Abstract
In this paper, we propose a new incremental model adaptation approach based on posterior distributions of model parameters. We consider a propagation mechanism of the posterior distributions whereby that the process of posterior refinement is modeled analytically. Then, we derive an incremental estimation algorithm based on a time evolution system, which explicitly includes a discrete stochastic process unlike the conventional Bayesian approaches. This algorithm is viewed as a general solution of the Kalman filter algorithm, where posterior distributions make a transition after every input of an utterance set, and where the evolutions of posterior distributions are represented on a macroscopic time scale.
| Original language | English |
|---|---|
| Title of host publication | 2007 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '07 |
| Volume | 4 |
| DOIs | |
| Publication status | Published - 2007 |
| Externally published | Yes |
| Event | 2007 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '07 - Honolulu, HI Duration: 2007 Apr 15 → 2007 Apr 20 |
Other
| Other | 2007 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '07 |
|---|---|
| City | Honolulu, HI |
| Period | 07/4/15 → 07/4/20 |
Fingerprint
Keywords
- Acoustic model
- Discrete stochastic process
- Incremental adaptation
- Macroscopic time evolution
- Speech recognition
ASJC Scopus subject areas
- Electrical and Electronic Engineering
- Signal Processing
- Acoustics and Ultrasonics
Cite this
Incremental adaptation based on a macroscopic time evolution system. / Watanabe, Shinji; Nakamura, Atsushi.
2007 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '07. Vol. 4 2007. 4218214.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Incremental adaptation based on a macroscopic time evolution system
AU - Watanabe, Shinji
AU - Nakamura, Atsushi
PY - 2007
Y1 - 2007
N2 - In this paper, we propose a new incremental model adaptation approach based on posterior distributions of model parameters. We consider a propagation mechanism of the posterior distributions whereby that the process of posterior refinement is modeled analytically. Then, we derive an incremental estimation algorithm based on a time evolution system, which explicitly includes a discrete stochastic process unlike the conventional Bayesian approaches. This algorithm is viewed as a general solution of the Kalman filter algorithm, where posterior distributions make a transition after every input of an utterance set, and where the evolutions of posterior distributions are represented on a macroscopic time scale.
AB - In this paper, we propose a new incremental model adaptation approach based on posterior distributions of model parameters. We consider a propagation mechanism of the posterior distributions whereby that the process of posterior refinement is modeled analytically. Then, we derive an incremental estimation algorithm based on a time evolution system, which explicitly includes a discrete stochastic process unlike the conventional Bayesian approaches. This algorithm is viewed as a general solution of the Kalman filter algorithm, where posterior distributions make a transition after every input of an utterance set, and where the evolutions of posterior distributions are represented on a macroscopic time scale.
KW - Acoustic model
KW - Discrete stochastic process
KW - Incremental adaptation
KW - Macroscopic time evolution
KW - Speech recognition
UR - http://www.scopus.com/inward/record.url?scp=34547552788&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=34547552788&partnerID=8YFLogxK
U2 - 10.1109/ICASSP.2007.367026
DO - 10.1109/ICASSP.2007.367026
M3 - Conference contribution
AN - SCOPUS:34547552788
SN - 1424407281
SN - 9781424407286
VL - 4
BT - 2007 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '07
ER -