Lagrangian Relaxation for Inference in Natural Language Processing.
Michael CollinsPublished in: IWPT (2011)
Keyphrases
- lagrangian relaxation
- natural language processing
- dual decomposition
- lower bound
- integer programming
- inference in graphical models
- feasible solution
- column generation
- branch and bound algorithm
- lower and upper bounds
- information extraction
- linear programming
- capacity constraints
- np hard
- machine learning
- dynamic programming
- map inference
- relaxation algorithm
- markov logic
- mixed integer program
- natural language
- upper bound
- graphical models
- coreference resolution
- lagrangian heuristic
- shortest path
- subgradient method
- knowledge representation
- bayesian networks
- markov random field
- neural network
- mixed integer linear programs
- feature space
- learning algorithm
- structured prediction
- reinforcement learning
- optimal solution
- energy minimization