MAP inference via Block-Coordinate Frank-Wolfe Algorithm
Machine Learning
2019-04-08 v2 Artificial Intelligence
Machine Learning
Abstract
We present a new proximal bundle method for Maximum-A-Posteriori (MAP) inference in structured energy minimization problems. The method optimizes a Lagrangean relaxation of the original energy minimization problem using a multi plane block-coordinate Frank-Wolfe method that takes advantage of the specific structure of the Lagrangean decomposition. We show empirically that our method outperforms state-of-the-art Lagrangean decomposition based algorithms on some challenging Markov Random Field, multi-label discrete tomography and graph matching problems.
Cite
@article{arxiv.1806.05049,
title = {MAP inference via Block-Coordinate Frank-Wolfe Algorithm},
author = {Paul Swoboda and Vladimir Kolmogorov},
journal= {arXiv preprint arXiv:1806.05049},
year = {2019}
}