Parallel Extreme Learning Machines Based on Frequency Multiplexing

In a recent work, we reported on an Extreme Learning Machine (ELM) implemented in photonic system based frequency multiplexing, where each wavelength of the light encodes different neuron state. present experimentally demonstrate parallelization potentialities this approach. We show that multiple combs centered frequencies can copropagate same system, resulting either independent ELMs executed parallel substrate or single ELM with increased number neurons. tested performances both these operation modes several classification tasks, employing up to three sources, which generates comb. also numerically evaluated configurations containing 15 sources.