HUGS: Combining Exact Inference and Gibbs Sampling in Junction Trees
This work addresses the challenge of solving complex inference problems that resist exact solutions, potentially expanding the class of tractable problems under bounded resources for researchers and practitioners in probabilistic graphical models.
The paper tackles the problem of performing inference in discrete Bayesian networks by combining exact inference and Gibbs sampling within junction trees, proposing an extension to the standard message passing scheme to enable this hybrid approach.
Dawid, Kjaerulff and Lauritzen (1994) provided a preliminary description of a hybrid between Monte-Carlo sampling methods and exact local computations in junction trees. Utilizing the strengths of both methods, such hybrid inference methods has the potential of expanding the class of problems which can be solved under bounded resources as well as solving problems which otherwise resist exact solutions. The paper provides a detailed description of a particular instance of such a hybrid scheme; namely, combination of exact inference and Gibbs sampling in discrete Bayesian networks. We argue that this combination calls for an extension of the usual message passing scheme of ordinary junction trees.