Automatic Speech Recognition: The Development of the Sphinx Recognition System
Automatic Speech Recognition: The Development of the Sphinx Recognition System
DARPA resource management benchmark test results June 1990
HLT '90 Proceedings of the workshop on Speech and Natural Language
| Hi-index | 0.00 |
This paper describes the tree-structured maximum mutual information (MMI) encoders used in SSI's Phonetic Engine® to perform large-vocabulary, continuous speech recognition. The MMI encoders are arranged into a two-stage cascade. At each stage, the encoder is trained to maximize the mutual information between a set of phonetic targets and corresponding codes. After each stage, the codes are compressed into segments. This step expands acoustic-phonetic context and reduces subsequent computation. We evaluated these MMI encoders by comparing them against a standard minimum distortion (MD) vector quantizer (encoder). Both encoders produced code streams, which were used to train speaker-independent discrete hidden Markov models in a simplified version of the Sphinx system [3]. We used data from the DARPA Resource Management (RM) task. The two-stage cascade of MMI encoders significantly outperforms the standard MD encoder in both speed and accuracy.