preprint
Generalized sequential tree-reweighted message passing
published
Vladimir
Kolmogorov
author 3D50B0BA-F248-11E8-B48F-1D18A9856A87
Thomas
Schoenemann
author
VlKo
department
This paper addresses the problem of approximate MAP-MRF inference in general graphical models. Following [36], we consider a family of linear programming relaxations of the problem where each relaxation is specified by a set of nested pairs of factors for which the marginalization constraint needs to be enforced. We develop a generalization of the TRW-S algorithm [9] for this problem, where we use a decomposition into junction chains, monotonic w.r.t. some ordering on the nodes. This generalizes the monotonic chains in [9] in a natural way. We also show how to deal with nested factors in an efficient way. Experiments show an improvement over min-sum diffusion, MPLP and subgradient ascent algorithms on a number of computer vision and natural language processing problems.
ArXiv2012
eng
arXiv
1205.6352
16
V. Kolmogorov, T. Schoenemann, ArXiv (2012).
Kolmogorov V, Schoenemann T. Generalized sequential tree-reweighted message passing. <i>arXiv</i>. 2012.
Kolmogorov, Vladimir, and Thomas Schoenemann. “Generalized Sequential Tree-Reweighted Message Passing.” <i>ArXiv</i>. ArXiv, 2012.
V. Kolmogorov and T. Schoenemann, “Generalized sequential tree-reweighted message passing,” <i>arXiv</i>. ArXiv, 2012.
Kolmogorov V, Schoenemann T. 2012. Generalized sequential tree-reweighted message passing. arXiv.
Kolmogorov, V., & Schoenemann, T. (2012). Generalized sequential tree-reweighted message passing. <i>ArXiv</i>. ArXiv.
Kolmogorov, Vladimir, and Thomas Schoenemann. “Generalized Sequential Tree-Reweighted Message Passing.” <i>ArXiv</i>, ArXiv, 2012.
29282018-12-11T12:00:23Z2020-05-12T08:50:48Z