The AT&T WATSON speech recognizer (2005)
| Venue: | in Proceedings of ICASSP |
| Citations: | 10 - 1 self |
BibTeX
@INPROCEEDINGS{Goffin05theat&t,
author = {V. Goffin and C. Allauzen and E. Bocchieri and D. Hakkani-tür and A. Ljolje and S. Parthasarathy and M. Rahim and G. Riccardi and M. Saraclar},
title = {The AT&T WATSON speech recognizer},
booktitle = {in Proceedings of ICASSP},
year = {2005}
}
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
This paper describes the AT&T WATSON real-time speech recognizer, the product of several decades of research at AT&T. The recognizer handles a wide range of vocabulary sizes and is based on continuous-density hidden Markov models for acoustic modeling and finite state networks for language modeling. The recognition network is optimized for efficient search. We identify the algorithms used for high-accuracy, real-time and low-latency recognition. We present results for small and large vocabulary tasks taken from the AT&T VoiceTone R ○ service, showing word accuracy improvement of about 5 % absolute and real-time processing speed-up by a factor between 2 and 3. 1.
Citations
| 67 | Heteroscedastic discriminant analysis and reduced rank HMMs for improved speech recognition - Kumar, Andreou - 1998 |
| 2 | Multiple task-domain acoustic models - Ljolje - 2003 |
| 1 | The at&t rt-02-stt gt10xrt system - Ljolje, Saraclar, et al. - 2002 |
