Abstract

A method of organizing an acoustic model for speech recognition is comprised of the steps of calculating a measure of acoustic dissimilarity of subphonetic units. A clustering technique is recursively applied to the subphonetic units based on the calculated measure of acoustic dissimilarity to automatically generate a hierarchically arranged model. Each application of the clustering technique produces another level of the hierarchy with the levels progressing from the least specific to the most specific. A technique for adapting the structure and size of a trained acoustic model to an unseen domain using only a small amount of adaptation data is also disclosed.

Other References

• Jurgen Fritsch, Michael Finke, "Acid/HNN: Clustering Hierarchies of Neural Networks for Context--Dependent Connectionist Acoustic Modeling," IEEE International conference on Acoustics, Speech and Signal Processing, Conference 23 (New York, New York), p. 505-508, (1998)
• J. Fritsch, M. Finke, A. Waibel, "Effective Structural Adaptation of LVCSR Systems to Unsen domains using hierarchical connectionist acoustic models," Proceedings of the International Conference on Spoken Language Processing, p. 2919-2922, (Nov. 30-Dec. 4, 1998)
• Paul, D.B., "Extensions to Phone-State Decision-Tree Clustering: Single Tree and Tagged Clustering," IEEE Comp. Soc. Press, IEEE International Conference on Acoustic, Speech, and Signal Processing (Los Alamitos, US), p. 1487-1490, ( 1997)
• H. Franco, "Context-Dependent Connectionist Probability Estimation in a Hybrid Markov Model-Neural Net Speech Recognition System," Computer Speech and Language, vol. 8 (No. 3), (Feb. 22, 1994)
• J Fritsch, et al., "Context-Dependent Hybrid HME/HMM Speech Recognition Using Polyphone Clustering Decision Trees," Proc. of ICASSP '97
• D.J. Kershaw, et al., "Context-Dependent Classes in a Hybrid Recurrent Network HMM Speech Recognition System," Tech. Rep. CUED/F-INFENG/TR217, CUED, Cambridge England 1995
• D.L. Thomson, "Ten Case Studies of the Effect of Field Conditions on Speech Recognition Errors," Proceedings of the IEEE ASRU Workshop, (Feb. 22, 1997)
• J. Schurmann and W. Doster, "A Decision Theoretic Approach to Hierarchical Classifier Design," Pattern Recognition 17(3), (Feb. 22, 1994)
• J. Fritsch, "Acid/HNN; A Framework for Hierarchical Connectionist Acoustic Modeling," Proceedsing of IEEE ASRU Workshop, (Feb. 22, 1997)
• C.J. Leggetter and P.C. Woodland, "Speaker Adaptation of HMMs using Linear Regression," Tech. Rep. CUED/F-INFENG/TR181, CUED, (Feb. 22, 1994)
• Franco, H., "Context-Dependent Connectionist Probability Estimation in a Hybrid Markov Model-Neural Net Speech Recognition System", Computer Speech and Language, vol. 8, No. 3, Jul. 1994
• Fritsch, J., et al, "Context-Dependent Hybrid MHE/HMM Speech Recognition Using Polyphone Clustering Decision Trees", Proc. Of ICASS '97, Apr. 21-24, 1997
• Kershaw, D. J., et al, "Contest-Dependent Classes in a Hybrid Recurrent Network HMM Speech Recognition System", Tech. Rep CUED/F-INFENG/TR217, CUED, Cambridge, England, Jul. 1995
• Thomson, D. L., "Ten Case Studies of the Effect of Field Conditions on Speech Recognition Errors", Proceedings of the IEEE ASRU Workshop, Dec. 17, 1997
• Schurmann, J., et al. "A Decision Theoretic Approach to Hierarchical Classifier Design", Pattern Recognition, 17 (3), 1984
• Fritsch., J., "ACIDHNN; A Framcwork for Hierarchical Connectionist Acoustic Modeling", Proceedings of IEEE ASRU Workshop, Dec. 14-17, 1997
• Leggetter, C.J., et al, "Speaker Adaptation of HMM's Using Linear Regression", Tech. Rep. CUED/F-INFENG/TR181, CUED, Jun. 199