Computing Divergences between Discrete Decomposable Models

There are many applications that benefit from computing the exact divergence between 2 discrete probability measures, including machine learning. Unfortunately, in absence of any assumptions on structure or independencies within these distributions, them is an intractable problem high dimensions. We show we able to compute a wide family functionals and divergences, such as alpha-beta divergence, two decomposable models, i.e. chordal Markov networks, time exponential treewidth models. The divergences include popular Kullback-Leibler Hellinger distance, chi-squared divergence. Thus, can accurately values this broad class extent which model distributions using