Learning Convex Inference of Marginals
Published on Jul 30, 20083291 Views
Graphical models trained using maximum likelihood are a common tool for probabilistic inference of marginal distributions. However, this approach suffers difficulties when either the inference process