Title :
A new class of upper bounds on the log partition function
Author :
Wainwright, Martin J. ; Jaakkola, Tommi S. ; Willsky, Alan S.
Author_Institution :
Dept. of Electr. Eng., Univ. of California, Berkeley, CA, USA
fDate :
7/1/2005 12:00:00 AM
Abstract :
We introduce a new class of upper bounds on the log partition function of a Markov random field (MRF). This quantity plays an important role in various contexts, including approximating marginal distributions, parameter estimation, combinatorial enumeration, statistical decision theory, and large-deviations bounds. Our derivation is based on concepts from convex duality and information geometry: in particular, it exploits mixtures of distributions in the exponential domain, and the Legendre mapping between exponential and mean parameters. In the special case of convex combinations of tree-structured distributions, we obtain a family of variational problems, similar to the Bethe variational problem, but distinguished by the following desirable properties: i) they are convex, and have a unique global optimum; and ii) the optimum gives an upper bound on the log partition function. This optimum is defined by stationary conditions very similar to those defining fixed points of the sum-product algorithm, or more generally, any local optimum of the Bethe variational problem. As with sum-product fixed points, the elements of the optimizing argument can be used as approximations to the marginals of the original model. The analysis extends naturally to convex combinations of hypertree-structured distributions, thereby establishing links to Kikuchi approximations and variants.
Keywords :
Markov processes; backpropagation; belief networks; decision theory; geometry; inference mechanisms; information theory; parameter estimation; trees (mathematics); Bethe-Kikuchi free energy; Legendre mapping; MRF; Markov random field; approximate inference; approximating marginal distribution; belief propagation; combinatorial enumeration; information geometry; log partition function; parameter estimation; statistical decision theory; sum-product algorithm; tree-structured distributions; Belief propagation; Decision theory; Graphical models; Inference algorithms; Information geometry; Markov random fields; Parameter estimation; Partitioning algorithms; Random variables; Upper bound; Approximate inference; Bethe/Kikuchi free energy; Markov random field (MRF); belief propagation; factor graphs; graphical models; information geometry; partition function; sum–product algorithm; variational method;
Journal_Title :
Information Theory, IEEE Transactions on
DOI :
10.1109/TIT.2005.850091