Abstract

i Abstract This thesis is concerned with the problem of Blind Source Separation. Specifically we considerthe Independent Component Analysis (ICA) model in which a set of observations are modelled by xt = Ast: (1) where A is an unknown mixing matrix and st is a vector of hidden source components attime t. The ICA problem is to find the sources given only a set of observations. In chapter 1, the blind source separation problem is introduced. In chapter 2 the methodof Ensemble Learning is explained. Chapter 3 applies Ensemble Learning to the ICA model and chapter 4 assesses the use of Ensemble Learning for model selection.Chapters 5-7 apply the Ensemble Learning ICA algorithm to data sets from physics (a medical imaging data set consisting of images of a tooth), biology (data sets from cDNAmicro-arrays) and astrophysics (Planck image separation and galaxy spectra separation).

Citations

6232	Maximum likelihood from incomplete data via the EM algorithm - Dempster, Laird, et al. - 1977
5666	Probabilistic reasoning in intelligent systems - Pearl - 1988
4404	The Mathematical Theory of Communication - Shannon, Weaver - 1963
3909	Neural Networks for Pattern Recognition - Bishop - 1995
1176	Cluster analysis and display of genome-wide expression patterns - Eisen, Spellman, et al. - 1998
1034	Independent component analysis, a new concept - Comon - 1994
1019	Oja E: Independent Component Analysis - Hyvarinen, Karhunen - 2001
896	BD: Pattern Recognition and Neural Networks - Ripley - 1996
872	Information theory and statistics - Kullback - 1959
698	Learning the parts of objects by non-negative matrix factorization - Lee, Seung - 1999
513	Bayesian Learning for Neural Networks - Neal - 1996
449	Probabilistic inference using Markov chain Monte Carlo methods - Neal - 1993
429	A new learning algorithm for blind signal separation - SI, Cichoki, et al. - 1996
427	Sparse coding with an overcomplete basis set: A strategy employed by V1? Vision Research - Olshausen, Field - 1997
426	D.: Markov chain Monte Carlo in practice - Gilks, Richardson, et al. - 1996
421	Fast learning in networks of locally-tuned processing units - Moody, Darken - 1989
377	Parallel distributed processing: Explorations in the microstructure of cognition - Rumelhart, McClelland - 1986
368	Blind separation of sources, part i: An adaptive algorithm based on neuromimetic architecture - Jutten, Herault - 1991
359	Probabilistic principal component analysis - Tipping, Bishop - 1999
325	Equivariant adaptative source separation - Cardoso, Labeld - 1996
273	Yakhini Z: Clustering Gene Expression Patterns - Ben-Dor, Shamir - 1999
215	Natural gradient works efficiently in learning - Amari - 1998
188	Learning overcomplete representations - Lewicki, Sejnowski - 2000
162	Differential-Geometrical Methods in Statistics - Amari - 1985
160	Connectionist learning of belief networks - Neal - 1992
150	Independent component analysis of electroencephalographic data - Makeig, Bell, et al. - 1996
147	al.Molecular classification of cancer -class discovery and class prediction by gene expression monitoring .Science,286(5439):531-537(1999 - Golub
146	Multidimensional independent component analysis - Cardoso
134	The evidence framework applied to classification networks - MacKay - 1992
132	Yooseph S: Exploring expression data: identification and analysis of coexpressed genes. Genome Res - LJ, Kruglyak - 1999
128	Separation of a mixture of independent signals using time delayed correlations - Molgedey, Schuster - 1994
120	Probability, frequency and reasonable expectation - Cox - 1946
120	M.: Variational inference for Bayesian mixtures of factor analysers - Ghahramani, Beal - 2000
113	Keeping neural networks simple by minimizing the description length of weights - Hinton, Camp - 1993
112	An Information Maximisation Approach to Blind Separation and - Bell, Sejnowski - 1995
106	Using graphical models and genomic expression data to statistically validate models of genetic regulatory networks - Hartemink, Gifford, et al. - 2001
105	Independant Component Analysis "Theory and Applications - Lee - 1998
102	Mean field theory for sigmoid belief networks - Zemel, Saul, et al. - 1996
101	Maximum likelihood and covariant algorithms for independent component analysis - MacKay - 1996
98	Autoencoders, minimum description length, and Helmholtz free energy - Hinton, Zemel - 1994
98	Adaptive online learning algorithms for blind separation: Maximum entropy and minimum mutual information - Yang, Amari - 1997
94	A theory of adaptive pattern classifiers - Amari - 1967
83	Exploiting tractable substructures in intractable networks - Saul, Jordan - 1996
71	Blind source separation of more sources than mixtures using overcomplete representation - Lee, Lewicki, et al. - 1999
71	separation of convolved sources based on information maximization - Torkkola, \Blind - 1996
70	Multichannel blind deconvolution and equalization using the natural gradient - Amari, Douglas, et al. - 1997
70	Nonlinear independent component analysis: Existence and uniqueness results - Hyvärinen, Pajunen - 1999
66	D.J.C.: Dasher - a data entry interface using continuous gestures and language models - Ward, Blackwell, et al. - 2000
64	Good codes based on very sparse matrices - MacKay, Neal - 1995
60	Maximum likelihood blind source separation: A context-sensitive generalization - Pearlmutter, Parra - 1997