The Convergence of Contrastive Divergences

This paper analyses the Contrastive Divergence algorithm for learning statistical parameters. We relate the algorithm to the stochastic approximation literature. This enables us to specify conditions under which the algorithm is guaranteed to converge to the optimal solution (with probability 1). This includes necessary and sufficient conditions for the solution to be unbiased.

• Type: Conference
• Year: 2004
• Where: NIPS
• Authors: Alan L. Yuille
• Keywords: Contrastive Divergence Algorithm, NIPS 2004, NIPS 2007, Statistical Parameters, Stochastic Approximation Literature