Belief propagation

< 1 >

Belief propagation is an algorithm for computing probabilities.

It is used in graphical models, where it updates what you believe about unknown variables by passing messages between connected nodes in a graph.

This algorithm was first proposed by Judea Pearl in 1982.