### Abstract

This paper presents an online/sequential linear regression adaptation framework for hidden Markov model (HMM) based speech recognition. Our attempt is to sequentially improve speaker-independent (SI) speech recognizer to meet nonstationary environments via linear regression adaptation of SI HMM's. A quasi-Bayes linear regression (QBLR) algorithm is developed to execute online adaptation where the regression matrix is estimated using QB theory. In the estimation, we moderately specify the prior density of regression matrix as a matrix variate normal distribution and exactly derive the pooled posterior density belonging to the same distribution family. Accordingly, the optimal regression matrix can be easily calculated. Also, the reproducible prior/posterior density pair provides meaningful mechanism for sequential learning of prior statistics. At each sequential epoch, only the updated prior statistics and the current observed data are required for adaptation. In general, the proposed QBLR is universal and can be reduced to well-known maximum likelihood linear regression (MLLR) and maximum a posteriori linear regression (MAPLR). Experiments show that the QBLR is effective for speaker adaptation in car environments.

Original language | English |
---|---|

Pages (from-to) | 329-332 |

Number of pages | 4 |

Journal | ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings |

Volume | 1 |

Publication status | Published - 2001 Sep 26 |

Event | 2001 IEEE Interntional Conference on Acoustics, Speech, and Signal Processing - Salt Lake, UT, United States Duration: 2001 May 7 → 2001 May 11 |

### Fingerprint

### All Science Journal Classification (ASJC) codes

- Software
- Signal Processing
- Electrical and Electronic Engineering

### Cite this

}

*ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings*, vol. 1, pp. 329-332.

**Online speaker adaptation based on quasi-Bayes linear regression.** / Chien, J. T.; Huang, Chih-Hsien.

Research output: Contribution to journal › Conference article

TY - JOUR

T1 - Online speaker adaptation based on quasi-Bayes linear regression

AU - Chien, J. T.

AU - Huang, Chih-Hsien

PY - 2001/9/26

Y1 - 2001/9/26

N2 - This paper presents an online/sequential linear regression adaptation framework for hidden Markov model (HMM) based speech recognition. Our attempt is to sequentially improve speaker-independent (SI) speech recognizer to meet nonstationary environments via linear regression adaptation of SI HMM's. A quasi-Bayes linear regression (QBLR) algorithm is developed to execute online adaptation where the regression matrix is estimated using QB theory. In the estimation, we moderately specify the prior density of regression matrix as a matrix variate normal distribution and exactly derive the pooled posterior density belonging to the same distribution family. Accordingly, the optimal regression matrix can be easily calculated. Also, the reproducible prior/posterior density pair provides meaningful mechanism for sequential learning of prior statistics. At each sequential epoch, only the updated prior statistics and the current observed data are required for adaptation. In general, the proposed QBLR is universal and can be reduced to well-known maximum likelihood linear regression (MLLR) and maximum a posteriori linear regression (MAPLR). Experiments show that the QBLR is effective for speaker adaptation in car environments.

AB - This paper presents an online/sequential linear regression adaptation framework for hidden Markov model (HMM) based speech recognition. Our attempt is to sequentially improve speaker-independent (SI) speech recognizer to meet nonstationary environments via linear regression adaptation of SI HMM's. A quasi-Bayes linear regression (QBLR) algorithm is developed to execute online adaptation where the regression matrix is estimated using QB theory. In the estimation, we moderately specify the prior density of regression matrix as a matrix variate normal distribution and exactly derive the pooled posterior density belonging to the same distribution family. Accordingly, the optimal regression matrix can be easily calculated. Also, the reproducible prior/posterior density pair provides meaningful mechanism for sequential learning of prior statistics. At each sequential epoch, only the updated prior statistics and the current observed data are required for adaptation. In general, the proposed QBLR is universal and can be reduced to well-known maximum likelihood linear regression (MLLR) and maximum a posteriori linear regression (MAPLR). Experiments show that the QBLR is effective for speaker adaptation in car environments.

UR - http://www.scopus.com/inward/record.url?scp=0034841235&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0034841235&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:0034841235

VL - 1

SP - 329

EP - 332

JO - Proceedings - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing

JF - Proceedings - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing

SN - 0736-7791

ER -