| |
Abstract:
We introduce a multilayer generalization of independent
component analysis (ICA). At each level, this network extracts
real-valued latent variables that are non-linear functions of the
input data with a highly adaptive functional form, resulting in a
hierarchical distributed representation of these data. The network
is based on a probabilistic generative model, constructed by
cascading single-layer local ICA models. Whereas exact
maximum-likelihood learning for this model is intractable, we
present and demonstrate an algorithm that maximizes a lower bound
on the likelihood. This algorithm is developed by formulating a
variational approach to hierarchical ICA networks.
|
