Probabilistic reasoning in intelligent systems: networks of plausible inference
Probabilistic reasoning in intelligent systems: networks of plausible inference
Decomposable negation normal form
Journal of the ACM (JACM)
Resolution versus Search: Two Strategies for SAT
Journal of Automated Reasoning
Mini-buckets: A general scheme for bounded inference
Journal of the ACM (JACM)
Using weighted MAX-SAT engines to solve MPE
Eighteenth national conference on Artificial intelligence
Compiling propositional weighted bases
Artificial Intelligence - Special issue on nonmonotonic reasoning
Journal of Artificial Intelligence Research
Compiling Bayesian networks with local structure
IJCAI'05 Proceedings of the 19th international joint conference on Artificial intelligence
Variable and value ordering for MPE search
IJCAI'09 Proceedings of the 21st international jont conference on Artifical intelligence
Structural relaxations by variable renaming and their compilation for solving MinCostSAT
CP'07 Proceedings of the 13th international conference on Principles and practice of constraint programming
A comparative study of energy minimization methods for markov random fields
ECCV'06 Proceedings of the 9th European conference on Computer Vision - Volume Part II
Relax, compensate and then recover: a theory of anytime, approximate inference
JELIA'10 Proceedings of the 12th European conference on Logics in artificial intelligence
Relax, compensate and then recover
JSAI-isAI'10 Proceedings of the 2010 international conference on New Frontiers in Artificial Intelligence
Hi-index | 0.00 |
We introduce a new approach to approximating weighted Max-SAT problems that is based on simplifying a given instance, and then tightening the approximation. First, we relax its structure until it is tractable for exact algorithms. Second, we compensate for the relaxation by introducing auxiliary weights. More specifically, we relax equivalence constraints from a given Max-SAT problem, which we compensate for by recovering a weaker notion of equivalence. We provide a simple algorithm for finding these approximations, that is based on iterating over relaxed constraints, compensating for them one-by-one. We show that the resulting Max-SAT instances have certain interesting properties, both theoretical and empirical.