Speaker adaptation based on lateral tying for large-vocabulary c

Details Speaker adaptation based on lateral tying for large-vocabulary c Speaker adaptation based on lateral tying for large-vocabulary c

1996-02-13

1998-04-07

Hafiz, Tariq R.

395 24, 395 255, 395 258, G10L 506

Patent

active

057374872

ABSTRACT:
A system and method for performing speaker adaptation in a speech recognition system which includes a set of reference models corresponding to speech data from a plurality of speakers. The speech data is represented by a plurality of acoustic models and corresponding sub-events, and each sub-event includes one or more observations of speech data. A degree of lateral tying is computed between each pair of sub-events, wherein the degree of tying indicates the degree to which a first observation in a first sub-event contributes to the remaining sub-events. When adaptation data from a new speaker becomes available, a new observation from adaptation data is assigned to one of the sub-events. Each of the sub-events is then populated with the observations contained in the assigned sub-event based on the degree of lateral tying that was computed between each pair of sub-events. The reference models corresponding to the populated sub-events are then adapted to account for speech pattern idiosyncrasies of the new speaker, thereby reducing the error rate of the speech recognition system.

REFERENCES:
patent: 5293452 (1994-03-01), Picone et al.
patent: 5377301 (1994-12-01), Rosenberg et al.
patent: 5469529 (1995-11-01), Bimbot et al.
patent: 5579436 (1996-11-01), Chou et al.
Proc: ICASSP. "Speaker Adaptation Through Vector Quantization", K. Shikano, K. Lee, R. Reddy, Tokyo, Japan, Apr. 1986, pp. 2643-2646.
Proc: ICASSP. "Speaker Adaptation for Large Vocabulary Speech Recognition Systems Using Speaker Markov Models." Rigoll, G., Glasgow, U.K., May 1989, pp. 5-8.
Proc. ICASP, "Speaker Adaptation From a speaker-Independent Training Corpus," Kubala F., Schwartz R., Barry C., Albuquerque, NM, Apr. 1990, pp. 137-140.
IEEE Trans. Speech Audio Proc., "Multonic Markov Word Models for Large Vocabulary Continuous Speech Recognition," Bahl L.R., et al., Jul. 1993, vol. SAP-1, No. 3, pp. 334-344.
IEEE Trans. Speech Audio Proc., "An Acoustic-Phonetic-Based Speaker Adaption tTechnique for Improving Speaker-Independent Continuous Speech Recognition, Zhao Y., Jul. 1994, vol. SAP-2, No. 3, pp. 380-394.
IEEE Trans Speech Audio Proc., "The Metamorphic Algorithm: Speaker Mapping Approach to Data Augmentatin," Bellegarda J.R. et al, Jul. 1994, vol. SAP-2, No. 3, pp. 413-420.
EUSIPCO, PROC., "Performance of the IBM Large Vocabulary Continuous Speech Recognition System on ARPA Wall Street Journal Task," Edinburgh, Scotland, Sep. 1994.
ICASSP, Proc 1995, "Experiments Using Data Augmentation for Speaker Adaptation," Bellgarda et al., Detroit, MI, May 1995.

Affiliated with

Also associated with

Rating

Speaker adaptation based on lateral tying for large-vocabulary c does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Speaker adaptation based on lateral tying for large-vocabulary c, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Speaker adaptation based on lateral tying for large-vocabulary c will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-21710