Learning a contingently acyclic, probabilistic relational model of a social network

ID
TR-2009-08
Authors
Peter Carbonetto, Jacek Kisynski, Michael Chiang and David Poole
Publishing date
April 06, 2009
Length
14 pages
Abstract
We demonstrate through experimental comparisons that modeling relations in a social network with a directed probabilistic model provides a viable alternative to the standard undirected graphical model approach. Our model incorporates special latent variables to guarantee acyclicity. We investigate the inference and learning challenges entailed by our approach.