Modelling pronunciation variation : Some preliminary results
Mirjam Wester, Judith M. Kessens, Catia Cucchiarini, Helmer Strik · ERA · 1997
In this paper we describe a method for improving the performance of a continuous speech recognizer by modelling pronunciation variation. Although the results obtained with this method are in line with those reported by other authors, the magnitude of the improvements is very small. In looking for possible explanations for these results, we computed various sorts of statistics about the material. Since these data proved to be very useful in understanding the effects of our method, they are discussed in this paper. Moreover, on the basis of these statistics we discuss how the system can be improved in the future. 1. Introduction At the Department of Language and Speech of the University of Nijmegen we are working on a Spoken Dialogue System (SDS) that will be employed to automate part of a public transport information service. This system was adapted from a German prototype developed by Philips Research Labs, and was further improved by means of a bootstrapping method (Strik et. al., 1...