| | |
Summary: An Augmented Lagrangian Approach to Constrained MAP Inference
Andr´e F. T. Martins
afm@cs.cmu.edu
M´ario A. T. Figueiredo
mario.figueiredo@lx.it.pt
Pedro M. Q. Aguiar aguiar@isr.ist.utl.pt
Noah A. Smith
nasmith@cs.cmu.edu
Eric P. Xing
epxing@cs.cmu.edu
School of Computer Science, Carnegie Mellon University, Pittsburgh, PA, USA
Instituto de Telecomunica¸c~oes / Instituto de Sistemas e Rob´otica, Instituto Superior T´ecnico, Lisboa, Portugal
Abstract
We propose a new algorithm for approxi-
mate MAP inference on factor graphs, which
combines augmented Lagrangian optimiza-
tion with the dual decomposition method.
Each slave subproblem is given a quadratic
|
