Quaternion-valued echo state networks.

IEEE Trans Neural Netw Learn Syst

Published: April 2015

Quaternion-valued echo state networks (QESNs) are introduced to cater for 3-D and 4-D processes, such as those observed in the context of renewable energy (3-D wind modeling) and human centered computing (3-D inertial body sensors). The introduction of QESNs is made possible by the recent emergence of quaternion nonlinear activation functions with local analytic properties, required by nonlinear gradient descent training algorithms. To make QENSs second-order optimal for the generality of quaternion signals (both circular and noncircular), we employ augmented quaternion statistics to introduce widely linear QESNs. To that end, the standard widely linear model is modified so as to suit the properties of dynamical reservoir, typically realized by recurrent neural networks. This allows for a full exploitation of second-order information in the data, contained both in the covariance and pseudocovariances, and a rigorous account of second-order noncircularity (improperness), and the corresponding power mismatch and coupling between the data components. Simulations in the prediction setting on both benchmark circular and noncircular signals and on noncircular real-world 3-D body motion data support the analysis.

Download full-text PDF

Source
http://dx.doi.org/10.1109/TNNLS.2014.2320715DOI Listing

Publication Analysis

Top Keywords

quaternion-valued echo
8
echo state
8
state networks
8
circular noncircular
8
networks quaternion-valued
4
networks qesns
4
qesns introduced
4
introduced cater
4
3-d
4
cater 3-d
4

Similar Publications

Delay-dependent stability analysis of the QUAD vector field fractional order quaternion-valued memristive uncertain neutral type leaky integrator echo state neural networks.

Neural Netw

September 2019

Institute of Engineering, Polytechnic of Porto, Department of Electrical Engineering, Rua Dr. António Bernardino de Almeida 431, 4249-015 Porto, Portugal. Electronic address:

This paper studies the robust stability analysis for a class of memristive-based neural networks (NN). The NN consists of a fractional order neutral type quaternion-valued leaky integrator echo state with parameter uncertainties and time-varying delays. First, the quaternion-valued leaky integrator echo state NN with QUAD vector field activation function is transformed into a real-valued system using a linear mapping function.

View Article and Find Full Text PDF

Quaternion-valued echo state networks (QESNs) are introduced to cater for 3-D and 4-D processes, such as those observed in the context of renewable energy (3-D wind modeling) and human centered computing (3-D inertial body sensors). The introduction of QESNs is made possible by the recent emergence of quaternion nonlinear activation functions with local analytic properties, required by nonlinear gradient descent training algorithms. To make QENSs second-order optimal for the generality of quaternion signals (both circular and noncircular), we employ augmented quaternion statistics to introduce widely linear QESNs.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!