Information processing by using mutually-coupled optoelectronic systems

We report information processing by using mutually-coupled optoelectronic system, which is called reservoir computing. Reservoir computing is a novel bio-inspired computing method and highly efficient approach for processing empirical data. We numerically show that the performance of reservoir computing is improved by using two Mach-Zehnder modulators. Chaotic time-series prediction task is used for evaluation of the performance. Our result shows that the reservoir using the two modulators has better performance than using a single one.