RNN architectures trained with Backpropagation and Reservoir Computing (RC) methods for forecasting high-dimensional chaotic dynamical systems.
