Next Generation Reservoir Computing

Abstract

Reservoir computing is a best-in-class machine learning algorithm for processing information generated by dynamical systems using observed time-series data. Importantly, it requires very small training data sets, uses linear optimization, and thus requires minimal computing resources. However, the algorithm uses randomly sampled matrices to define the underlying recurrent neural network and has a multitude of metaparameters that must be optimized. Recent results demonstrate the equivalence of reservoir computing to nonlinear vector autoregression, which requires no random matrices, fewer metaparameters, and provides interpretable results. Here, we demonstrate that nonlinear vector autoregression excels at reservoir computing benchmark tasks and requires even shorter training data sets and training time, heralding the next generation of reservoir computing.

New supplementary materials on noise sensitivity, dependence on training data set size and other smaller changes

References in corpus (2)

• Using Machine Learning to Replicate Chaotic Attractors and Calculate Lyapunov Exponents from Data
• Forecasting Chaotic Systems with Very Low Connectivity Reservoir Computers

Cited by in corpus (4)

• Combining machine learning and data assimilation to forecast dynamical systems from noisy partial observations
• Optimizing Memory in Reservoir Computers
• Neural Echo State Network using oscillations of gas bubbles in water
• Model-free inference of unseen attractors: Reconstructing phase space features from a single noisy trajectory using reservoir computing