Rounding-based moves for semi-metric labeling

Journal article

Semi-metric labeling is a special case of energy minimization for pairwise Markov random fields. The energy function consists of arbitrary unary potentials, and pairwise potentials that are proportional to a given semi-metric distance function over the label set. Popular methods for solving semi-metric labeling include (i) move-making algorithms, which iteratively solve a minimum st-cut problem; and (ii) the linear programming (LP) relaxation based approach. In order to convert the fractional solution of the LP relaxation to an integer solution, several randomized rounding procedures have been developed in the literature. We consider a large class of parallel rounding procedures, and design move-making algorithms that closely mimic them. We prove that the multiplicative bound of a move-making algorithm exactly matches the approximation factor of the corresponding rounding procedure for any arbitrary distance function. Our analysis includes all known results for move-making algorithms as special cases.

Authors: Kumar, M; Dokania, P

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: MIT Press
• Journal: Journal of Machine Learning Research
• Publication date: 2016-04-01
• Acceptance date: 2016-04-23
• EISSN: 1533-7928
• ISSN: 1532-4435
• Keywords: multiplicative bounds; linear programming relaxation; semi-metric labeling; move-making algorithms