Structured prediction, dual extragradient and Bregman projections

Structured prediction, dual extragradient and Bregman projections (2006)

by Ben Taskar , Simon Lacoste-julien , Michael I. Jordan

Venue:	Journal of Machine Learning Research
Citations:	30 - 2 self

Datum	Value	Source
TITLE	Structured prediction, dual extragradient and Bregman projections	INFERENCE
AUTHOR NAME	Ben Taskar	SVM HeaderParse 0.2
AUTHOR NAME	Simon Lacoste-julien	SVM HeaderParse 0.2
AUTHOR NAME	Michael I. Jordan	SVM HeaderParse 0.2
AUTHOR AFFIL	Tech Report 697; Department of Statistics; University of California, Berkeley	SVM HeaderParse 0.2
ABSTRACT	We present a simple and scalable algorithm for maximum-margin estimation of structured output models, including an important class of Markov networks and combinatorial models. We formulate the estimation problem as a convex-concave saddle-point problem that allows us to use simple projection methods based on the dual extragradient algorithm (Nesterov, 2003). The projection step can be solved using dynamic programming or combinatorial algorithms for min-cost convex flow, depending on the structure of the problem. We show that this approach provides a memory-efficient alternative to formulations based on reductions to a quadratic program (QP). We analyze the convergence of the method and present experiments on two very different structured prediction tasks: 3D image segmentation and word alignment, illustrating the favorable scaling properties of our algorithm. 1 1.	SVM HeaderParse 0.2
YEAR	2006	INFERENCE
VENUE	Journal of Machine Learning Research	INFERENCE
VENUE TYPE	JOURNAL	INFERENCE
PAGES	1627--1653	INFERENCE
VOLUME	7	INFERENCE
TECH	Technical report	INFERENCE
CITATIONS	38 found	ParsCit 1.0