Echo state networks (ESN) are a novel approach to recurrent neural network training. An ESN consists of a large, fixed, recurrent "reservoir" network, from which the desired output is obtained by training suitable output connection weights. Determination of optimal output weights becomes a linear, uniquely solvable task of MSE minimization. This article reviews the basic ideas and describes an ...

This paper presents the design of feedback linearization and neural network based feedback linearization (NARMAL2) controller for a magnetic levitation system. The magnetic levitation system is one of the classical nonlinear systems. The paper provides simulation results to validate the theoretical design.

Greedy algorithms form an essential tool for compressed sensing. However, their inherent batch mode discourages their use in time-varying environments due to significant complexity and storage requirements. In this paper a powerful greedy scheme developed in [1, 2] is converted into an adaptive algorithm which is applied to estimation of nonlinear channels. Performance is assessed via computer ...

Current synthetic chemical systems lack the ability to selfmodify and learn to solve desired tasks. In this paper we introduce a new parallel model of a chemical delay line, which stores past concentrations over time with minimal latency. To enable temporal processing, we integrate the delay line with our previously proposed analog chemical perceptron. We show that we can successfully train our...