Extended anti-Hebbian adaptation for unsupervised source extraction

Author

Malouche, Zied ; Macchi, Odile

Author_Institution

Lab. des Signaux et Syst., CNRS, Gif-sur-Yvette, France

Volume

3

fYear

1996

fDate

7-10 May 1996

Firstpage

1664

Abstract

We propose a new adaptive algorithm to separate a linear mixture of sources using an extended anti-Hebbian rule. This solution can be viewed as a stochastic gradient way to minimize certain output high order statistics. The system is modular: it is decomposed into parallel and independent subsystems. Each one is capable of extracting one source with negative kurtosis out of the mixture, provided the number of observations is greater or equal to the number of sources and provided it is appropriately initialized

Keywords

adaptive signal processing; array signal processing; higher order statistics; neural nets; stochastic processes; unsupervised learning; adaptive algorithm; array signal processing; extended anti-Hebbian rule; independent subsystems; linear source mixture; modular system; negative kurtosis; neural networks; observations; output high order statistics; parallel subsystems; source separation; source statistics; stochastic algorithm; stochastic gradient method; unsupervised source extraction; Adaptive algorithm; Array signal processing; Computer networks; Costs; Neural networks; Polynomials; Principal component analysis; Source separation; Statistics; Stochastic processes;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1996. ICASSP-96. Conference Proceedings., 1996 IEEE International Conference on

Conference_Location

Atlanta, GA

ISSN

1520-6149

Print_ISBN

0-7803-3192-3

Type

conf

DOI

10.1109/ICASSP.1996.544125

Filename

544125