Metamorphosis Networks: An Alternative to Constructive Methods

Given a set of training examples, determining the appropriate number of free parameters is a challenging problem. Constructive learning algorithms attempt to solve this problem automatically by adding hidden units, and therefore free parameters, during learning. We explore an alternative class of algorithms|called metamorphosis algorithms|in which the number of units is xed, but the number of free parameters gradually increases during learning. The architecture we investigate is composed of RBF units on a lattice , which imposes exible constraints on the parameters of the network. Virtues of this approach include variable subset selection , robust parameter selection, multiresolution processing, and interpolation of sparse training data.