Distributed MAP Inference for Undirected Graphical Models

In this work, we distribute the MCMC-based MAP inference using the Map-Reduce framework. The variables are assigned randomly to machines, which leads to some factors that neighbor vari- ables on separ