Comments on "Efficient training algorithms for HMM's using incremental estimation"

Authors
Byrne, W Gunawardana, A
Citation
W. Byrne et A. Gunawardana, Comments on "Efficient training algorithms for HMM's using incremental estimation", IEEE SPEECH, 8(6), 2000, pp. 751-754
Citations number
13
Categorie Soggetti
Eletrical & Eletronics Engineeing
Journal title
IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING
ISSN journal
10636676 → ACNP
Volume
8
Issue
6
Year of publication
2000
Pages
751 - 754
Database
ISI
SICI code
1063-6676(200011)8:6<751:CO"TAF>2.0.ZU;2-H
Abstract
"Efficient training algorithms for HMM's using incremental estimation" [1] investigates EM procedures that increase training speed. The authors' claim that these are GEM [2] procedures Ts incorrect. We discuss why this is so, provide an example of nonmonotonic convergence to a local maximum in Likel ihood, and outline conditions that guarantee such convergence.