WERNER, T.: Revisiting the Linear Programming Relaxation Approach to Gibbs Energy Minimization and Weighted Constraint Satisfaction. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2010, vol. 32, no. 8, p. 1474-1488. ISSN 0162-8828. DOI: 10.1109/TPAMI.2009.134
We present a number of contributions to the LP relaxation approach to weighted constraint satisfaction (= Gibbs energy minimization). We link this approach to many works from constraint programming, which relation has so far been ignored in machine vision and learning. While the approach has been mostly considered only for binary constraints, we generalize it to n-ary constraints in a simple and natural way. This includes a simple algorithm to minimize the LP-based upper bound, n-ary max-sum diffusion – however, we consider using other bound-optimizing algorithms as well. The diffusion iteration is tractable for a certain class of higharity constraints represented as a black-box, which is analogical to propagators for global constraints CSP. Diffusion exactly solves permuted n-ary supermodular problems. A hierarchy of gradually tighter LP relaxations is obtained simply by adding various zero constraints and coupling them in various ways to existing constraints. Zero constraints can be added incrementally, which leads to a cutting plane algorithm. The separation problem is formulated as finding an unsatisfiable subproblem of a CSP.
cze
dc.language.iso
en
cze
dc.publisher
IEEE
cze
dc.relation.ispartof
IEEE Transactions on Pattern Analysis and Machine Intelligence. 2010, vol. 32, no. 8
(c) 2010 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other users, including reprinting/ republishing this material for advertising or promotional purposes, creating new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted components of this work in other works.
eng
dc.subject
Research Subject Categories::TECHNOLOGY::Electrical engineering, electronics and photonics
cze
dc.subject
weighted constraint satisfaction
eng
dc.subject
Gibbs distribution
eng
dc.subject
graphical model
eng
dc.subject
Markov random field
eng
dc.subject
linear programming relaxation
eng
dc.subject
marginal polytope
eng
dc.subject
cut polytope
eng
dc.subject
cutting plane algorithm
eng
dc.subject
global constraint
eng
dc.subject
supermodularity
eng
dc.subject
tree-reweighted max-product
eng
dc.title
Revisiting the Linear Programming Relaxation Approach to Gibbs Energy Minimization and Weighted Constraint Satisfaction