Partial Optimality by Pruning for MAP-Inference with General Graphical Models
Autor: | Bogdan Savchynskyy, Paul Swoboda, Jörg Hendrik Kappes, Christoph Schnörr, Alexander Shekhovtsov |
---|---|
Rok vydání: | 2015 |
Předmět: |
FOS: Computer and information sciences
Mathematical optimization Computer Science - Artificial Intelligence Markov process 02 engineering and technology 010501 environmental sciences 01 natural sciences symbols.namesake Artificial Intelligence 0202 electrical engineering electronic engineering information engineering Graphical model Pruning (decision trees) Time complexity 0105 earth and related environmental sciences Mathematics Random field Markov chain Applied Mathematics Solver Artificial Intelligence (cs.AI) Computational Theory and Mathematics symbols 020201 artificial intelligence & image processing Computer Vision and Pattern Recognition Minification Software |
Zdroj: | IEEE transactions on pattern analysis and machine intelligence. 38(7) |
ISSN: | 1939-3539 |
Popis: | We consider the energy minimization problem for undirected graphical models, also known as MAP-inference problem for Markov random fields which is NP-hard in general. We propose a novel polynomial time algorithm to obtain a part of its optimal non-relaxed integral solution. Our algorithm is initialized with variables taking integral values in the solution of a convex relaxation of the MAP-inference problem and iteratively prunes those, which do not satisfy our criterion for partial optimality. We show that our pruning strategy is in a certain sense theoretically optimal. Also empirically our method outperforms previous approaches in terms of the number of persistently labelled variables. The method is very general, as it is applicable to models with arbitrary factors of an arbitrary order and can employ any solver for the considered relaxed problem. Our method's runtime is determined by the runtime of the convex relaxation solver for the MAP-inference problem. 16 pages, 4 tables and 4 figures |
Databáze: | OpenAIRE |
Externí odkaz: |