Summary
International Symposium on Nonlinear Theory and Its Applications
2022
Session Number:D1L-C
Session:
Number:D1L-C-02
Application of a Multi-Layer Reservoir Neural Network to the Prediction of Spatiotemporal Chaos
Shoki Yabu , Yo Horikawa,
pp.609-612
Publication Date:12/12/2022
Online ISSN:2188-5079
DOI:10.34385/proc.71.D1L-C-02
PDF download (884KB)
Summary:
reservoir computing (rc) is one of recurrent neural networks that can process temporal information. in the rc, the weights of input and reservoir layers are fixed and only the weights of an output layer are learned. in this study, we apply a multi-layer reservoir neural network to the prediction of spatiotemporal chaos. the model has hierarchical structure with parallel reservoirs stacked in a pyramid-like shape. we then compare its performance with a single-layer reservoir network.