The Polysemy Problem, an Important Issue in a Chinese to Taiwanese TTS System

Computational Linguistics and Chinese Language Processing Vol. 17, No. 1, March 2012, pp. 43-64 43 © The Association for Computational Linguistics and Chinese Language Processing The Polysemy Problem, an Important Issue in a Chinese to Taiwanese TTS System Ming-Shing Yu∗ and Yih-Jeng Lin+ Abstract This paper brings up an important issue, 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 “我們” (we), “不” (no), “你” (you), “我” (I), and “要” (want). We first will show the importance of the polysemy problem in a Chinese to Taiwanese (C2T) TTS system. Then, we will propose some approaches to a difficult case of such problems by determining the pronunciation of “我們” (we) in a C2T TTS system. There are two pronunciations of the word “我們” (we) in Taiwanese, /ghun/ and /lan/. The corresponding Chinese words are “阮” (we1) and “咱” (we2). We propose two approaches and a combination of the two to solve the problem. The results show that we have a 93.1% precision in finding the correct pronunciation of the word “我們” (we). Compared to the results of the layered approach, which has been shown to work well in solving other polysemy problems, the results of the combined approach are an improvement. Keywords: Polysemy, Taiwanese, Chinese to Taiwanese TTS System, Layered Approach 1. Introduction Besides Mandarin, Taiwanese is the most widely spoken dialect in Taiwan. According to Liang et al. (2004), about 75% of the population in Taiwan speaks Taiwanese. Currently, it is government policy to encourage people to learn one’s mother tongue in schools because local languages are a part of local culture. Researchers (Bao et al., 2002; Chen et al., 1996; Lin et al., 1998; Lu, 2002; Shih et al., 1996; Wu et al., 2007; Yu et al., 2005) have had outstanding results in developing Mandarin ∗ Department of Computer Science and Engineering, National Chung-Hsing University, Taichung 40227, Taiwan. + Department of Information Management, Chien-Kuo Technology University, Chang-hua 500, Taiwan. E-mail: yclin@ctu.edu.tw 44 Ming-Shing Yu and Yih-Jeng Lin text-to-speech (TTS) systems over the past ten years. Other researchers (Ho, 2000; Huang, 2001; Hwang, 1996; Lin et al., 1999; Pan & Yu, 2008; Pan, Yu, & Tsai, 2008; Yang, 1999; Zhong, 1999) have just begun to develop Taiwanese TTS systems. There are no formal characters for Taiwanese, so Chinese characters are officially used in Taiwan. Consequently, many researchers have focused on Chinese to Taiwanese (C2T) TTS systems. This means that the input of a so-called Taiwanese TTS system is Chinese text. Yang (1999) developed a method based on machine translation to help solve this problem. Since there are differences between Mandarin and Taiwanese, a C2T TTS system should have a text analysis module that can solve problems specific to Taiwanese. For instance, there is only one pronunciation for “我們” (we) in Chinese, but there are two pronunciations for “我們” (we) in Taiwanese. Figure 1 shows a common structure of a C2T TTS system. In general, a …

Content truncated to 500 words