Efficient Optimisation of Physical Reservoir Computers using only a Delayed Input

Reservoir Computing (RC) [1] is a possible promising approach. RC diverges from conventional neural networks by introducing a distinct architecture characterized by a fixed, randomly initialized recurrent layer, called "reservoir", coupled with a simple linear readout layer: its characteristic structure reduces the training computational complexity, and consequently, the energy consumption. RC possesses a remarkable capacity to process temporal data, distinguishing itself across a large variety of tasks, such as equalization of distorted nonlinear communication channels [1, 2], audio processing [3, 4], weather prediction [5], image [6, 7] and video [8, 9] classification. RC has also gathered attention for its feasibility to being implemented on a wide range of physical substrates [10, 11]. In particular, photonic RC stands out as one of the most promising hardware platforms, thanks to its advantages in parallelization [12, 13], high speed [14] and minimal hardware requirements [15].