Sciweavers

Share
ICA
2004
Springer

Post-nonlinear Independent Component Analysis by Variational Bayesian Learning

11 years 5 months ago
Post-nonlinear Independent Component Analysis by Variational Bayesian Learning
Post-nonlinear (PNL) independent component analysis (ICA) is a generalisation of ICA where the observations are assumed to have been generated from independent sources by linear mixing followed by component-wise scalar nonlinearities. Most previous PNL ICA algorithms require the post-nonlinearities to be invertible functions. In this paper, we present a variational Bayesian approach to PNL ICA that also works for non-invertible post-nonlinearities. The method is based on a generative model with multi-layer perceptron (MLP) networks to model the post-nonlinearities. Preliminary results with a difficult artificial example are encouraging.
Alexander Ilin, Antti Honkela
Added 01 Jul 2010
Updated 01 Jul 2010
Type Conference
Year 2004
Where ICA
Authors Alexander Ilin, Antti Honkela
Comments (0)
books