Probabilistic Modeling with Bayesian Networks for Automatic Speech Recognition
Download:
|
|
by Geoffrey Zweigy, Stuart Russellz
http://www.cs.berkeley.edu/~russell/papers/ajiips-speech.ps
Add To MetaCart
Abstract:
Bayesian networks are an extremely general probabilistic modeling framework, and are increasingly being applied to complex real-world problems. In this paper, we describe the use of a Bayesian network system in large vocabulary isolated word recognition. We briefly review the algorithms and network structures used, and present results showing that significant improvements in word error rate result from modeling acoustic and articulatory context with a multivalued context variable. The network parameters are highly correlated with simply defined acoustic characteristics, and utterance clustering results in a partitioning according to both speaker gender and the presence of liquid consonants. 1
Citations
| 88 | Probabilistic temporal reasoning – Dean, Kanazawa - 1988 |
