Lagrangian Relaxation for MAP Estimation in Graphical Models
Jason K. JohnsonDmitry M. MalioutovAlan S. WillskyPublished in: CoRR (2007)
Keyphrases
- lagrangian relaxation
- graphical models
- map estimation
- maximum a posteriori
- integer programming
- markov networks
- feasible solution
- lower bound
- probabilistic model
- belief propagation
- markov random field
- dynamic programming
- linear programming
- branch and bound algorithm
- random variables
- probabilistic graphical models
- maximum likelihood
- np hard
- lower and upper bounds
- column generation
- approximate inference
- bayesian networks
- map inference
- structure learning
- em algorithm
- image reconstruction
- expectation maximization
- message passing
- shortest path
- conditional random fields
- optimal solution
- energy function
- pairwise
- combinatorial optimization
- special case
- statistical relational learning