Mean-Field Networks

TL;DR

This paper introduces mean field networks (MFNs), which convert the mean field inference algorithm into a neural network framework, enabling efficient inference and improved discriminative performance.

Contribution

The paper presents a novel neural network formulation of mean field inference, allowing for learned, flexible, and more effective approximate inference methods.

Findings

MFNs can learn to perform inference efficiently.

MFNs outperform traditional mean field in discriminative tasks.

Preliminary results show significant performance improvements.

Abstract

The mean field algorithm is a widely used approximate inference algorithm for graphical models whose exact inference is intractable. In each iteration of mean field, the approximate marginals for each variable are updated by getting information from the neighbors. This process can be equivalently converted into a feedforward network, with each layer representing one iteration of mean field and with tied weights on all layers. This conversion enables a few natural extensions, e.g. untying the weights in the network. In this paper, we study these mean field networks (MFNs), and use them as inference tools as well as discriminative models. Preliminary experiment results show that MFNs can learn to do inference very efficiently and perform significantly better than mean field as discriminative models.