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)
“A taxidermist stuffed a bear with such brio that he cried aloud “I love it!” A passing goddess, imperfectly understanding, but wishing to be responsive, kindly brought the beast to life. It consumed the artisan forthwith. Moral: Say what you mean.”
—Stan Washburn (b. 1943)