We present a novel application of an echo state neural network (ESN) to noninvasive foetal electrocardiogram (FECG) extraction. Extraction of the FECG is performed on abdominal recordings of pregnant women via maternal ECG cancellation. The FECG can then be used for foetal health monitoring by extracting clinically interpretable features. We show that optimising an ESN by random search gives al...

Autonomous, self* sensor networks require sensor nodes with a certain degree of “intelligence”. An elementary component of such an “intelligence” is the ability to learn online predicting sensor values. We consider recurrent neural network (RNN) models trained with an extended Kalman filter algorithm based on real time recurrent learning (RTRL) with teacher forcing. We compared the performance ...

Echo state networks are powerful recurrent neural networks. However, they are often unstable and shaky, making the process of finding an good ESN for a specific dataset quite hard. Obtaining a superb accuracy by using the Echo State Network is a challenging task. We create, develop and implement a family of predictably optimal robust and stable ensemble of Echo State Networks via regularizing t...

This paper use echo state network (ESN), feedforward echo state network (FE-ESN) and tapped delay line with inputs (TDL-I) to predict the sea clutter time series and detect target embedded in sea clutter. The performance of predicting and detecting using these methods is compared. A set of time series from IPIX radar data is tested. Numerical experiments reveal that FE-ESN and TDL-I show high p...

In this report, we developed a new recurrent neural network toolbox, including the recurrent multilayer perceptron structure and its companying extended Kalman filter based training algorithms: BPTT-GEKF and BPTT-DEKF. Besides, we also constructed programs for designing echo state network with single reservoir, together with the offline linear regression based training algorithm. We name this t...

Recently, echo state network (ESN) has attracted a great deal of attention due to its high accuracy and efficient learning performance. Compared with the traditional random structure and classical sigmoid units, simple circle topology and leaky integrator neurons have more advantages on reservoir computing of ESN. In this paper, we propose a new model of ESN with both circle reservoir structure...

Abstract As a recurrent neural network, ESN has attracted wide attention because of its simple training process and unique reservoir structure, been applied to time series prediction other fields. However, also some shortcomings, such as the optimization collinearity. Many researchers try optimize structure performance deep by constructing ESN. with increase number network layers, problem low c...

In this paper we introduce an efficient approach to Recursive Neural Networks (RecNNs) modeling, the Tree Echo State Network (TreeESN), extending the Echo State Network (ESN) model from sequential to tree structured domains processing. For structure-to-element transductions, the state mapping (i.e. the way in which the state values for the whole structure are selected/collected) turns out to ha...

This tutorial is a worked-out version of a 5-hour course originally held at AIS in September/October 2002. It has two distinct components. First, it contains a mathematically-oriented crash course on traditional training methods for recurrent neural networks, covering back-propagation through time (BPTT), real-time recurrent learning (RTRL), and extended Kalman filtering approaches (EKF). This ...