Large margin methods for structured and interdependent output variables

Large margin methods for structured and interdependent output variables (2005)

Cached

Download Links

[jmlr.csail.mit.edu]
[www.jmlr.org]

Other Repositories/Bibliography

DBLP

by Ioannis Tsochantaridis , Thorsten Joachims , Thomas Hofmann , Yasemin Altun

Venue:	JOURNAL OF MACHINE LEARNING RESEARCH
Citations:	208 - 10 self

Versions

Version History

Metadata Version 2

User correction supplied by mph

Datum	Value	Source
TITLE	Large margin methods for structured and interdependent output variables	INFERENCE
AUTHOR NAME	Ioannis Tsochantaridis	user correction
AUTHOR AFFIL	Google Inc.	user correction
AUTHOR ADDR	Mountain View, CA 94043, USA	user correction
AUTHOR NAME	Thorsten Joachims	user correction
AUTHOR AFFIL	Department of Computer Science; Cornell University	user correction
AUTHOR ADDR	Ithaca, NY 14853, USA	user correction
AUTHOR NAME	Thomas Hofmann	user correction
AUTHOR AFFIL	Darmstadt University of Technology; Fraunhofer IPSI	user correction
AUTHOR ADDR	Darmstadt, Germany	user correction
AUTHOR NAME	Yasemin Altun	user correction
AUTHOR AFFIL	Toyota Technological Institute	user correction
AUTHOR ADDR	Chicago, IL 60637, USA	user correction
ABSTRACT	Learning general functional dependencies between arbitrary input and output spaces is one of the key challenges in computational intelligence. While recent progress in machine learning has mainly focused on designing flexible and powerful input representations, this paper addresses the complementary issue of designing classification algorithms that can deal with more complex outputs, such as trees, sequences, or sets. More generally, we consider problems involving multiple dependent output variables, structured output spaces, and classification problems with class attributes. In order to accomplish this, we propose to appropriately generalize the well-known notion of a separation margin and derive a corresponding maximum-margin formulation. While this leads to a quadratic program with a potentially prohibitive, i.e. exponential, number of constraints, we present a cutting plane algorithm that solves the optimization problem in polynomial time for a large class of problems. The proposed method has important applications in areas such as computational biology, natural language processing, information retrieval/extraction, and optical character recognition. Experiments from various domains involving different types of output spaces emphasize the breadth and generality of our approach.	user correction
YEAR	2005	INFERENCE
VENUE	JOURNAL OF MACHINE LEARNING RESEARCH	user correction
VENUE TYPE	JOURNAL	INFERENCE
PAGES	1453--1484	INFERENCE
VOLUME	6	INFERENCE
CITATIONS	33 found	ParsCit 1.0