Belief Propagation

This is a method proposed in Judea Pearl, 1982: Reverend Bayes on inference engines: A distributed hierarchical approach, AAAI 1982.

Belief Propagation is a message passing inference method for statistical graphical models (e.g. Markov random fields), especially when the graphical model is both a factor graph and a tree (it can compute exact marginals). The basic idea of belief propagation is to compute the marginal distribution of unobserved nodes, based on the conditional distribution of observed nodes.