Efficient search using posterior phone probability estimates
- 19 November 2002
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 1 (15206149) , 596-599
- https://doi.org/10.1109/icassp.1995.479668
Abstract
We present a novel, efficient search strategy for large vocabulary continuous speech recognition (LVCSR). The search algorithm, based on stack decoding, uses posterior phone probability estimates to substantially increase its efficiency with minimal effect on accuracy. In particular, the search space is dramatically reduced by phone deactivation pruning where phones with a small local posterior probability are deactivated. This approach is particularly well-suited to hybrid connectionist/hidden Markov model systems because posterior phone probabilities are directly computed by the acoustic model. On large vocabulary tasks, using a trigram language model, this increased the search speed by an order of magnitude, with 2% or less relative search error. Results from a hybrid system are presented using the Wall Street Journal LVCSR database for a 20,000 word task using a backed-off trigram language model. For this task, our single-pass decoder took around 15x realtime on an HP735 workstation. At a cost of 7% relative search error, the decoding time can be speeded up to approximately realtime.Keywords
This publication has 12 references indexed in Scilit:
- Optimizing recognition and rejection performance in wordspotting systemsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- A channel-bank-based phone detection strategyPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Improvements in beam search for 10000-word continuous-speech recognitionIEEE Transactions on Speech and Audio Processing, 1994
- An application of recurrent nets to phone probability estimationIEEE Transactions on Neural Networks, 1994
- Connectionist probability estimators in HMM speech recognitionIEEE Transactions on Speech and Audio Processing, 1994
- Connectionist Speech RecognitionPublished by Springer Nature ,1994
- A*-admissible heuristics for rapid lexical accessIEEE Transactions on Speech and Audio Processing, 1993
- An efficient A* stack decoder algorithm for continuous speech recognition with a stochastic language modelPublished by Institute of Electrical and Electronics Engineers (IEEE) ,1992
- Decoding for channels with insertions, deletions, and substitutions with applications to speech recognitionIEEE Transactions on Information Theory, 1975
- Fast Sequential Decoding Algorithm Using a StackIBM Journal of Research and Development, 1969