ICA-CNL website, new ICA paper & matlab code.

[Te-Won Lee]

Dear Connectionists,

*** Visit the ICA - CNL website! ***

	http://www.cnl.salk.edu/~tewon/ica_cnl.html

Independent Component Analysis (ICA) has received a lot
of attention recently. The ICA - CNL website shows a 
variety of ICA research performed in the Computational
Neuroscience Laborabory (CNL) at the Salk Institute
(Terry Sejnowski's Lab).

The goal of this webpage is to provide detailed services to
scientists, engineers and industrials about ICA 
or Blind Source Separation (BSS).  The following
services are available:

- Introduction to ICA
- ICA Projects
- Researchers
- Papers
- Matlab Code
- ICA Demos
- ICA News
- Links
- ICA FAQs

If you have questions or comments please send email 
to tewon at salk.edu


*** New ICA paper and Matlab code available ***

	"Independent component analysis using an extended infomax algorithm 
	for mixed sub-Gaussian and super-Gaussian sources"

	T-W. Lee, M. Girolami and T.J. Sejnowski.
	to appear in Neural Computation, MIT Press.

	Paper:	
	http://www.cnl.salk.edu/~tewon/Public/nc97.ps.gz
	(1470k, 33 pages)

	Matlab code:
	http://www.cnl.salk.edu/~tewon/ica_cnl.html
	and go to Matlab Code and download the extended infomax algorithm


Abstract:

An extension of the infomax algorithm of Bell and Sejnowski (1995) 
is presented that is able to blindly separate mixed signals with sub- and 
super-Gaussian source distributions.  This was achieved by using a
simple type of learning rule first derived by Girolami (1997)
by choosing negentropy as a projection pursuit index. Parameterized
probability distributions that have sub- and super-Gaussian regimes
were used to derive a general learning rule that preserves the simple
architecture proposed by Bell and Sejnowski (1995), is optimized using the
natural gradient by Amari (1997), and uses the stability analysis
of Cardoso and Laheld (1996) to switch between sub- and super-Gaussian
regimes. We demonstrate that the extended infomax algorithm is able to
easily separate 20 sources with a variety of source distributions.
Applied to high-dimensional data from electroencephalographic (EEG)
recordings, it is effective at separating artifacts such as eye blinks
and line noise from weaker electrical signals that arise from sources
in the brain.




-- 
----------------------------------------------------------------------
Dr. Te-Won Lee                   EMAIL: tewon at salk.edu
Computational Neurobiology Lab,   WORK: (619) 453-4100 x1215 
Salk Institute,                   HOME: (619) 450-9036 
10010 N. Torrey Pines Rd.          FAX: (619) 587-0417
La Jolla, CA  92037                WEB: http://www.cnl.salk.edu/~tewon 
----------------------------------------------------------------------