and time multiplexing are combined, suggesting a promising
strategy for future experimental implementations.