Nonlinear photonic dynamical systems for unconventional computing

Driven by the remarkable breakthroughs during past decade, photonics neural networks have experienced a revival. Here, we provide general overview of progress over and sketch roadmap important future developments. We focus on photonic implementations reservoir computing machine learning paradigm, which offers conceptually simple approach that is amenable to hardware implementations. In particular, an implemented via either spatio temporal or delay dynamical systems. Going beyond computing, discuss recent advances challenges deep networks, quest for methods are hardware-friendly as well realizing autonomous i.e. with minimal digital electronic auxiliary hardware.