We develop a novel algorithm, called VIP*, for structured variational approximate inference. This algorithm extends known algorithms to allow efficient multiple potential updates for overlapping clusters, and overcomes the difficulties imposed by deterministic constraints. The algorithms convergence is proven and its applicability demonstrated for genetic linkage analysis.

In: Journal of Artificial Intelligence Research

Details