Temporal Convolution Derived Multi-Layered Reservoir Computing

TL;DR

This paper introduces novel reservoir computing architectures with a new data mapping method, significantly improving time series prediction accuracy, especially for chaotic data, while reducing reliance on randomness and hyper-parameter tuning.

Contribution

The authors propose a new input mapping and two innovative network architectures that enhance parallelizability, depth, and predictive performance in reservoir computing for time series forecasting.

Findings

Error reduction of up to 85.45% for chaotic series compared to Echo State Networks

Error reduction of up to 99.99% for non-chaotic series compared to existing methods

Significant improvements in prediction accuracy across diverse time series datasets.

Abstract

The prediction of time series is a challenging task relevant in such diverse applications as analyzing financial data, forecasting flow dynamics or understanding biological processes. Especially chaotic time series that depend on a long history pose an exceptionally difficult problem. While machine learning has shown to be a promising approach for predicting such time series, it either demands long training time and much training data when using deep Recurrent Neural Networks. Alternative, when using a Reservoir Computing approach it comes with high uncertainty and typically a high number of random initializations and extensive hyper-parameter tuning. In this paper, we focus on the Reservoir Computing approach and propose a new mapping of input data into the reservoir's state space. Furthermore, we incorporate this method in two novel network architectures increasing parallelizability, depth and predictive capabilities of the neural network while reducing the dependence on randomness. For the evaluation, we approximate a set of time series from the Mackey-Glass equation, inhabiting non-chaotic as well as chaotic behavior as well as the SantaFe Laser dataset and compare our approaches in regard to their predictive capabilities to Echo State Networks, Autoencoder connected Echo State Networks and Gated Recurrent Units. For the chaotic time series, we observe an error reduction of up to $85.45\%$ compared to Echo State Networks and $90.72\%$ compared to Gated Recurrent Units. Furthermore, we also observe tremendous improvements for non-chaotic time series of up to $99.99\%$ in contrast to the existing approaches.