Proceedings of the Acoustical Society of Korea Conference (한국음향학회:학술대회논문집)
- 1994.06a
- /
- Pages.997-1002
- /
- 1994
A GPD-BASED DISCRIMINATIVE TRAINING ALGORITHM FOR PREDICTIVE NEURAL NETWORK MODELS
-
Na, Kyung-Min
(Department of Electronics Engineering Seoul National University) ;
- Rheem, Jae-Yeol (Department of Electronics Engineering Seoul National University) ;
- Ann, Sou-Guil (Department of Electronics Engineering Seoul National University)
- Published : 1994.06.01
Abstract
Predictive neural network models are powerful speech recognition models based on a nonlinear pattern prediction. Those models can effectively normalize the temporal and spatial variability of speech signals. But those models suffer from poor discrimination between acoustically similar words. In this paper, we propose a discriminative training algorithm for predictive neural network models based on a generalized probabilistic descent (GPD) algorithm and minimum classification error formulation (MCEF). The Evaluation of our training algorithm on ten Korean digits shows its effectiveness by 40% reduction of recognition error.
Keywords