A Layered Approach to the Polysemy Problems in a Chinese to Taiwanese TTS Systems

This paper proposes a novel approach to the polysemy problems in a Chinese to Taiwanese TTS (text-to-speech) system. Polysemy means there are words with more than one meaning or pronunciation, such as “不”(no), “上”(up), and “下”(down), and so on. The correct pronunciation of a word affects the comprehensibility (or intelligibility) and fluency of Taiwanese speech. We applied our approach to solve the polysemy problems of the above mentioned three words. These words are difficult polysemy cases since they have many (6, 3, and 4, respectively) possible pronunciations. Results show that the proposed layered approach is over some language models and decision list classifier. Our approach has high accuracy and it can be extended when more training data is available.