Fast Convergent Algorithms for Expectation Propagation Approximate Bayesian Inference

Abstract

We propose a novel algorithm to solve the expectation propagation relaxation of Bayesian inference for continuous-variable graphical models. In contrast to most previous algorithms, our method is provably convergent. By marrying convergent EP ideas from (Opper&Winther 05) with covariance decoupling techniques (Wipf&Nagarajan 08, Nickisch&Seeger 09), it runs at least an order of magnitude faster than the most commonly used EP solver.

16 pages, 3 figures, submitted for conference publication

References in corpus (1)

• Expectation Propagation for approximate Bayesian inference

Cited by in corpus (5)

• Gaussian Probabilities and Expectation Propagation
• Expectation Propagation for Neural Networks with Sparsity-promoting Priors
• Gaussian Process Regression with a Student-t Likelihood
• Perturbative Corrections for Approximate Inference in Gaussian Latent Variable Models
• Bayes-Newton Methods for Approximate Bayesian Inference with PSD Guarantees