Abstract | ||
---|---|---|
The BBN BYBLOS continuous speech recognition system has been used to develop a method of speaker adaptation from limited training. The key step in the method is the estimation of a probabilistic spectral mapping between a prototype speaker, for whom there exists a well-trained speaker-dependent hidden Markov model (HMM), and a target speaker for whom there is only a small amount of training speech available. The mapping defines a set of transformation matrices which are used to modify the parameters of the prototype model. The resulting transformed model is then used as an approximation to a well-trained model for the target speaker. We review the techniques employed to accomplish this transformation and present experimental results conducted on the DARPA Resource Management database. |
Year | DOI | Venue |
---|---|---|
1989 | 10.3115/100964.100969 | HLT |
Keywords | Field | DocType |
markov model,bbn byblos speech recognition,well-trained model,prototype model,prototype speaker,training speech,probabilistic spectral mapping,bbn byblos continuous speech,limited training,speaker adaptation,target speaker,speech recognition,hidden markov model,resource manager | Resource management,Existential quantification,Computer science,Speech recognition,Speaker recognition,Speaker diarisation,Natural language processing,Artificial intelligence,Probabilistic logic,Transformation matrix,Hidden Markov model,Speaker adaptation | Conference |
Citations | PageRank | References |
3 | 3.05 | 3 |
Authors | ||
4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Francis Kubala | 1 | 410 | 99.88 |
Ming-Whei Feng | 2 | 72 | 11.78 |
J. Makhoul | 3 | 1097 | 233.37 |
Richard M. Schwartz | 4 | 2839 | 717.76 |