Loading...
Please wait, while we are loading the content...
Similar Documents
An augmented Lagrangian approach to constrained MAP inference (2011)
| Content Provider | CiteSeerX |
|---|---|
| Author | Xing, Eric P. Aguiar, Pedro M. Q. Smith, Noah A. Figueiredo, Mário A. T. Martins, André F. T. |
| Abstract | We propose a new algorithm for approximate MAP inference on factor graphs, by combining augmented Lagrangian optimization with the dual decomposition method. Each slave subproblem is given a quadratic penalty, which pushes toward faster consensus than in previous subgradient approaches. Our algorithm is provably convergent, parallelizable, and suitable for fine decompositions of the graph. We show how it can efficiently handle problems with (possibly global) structural constraints via simple sort operations. Experiments on synthetic and real-world data show that our approach compares favorably with the state-of-the-art. 1. |
| File Format | |
| Publisher Date | 2011-01-01 |
| Publisher Institution | In Proceedings of the 28th International Conference on Machine Learning (ICML |
| Access Restriction | Open |
| Subject Keyword | Augmented Lagrangian Optimization Augmented Lagrangian Approach Slave Subproblem Fine Decomposition Real-world Data Show Approximate Map Inference New Algorithm Map Inference Structural Constraint Factor Graph Simple Sort Operation Previous Subgradient Approach Quadratic Penalty Dual Decomposition Method |
| Content Type | Text |
| Resource Type | Proceeding |
