Evolving reservoir weights in the frequency domain - Archive ouverte HAL Access content directly
Conference Papers Year : 2021

Evolving reservoir weights in the frequency domain

(1) , (2)
1
2

Abstract

Reservoir Computing models are a class of recurrent neural networks that have enjoyed recent attention, in particular, their main family, Echo State Networks (ESNs). These models have a large number of hidden-hidden weights (in the so-called reservoir) forming a recurrent topology. The reservoir is randomly connected with fixed weights during learning: only readout parameters (from reservoir to output neurons) are trained; the reservoir weights are frozen after initialized. Since the reservoir structure is fixed during learning, only its initialization process has an impact on the model's performance. In this work, we introduce an evolutionary method for adjusting the reservoir non-null weights. Moreover, the evolutionary process runs on the frequency space corresponding to a Fourier transformation of the weights. We combine an evolutionary search in the Fourier space with supervised learning for the readout weights. The resulting algorithm, called EvoESN (Evolutionary ESN), obtains promising results modeling two well-known problems of the chaotic time-series area.
Fichier principal
Vignette du fichier
Poster_V2.pdf (446.63 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03508409 , version 1 (03-01-2022)

Identifiers

Cite

Sebastián Basterrech, Gerardo Rubino. Evolving reservoir weights in the frequency domain. GECCO 2021 - Genetic and Evolutionary Computation Conference, Jul 2021, Lille, France. pp.271-272, ⟨10.1145/3449726.3459457⟩. ⟨hal-03508409⟩
14 View
14 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More