31 search results - page 1 / 7 » Expectation Consistent Free Energies for Approximate Inferen... |
NIPS
2004
13 years 5 months ago
2004
We propose a novel a framework for deriving approximations for intractable probabilistic models. This framework is based on a free energy (negative log marginal likelihood) and ca...
UAI
2003
13 years 5 months ago
2003
Loopy and generalized belief propagation are popular algorithms for approximate inference in Markov random fields and Bayesian networks. Fixed points of these algorithms correspo...
CVPR
2008
IEEE
14 years 6 months ago
2008
IEEE
Belief Propagation (BP) can be very useful and efficient for performing approximate inference on graphs. But when the graph is very highly connected with strong conflicting intera...
JAIR
2006
13 years 4 months ago
2006
Loopy and generalized belief propagation are popular algorithms for approximate inference in Markov random fields and Bayesian networks. Fixed points of these algorithms have been...
UAI
2001
13 years 5 months ago
2001
This paper presents a new deterministic approximation technique in Bayesian networks. This method, "Expectation Propagation," unifies two previous techniques: assumed-de...