Name
Title | ||
|---|---|---|
Different strategies for distribution clustering using discrete, semicontinuous and continuous HMMs in CSR |
Abstract | ||
|---|---|---|
The authors present an overview of different strategies and refinements to share parameters in HMM models at distribution (state) level for continuous speech recognition, showing the advantages and drawbacks of the different kinds of modeling. They compare them with sharing at the model level, achieving an error reduction close to 20%. Discrete, semicontinuous and continuous HMM models are also compared using these approaches. They consider two ways to smooth discrete distributions (interpolate detailed context dependent with robust context independent) derived from deleted interpolation and co-occurrence smoothing |
Year | DOI | Venue |
|---|---|---|
1996 | 10.1109/ICSLP.1996.607798 | Spoken Language, 1996. ICSLP 96. Proceedings., Fourth International Conference |
Keywords | Field | DocType |
hidden Markov models,interpolation,smoothing methods,speech recognition,co-occurrence smoothing,continuous HMM,continuous speech recognition,deleted interpolation,discrete HMM,discrete distribution smoothing,distribution clustering,error reduction,modeling,parameter sharing,semicontinuous HMM | Pattern recognition,Interpolation,Robustness (computer science),Context model,Smoothing,Artificial intelligence,Context independent,Loudspeaker,Cluster analysis,Hidden Markov model,Mathematics | Conference |
Volume | ISBN | Citations |
2 | 0-7803-3555-4 | 2 |
PageRank | References | Authors |
0.60 | 5 | 3 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Ricardo De Córdoba | 1 | 142 | 25.58 |
| José Manuel Pardo | 2 | 152 | 30.36 |
| de Cordoba, R. | 3 | 2 | 0.60 |