Nonlinear multilayer principal component type subspace learning algorithms

Author

Joutsensalo, Jyrki ; Karhunen, Juha

Author_Institution

Lab. of Comput. & Inf. Sci., Helsinki Univ. of Technol., Espoo, Finland

fYear

1993

fDate

6-9 Sep 1993

Firstpage

68

Lastpage

77

Abstract

A hidden layer is introduced into nonlinear principal component type learning algorithms. The algorithms are derived from nonlinear optimization criteria. Both subspace type and hierarchical versions are considered. The algorithms are tested in context with harmonic retrieval and directions-of-arrival estimation problems using impulsive and colored noise. Some of the nonlinear algorithms have interesting signal separation properties, i.e., the neurons become sensitive to independent sinusoidal signals

Keywords

direction-of-arrival estimation; learning (artificial intelligence); multilayer perceptrons; nonlinear programming; colored noise; harmonic retrieval; hierarchical learning; impulsive noise; independent sinusoidal signal sensitivity; nonlinear multilayer principal-component-type subspace learning algorithms; nonlinear optimization criteria; signal separation; Array signal processing; Covariance matrix; Information science; Laboratories; Neurons; Nonhomogeneous media; Principal component analysis; Signal processing; Signal processing algorithms; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks for Processing [1993] III. Proceedings of the 1993 IEEE-SP Workshop

Conference_Location

Linthicum Heights, MD

Print_ISBN

0-7803-0928-6

Type

conf

DOI

10.1109/NNSP.1993.471882

Filename

471882