Gauges, Loops, and Polynomials for Partition Functions of Graphical Models

11/12/2018
by   Michael Chertkov, et al.
0

We suggest a new methodology for analysis and approximate computations of the Partition Functions (PF) of Graphical Models (GM) in the Normal Factor Graph representation that combines the gauge transformation (GT) technique from (Chertkov, Chernyak 2006) with the technique developed in (Straszak, Vishnoi 2017) based on the recent progress in the field of real stable polynomials. We show that GTs (while keeping PF invariant) allow representation of PF as a sum of polynomials of variables associated with edges of the graph. A special belief propagation (BP) gauge makes a single out term of the series least sensitive to variations then resulting in the loop series for PF introduced in (Chertkov, Chernyak 2006). In addition to restating the known results in the polynomial form, we also discover a new relation between the computationally tractable BP term (single out term of the loop series evaluated at the BP gauge) and the PF: sequential application of differential operators, each associated with an edge of the graph, to the BP polynomial results in the PF. Each term in the sequence corresponds to a BP polynomial of a modified GM derived by contraction of an edge. Even though complexity of computing factors in the derived GMs grow exponentially with the number of eliminated edges, polynomials associated with the new factors remain real stable if the original factors have this property. Moreover, we show that BP estimations for the PF do not decrease with eliminations, thus resulting overall in the proof that the BP solution of the original GM gives a lower bound for PF. The proof extends results of (Straszak, Vishnoi 2017) from bipartite to general graphs, however, it is limited to the case when the BP solution is feasible.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
03/03/2017

Gauging Variational Inference

Computing partition function is the most important statistical inference...
research
08/08/2017

Belief Propagation, Bethe Approximation and Polynomials

Factor graphs are important models for succinctly representing probabili...
research
08/09/2014

Approximate inference on planar graphs using Loop Calculus and Belief Propagation

We introduce novel results for approximate inference on planar graphical...
research
05/09/2012

Counting Belief Propagation

A major benefit of graphical models is that most knowledge is captured i...
research
09/08/2011

The Complexity of Approximating a Bethe Equilibrium

This paper resolves a common complexity issue in the Bethe approximation...
research
05/29/2016

MCMC assisted by Belief Propagaion

Markov Chain Monte Carlo (MCMC) and Belief Propagation (BP) are the most...
research
02/24/2023

Maximal Clique and Edge-Ranking Bounds of Biclique Cover Number

The biclique cover number (bc) of a graph G denotes the minimum number o...

Please sign up or login with your details

Forgot password? Click here to reset