Genones: optimizing the degree of mixture tying in a large vocabulary hidden Markov model based speech recognizer
Authors:
- 17 December 2002
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. i (15206149) , I/537-I/540
Abstract
We propose a scheme that improves the robustness of continuous HMM systems that use mixture observation densities by sharing the same mixture components among different HMM states. The sets of HMM states that share the same mixture components are determined automatically using agglomerative clustering techniques. Experimental results on the Wall-Street Journal Corpus show that our new form of output distributions achieves a 25% reduction in error rate over typical tied-mixture systems.Keywords
This publication has 12 references indexed in Scilit:
- Speaker stress-resistant continuous speech recognition2003
- On the use of tied-mixture distributions1993
- Benchmark tests for the DARPA Spoken Language Program1993
- Context-Dependent Phonetic Hidden Markov Models for Speaker-Independent Continuous Speech Recognition1992
- Linear discriminant analysis for improved large vocabulary continuous speech recognition1992
- Subphonetic modeling with Markov states-Senone1992
- Bayesian learning of Gaussian mixture densities for hidden Markov models1991
- Tied mixture continuous parameter modeling for speech recognitionIEEE Transactions on Acoustics, Speech, and Signal Processing, 1990
- Acoustic modeling for large vocabulary speech recognitionComputer Speech & Language, 1990
- Speaker-independent phone recognition using hidden Markov modelsIEEE Transactions on Acoustics, Speech, and Signal Processing, 1989