Bayesian inference for finite-state transducers

Bayesian inference for finite-state transducers (2010)

Cached

Download Links

by David Chiang , Jonathan Graehl , Kevin Knight , Adam Pauls , Sujith Ravi

Venue:	in HLT-NAACL
Citations:	7 - 4 self

Versions

Version History

Metadata Version 1

Datum	Value	Source
TITLE	Bayesian inference for finite-state transducers	INFERENCE
AUTHOR NAME	David Chiang	SVM HeaderParse 0.2
AUTHOR AFFIL	1; Information Sciences Institute; University of Southern California	SVM HeaderParse 0.2
AUTHOR ADDR	4676 Admiralty Way, Suite 1001; Marina del Rey, CA 90292	SVM HeaderParse 0.2
AUTHOR NAME	Jonathan Graehl	SVM HeaderParse 0.2
AUTHOR AFFIL	1; Information Sciences Institute; University of Southern California	SVM HeaderParse 0.2
AUTHOR ADDR	4676 Admiralty Way, Suite 1001; Marina del Rey, CA 90292	SVM HeaderParse 0.2
AUTHOR NAME	Kevin Knight	SVM HeaderParse 0.2
AUTHOR AFFIL	1; Information Sciences Institute; University of Southern California	SVM HeaderParse 0.2
AUTHOR ADDR	4676 Admiralty Way, Suite 1001; Marina del Rey, CA 90292	SVM HeaderParse 0.2
AUTHOR NAME	Adam Pauls	SVM HeaderParse 0.2
AUTHOR AFFIL	1; Information Sciences Institute; University of Southern California	SVM HeaderParse 0.2
AUTHOR ADDR	4676 Admiralty Way, Suite 1001; Marina del Rey, CA 90292	SVM HeaderParse 0.2
AUTHOR NAME	Sujith Ravi	SVM HeaderParse 0.2
AUTHOR AFFIL	1; Information Sciences Institute; University of Southern California	SVM HeaderParse 0.2
AUTHOR ADDR	4676 Admiralty Way, Suite 1001; Marina del Rey, CA 90292	SVM HeaderParse 0.2
ABSTRACT	We describe a Bayesian inference algorithm that can be used to train any cascade of weighted finite-state transducers on end-toend data. We also investigate the problem of automatically selecting from among multiple training runs. Our experiments on four different tasks demonstrate the genericity of this framework, and, where applicable, large improvements in performance over EM. We also show, for unsupervised part-of-speech tagging, that automatic run selection gives a large improvement over previous Bayesian approaches. 1	SVM HeaderParse 0.2
YEAR	2010	INFERENCE
VENUE	in HLT-NAACL	INFERENCE
VENUE TYPE	CONFERENCE	INFERENCE
CITATIONS	25 found	ParsCit 1.0