Echo State Networks trained by Tikhonov least squares are L2(μ) approximators of ergodic dynamical systems

PHYSICA D-NONLINEAR PHENOMENA

卷:421

Article

Hart, Allen G.¹  Hook, James L.¹  Dawes, Jonathan H. P.¹ 

[1] Univ Bath, Bath, Avon, England

关键词: Reservoir computing;    Liquid state machine;    Time series analysis;    Lorenz equations;    Delay embedding;    Recurrent neural networks;   

DOI  :  10.1016/j.physd.2021.132882

来源: Elsevier

【 摘 要 】

Echo State Networks (ESNs) are a class of single-layer recurrent neural networks with randomly generated internal weights, and a single layer of tuneable outer weights, which are usually trained by regularised linear least squares regression. Remarkably, ESNs still enjoy the universal approximation property despite the training procedure being entirely linear. In this paper, we prove that an ESN trained on a sequence of observations from an ergodic dynamical system (with invariant measure mu) using Tikhonov least squares regression against a set of targets, will approximate the target function in the L-2(mu) norm. In the special case that the targets are future observations, the ESN is learning the next step map, which allows time series forecasting. We demonstrate the theory numerically by training an ESN using Tikhonov least squares on a sequence of scalar observations of the Lorenz system. (C) 2021 Elsevier B.V. All rights reserved.

【 授权许可】

Free