๐”– Bobbio Scriptorium
โœฆ   LIBER   โœฆ

Simulated annealing for maximum a posteriori parameter estimation of hidden Markov models

โœ Scribed by Andrieu, C.; Doucet, A.


Book ID
114541576
Publisher
IEEE
Year
2000
Tongue
English
Weight
307 KB
Volume
46
Category
Article
ISSN
0018-9448

No coin nor oath required. For personal study only.


๐Ÿ“œ SIMILAR VOLUMES


The semi-relaxed algorithm for estimatin
โœ Li Deng ๐Ÿ“‚ Article ๐Ÿ“… 1991 ๐Ÿ› Elsevier Science ๐ŸŒ English โš– 489 KB

For hidden Markov model (HMM) based speech recognition where the basic speech unit is smaller than the recognizer's output unit, the standard full Baum-Welch re-estimation procedure for the HMM training is very costly in computation. This is hecause it requires evaluation of the HMM output densities