A New
Learning Algorithm for Mean Field Boltzmann Machines
Max Welling and G.E. Hinton
Gatsby Computational Neuroscience Unit
GCNU TR 2001-002
Abstract
We present a new learning algorithm for Mean Field Boltzmann
Machines based on the contrastive divergence optimization criterion. In addition to
mini-mizing the divergence between the data distribution and the equilibrium
dis-tribution that the network believes in, we maximize the divergence between
one-step reconstructions of the data and the equilibrium distribution. This eliminates the
need to estimate equilibrium statistics, so we do not need to ap-proximate the multimodal
probablility distribution of the free network with the unimodal mean field distribution.
We test the learning algorithm on the classification of digits.
Download: [ps.gz] or [pdf]