Summary
International Symposium on Nonlinear Theory and Its Applications
2022
Session Number:B1L-C
Session:
Number:B1L-C-02
Characteristics Representation of Reservoir Set Based on Memory Capacity and Nonlinearity
Tomoya Kitamura , Kazuyuki Yoshimura,
pp.221-224
Publication Date:12/12/2022
Online ISSN:2188-5079
DOI:10.34385/proc.71.B1L-C-02
PDF download (2.6MB)
Summary:
Reservoir Computing has various degrees of freedom in terms of reservoir parameters, topology and activation functions; however, the design strategy of a good reservoir is still unclear. In this study, we focus on memory capacity and nonlinearity, where the memory and nonlinearity indices represent the characteristics of the reservoir dynamics. And we show that the two indices correlate with the information processing performance of the reservoir for a particular benchmark task.