Language adaptive LVCSR through polyphone decision tree specialization
Tanja Schultz, Alex Waibel · 1999
With the distribution of speech technology products all over the world, the fast and efficient portability to new target languages becomes a practical concern. In this paper we explore the relative effectiveness of porting multilingual recognition systems to new target languages with very limited adaptation data. For this purpose we introduce a polyphone decision tree specialization method. Several recognition results are presented based on mono- and multilingual recognizers developed in the framework of the project GlobalPhone which investigates LVCSR systems in 15 languages. 1. Introduction With the distribution of speech technology products all over the world, the fast and efficient portability to new target languages becomes a practical concern. So far one major time and costs limitation in developingLVCSR systems in new languages is the need of large training data. According to the amount of data used for porting acoustic models to a new target language we differentiate three as...