Bayesian Nonlinear Independent Component Analysis by Multi-Layer Perceptrons

In Advances in Independent Component Analysis, ed. Mark Girolami (Springer, 2000). In press.

Harri Lappalainen - Antti Honkela

Helsinki University of Technology, Neural Networks Research Centre, P.O. Box 5400, FIN-02015 HUT, Finland

Abstract:

In this chapter, a nonlinear extension to independent component analysis is developed. The nonlinear mapping from source signals to observations is modelled by a multi-layer perceptron network and the distributions of source signals are modelled by mixture-of-Gaussians. The observations are assumed to be corrupted by Gaussian noise and therefore the method is more adequately described as nonlinear independent factor analysis. The nonlinear mapping, the source distributions and the noise level are estimated from the data. Bayesian approach to learning avoids problems with overlearning which would otherwise be severe in unsupervised learning with flexible nonlinear models.