A Linear Feedforward Neural Network with Lateral Feedback Connections for Blind Source Separation
نویسندگان
چکیده
We presents a new necessary and sufficient condition for the blind separation of sources having non-zero kurtosis, from their linear mixtures. It is shown here that a new blind separation criterion based on both odd ( ) and even ( ) functions, presents desirable solutions, provided that all source signals have negative kurtosis (sub-Gaussian) or have positive kurtosis (super-Gaussian). Based on this new separation criterion, a linear feedforward network with lateral feedback connections is constructed. Both theoretical and computer simulation results are presented.
منابع مشابه
Mixed Signal Learning by Spike Correlation Propagation in Feedback Inhibitory Circuits
The brain can learn and detect mixed input signals masked by various types of noise, and spike-timing-dependent plasticity (STDP) is the candidate synaptic level mechanism. Because sensory inputs typically have spike correlation, and local circuits have dense feedback connections, input spikes cause the propagation of spike correlation in lateral circuits; however, it is largely unknown how thi...
متن کاملBlind separation of circularly distributed sources by neural extended APEX algorithm
The aim of this work is to present a generalized Hebbian learning theory for complex-weighted linear feed-forward network endowed with lateral inhibitory connections, and to show how it can be applied to blind separation from complex-valued mixtures. We start by stating an optimization principle for Kung-Diamantaras' network which leads to a generalized APEX-like learning theory relying on some...
متن کاملA class of neural networks for independent component analysis
Independent component analysis (ICA) is a recently developed, useful extension of standard principal component analysis (PCA). The ICA model is utilized mainly in blind separation of unknown source signals from their linear mixtures. In this application only the source signals which correspond to the coefficients of the ICA expansion are of interest. In this paper, we propose neural structures ...
متن کاملRedundancy Reduction and Independent Component Analysis: Conditions on Cumulants and Adaptive Approaches Submitted to Neural Computation
In the context of both sensory coding and signal processing, building fac-torized codes has been shown to be an eecient strategy. In a wide variety of situations, the signal to be processed is a linear mixture of statistically independent sources. Building a factorized code is then equivalent to performing blind source separation. Thanks to the linear structure of the data, this can be done, in...
متن کاملAlgorithms for Blind Signal Separation and Recovery in Static and Dynamic Environments
We propose update laws for the problem of blind separation in static and dynamic environments. The energy function is based on an approximation of the mutual information as a measure of independence. Both feedforward and feedback structures of the neural network are considered. A general framework to develop the update law for the dynamic model is proposed. Computer simulations to support the a...
متن کامل