Variational message passing (VMP) is an approximate inference technique for continuous- or discrete-valued Bayesian networks, with conjugate-exponential parents, developed by John Winn. VMP was developed as a means of generalizing the approximate variational methods used by such techniques as Latent Dirichlet allocation and works by updating an approximate distribution at each node through messages in the node's Markov blanket.
Read more about Variational Message Passing: Likelihood Lower Bound, Determining The Update Rule, Messages in Variational Message Passing, Relationship To Exponential Families, VMP Algorithm, Constraints
Famous quotes containing the words message and/or passing:
“Never miss an opportunity to allow a child to do something she can and wants to on her own. Sometimes we’re in too much of a rush—and she might spill something, or do it wrong. But whenever possible she needs to learn, error by error, lesson by lesson, to do better. And the more she is able to learn by herself the more she gets the message that she’s a kid who can.”
—Polly Berrien Berends (20th century)
“The business eternally passing thro’ my mind and occupying it exclusively, wipes out at once the recollection of things which have been presented to it but once, and on which it has had no occasion to recur.”
—Thomas Jefferson (1743–1826)