Tighter relaxations for MAP-MRF optimization via Singleton Arc Consistency

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lev-Ran, Asaf, Arkhipov, Pavel, Kolmogorov, Vladimir
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918499615506432
author Lev-Ran, Asaf
Arkhipov, Pavel
Kolmogorov, Vladimir
author_facet Lev-Ran, Asaf
Arkhipov, Pavel
Kolmogorov, Vladimir
contents We consider the MAP-MRF inference task, that is, minimizing a function of discrete variables represented as a sum of unary and pairwise terms. A prominent approach for tackling this NP-hard problem in practice is to solve its natural LP relaxation and then iteratively tighten the relaxation by adding clusters. Based on some theoretical observations, we propose a new technique for identifying such clusters. It works by running the Singleton Arc Consistency algorithm in a certain CSP instance. Experimental results indicate that the new tightening technique outperforms the previous approach by [Sontag et al. UAI 2012] that searches for frustrated cycles. Our code will be made available at https://github.com/vnk-ist/MAP-MRF/.
format Preprint
id arxiv_https___arxiv_org_abs_2605_13392
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Tighter relaxations for MAP-MRF optimization via Singleton Arc Consistency
Lev-Ran, Asaf
Arkhipov, Pavel
Kolmogorov, Vladimir
Data Structures and Algorithms
We consider the MAP-MRF inference task, that is, minimizing a function of discrete variables represented as a sum of unary and pairwise terms. A prominent approach for tackling this NP-hard problem in practice is to solve its natural LP relaxation and then iteratively tighten the relaxation by adding clusters. Based on some theoretical observations, we propose a new technique for identifying such clusters. It works by running the Singleton Arc Consistency algorithm in a certain CSP instance. Experimental results indicate that the new tightening technique outperforms the previous approach by [Sontag et al. UAI 2012] that searches for frustrated cycles. Our code will be made available at https://github.com/vnk-ist/MAP-MRF/.
title Tighter relaxations for MAP-MRF optimization via Singleton Arc Consistency
topic Data Structures and Algorithms
url https://arxiv.org/abs/2605.13392