Semi-supervised Learning of Classifiers: Theory, Algorithms and Their Application to Human-Computer Interaction

Semi-supervised Learning of Classifiers: Theory, Algorithms and Their Application to Human-Computer Interaction (2004)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Ira Cohen , Fabio G. Cozman , Nicu Sebe , Marcelo C. Cirelo , Thomas S. Huang

Venue:	IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
Citations:	47 - 14 self

Versions

Version History

Metadata Version 1

Datum	Value	Source
TITLE	Semisupervised learning of classifiers: Theory, algorithms, and their application to human-computer interaction	INFERENCE
AUTHOR NAME	Ira Cohen	user correction - Legacy Corrections
AUTHOR AFFIL	Ira Cohen: Hewlett-Packard Labs, Palo Alto, CA, USA,	user correction - Legacy Corrections
AUTHOR NAME	Nicu Sebe	user correction - Legacy Corrections
AUTHOR AFFIL	Ira Cohen: Hewlett-Packard Labs, Palo Alto, CA, USA,	user correction - Legacy Corrections
AUTHOR NAME	Marcelo C. Cirelo	user correction - Legacy Corrections
AUTHOR AFFIL	Ira Cohen: Hewlett-Packard Labs, Palo Alto, CA, USA,	user correction - Legacy Corrections
AUTHOR NAME	Thomas S. Huang	user correction - Legacy Corrections
AUTHOR AFFIL	Ira Cohen: Hewlett-Packard Labs, Palo Alto, CA, USA,	user correction - Legacy Corrections
AUTHOR NAME	Fabio G. Cozman	user correction - Legacy Corrections
AUTHOR AFFIL	; Thomas S. Huang: Beckman Institute, University of Illinois at Urbana-Champaign, USA.; 1	user correction - Legacy Corrections
AUTHOR ADDR	; Nicu Sebe: Faculty of Science, University of Amsterdam, The Netherlands.	user correction - Legacy Corrections
AUTHOR NAME	Fabio G. Cozman	user correction - Legacy Corrections
AUTHOR AFFIL	; Thomas S. Huang: Beckman Institute, University of Illinois at Urbana-Champaign, USA.; 1	user correction - Legacy Corrections
AUTHOR ADDR	; Nicu Sebe: Faculty of Science, University of Amsterdam, The Netherlands.	user correction - Legacy Corrections
ABSTRACT	Automatic classification is one of the basic tasks required in any pattern recognition and human computer interaction application. In this paper we discuss training probabilistic classifiers with labeled and unlabeled data. We provide a new analysis that shows under what conditions unlabeled data can be used in learning to improve classification performance. We also show that if the conditions are violated, using unlabeled data can be detrimental to classification performance. We discuss the implications of this analysis to a specific type of probabilistic classifiers, Bayesian networks, and propose a new structure learning algorithm that can utilize unlabeled data to improve classification. Finally, we show how the resulting algorithms are successfully employed in two applications related to human-computer interaction and pattern recognition; facial expression recognition and face detection.	user correction - Legacy Corrections
YEAR	2004	INFERENCE
VENUE	IEEE Transactions on Pattern Analysis and Machine Intelligence	INFERENCE
VENUE TYPE	JOURNAL	INFERENCE
PAGES	1553--1567	INFERENCE
VOLUME	26	INFERENCE
CITATIONS	63 found	ParsCit 1.0