Publications [#329999] of Alan E. Gelfand
search .Papers Published
- Gelfand, AE; Welling, M. "Generalized belief propagation on tree robust structured region graphs." Uncertainty in Artificial Intelligence Proceedings of the 28th Conference, Uai 2012 (December, 2012): 296-305.
(last updated on 2023/06/01)Abstract:
This paper provides some new guidance in the construction of region graphs for Generalized Belief Propagation (GBP). We connect the problem of choosing the outer regions of a Loop- Structured Region Graph (SRG) to that of finding a fundamental cycle basis of the corresponding Markov network. We also define a new class of tree-robust Loop-SRG for which GBP on any induced (spanning) tree of the Markov network, obtained by setting to zero the off-tree interactions, is exact. This class of SRG is then mapped to an equivalent class of tree-robust cycle bases on the Markov network. We show that a treerobust cycle basis can be identified by proving that for every subset of cycles, the graph obtained from the edges that participate in a single cycle only, is multiply connected. Using this we identify two classes of tree-robust cycle bases: planar cycle bases and star cycle bases. In experiments we show that tree-robustness can be successfully exploited as a design principle to improve the accuracy and convergence of GBP.