Improving speech recognition performance via phone-dependent VQ codebooks and adaptive language models in SPHINX-II

Mei-Yuh Hwang, Ronald Rosenfeld, Eric H. Thayer, Mosur Ravishankar, Lin Lawrence Chase, Robert Weide, Xuedong Huang, Fil Alleva

Abstract

This paper presents improvements in acoustic and language modeling for automatic speech recognition. Specifically, semi-continuous HMMs (SCHMMs) with phone-dependent VQ codebooks are presented and incorporated into the SPHINX-II speech recognition system. The phone-dependent VQ codebooks relax the density-tying constraint in SCHMMs in order to obtain more detailed models. A 6% error rate reduction was achieved on the speaker-independent 20000-word Wall Street Journal (WSJ) task. Dynamic adaptation of the language model in the context of long documents is also explored. A maximum entropy framework is used to exploit long distance trigrams and trigger effects. A 10%-15% word error rate reduction is reported on the same WSJ task using the adaptive language modeling technique.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>
BibTeX
@inproceedings{hwang1994improving,
  author = {Mei-Yuh Hwang and Ronald Rosenfeld and Eric H. Thayer and Mosur Ravishankar and Lin Lawrence Chase and Robert Weide and Xuedong Huang and Fil Alleva},
  title = {{Improving speech recognition performance via phone-dependent VQ codebooks and adaptive language models in SPHINX-II}},
  booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing},
  year = {1994},
  doi = {10.1109/icassp.1994.389235},
}

Loading related content...

Figure