Hardware implementations of trainable nonlinear oscillators via reservoir computing

Hardware-implemented reservoir computing (RC) has been gaining considerable interest in recent years, due to the gain in fundamental insights and its appeal for applications including classification and nonlinear-prediction tasks. In this talk, I will start giving an overview of a few key publications that paved the way for the realization of trainable autonomous nonlinear oscillators. Then, I will present how noise can play a constructive role in the replication of chaotic dynamics when using these autonomous nonlinear oscillators. Finally, I will discuss current limitations and novel applications and perspectives for this reservoir computing-based approach.