Malay-English Code Switching Speech Recognition

Malaysia is a multilingual society where most people are capable of speaking more than one language. Code switching is a common phenomenon in Malaysian conversation where more than one language is used at the same time. In Malay conversation for example, speakers often switch between Malay and English. In a code switching conversation, a speaker can switch between two or more languages. One or more words or sentences may be switched.

In this research, we study the framework, acoustic modeling, pronunciation modeling and language modeling in automatic speech recognition system

Malay Dialects Speech Synthesis

Malay is one of the most widely spoken languages in the world. Malay is the official language in Malaysia, Indonesia, Singapore, and Brunei. However, Malay spoken in different countries and even within a country itself might vary in terms of pronunciation, vocabulary and may be grammar from one place to another. Therefore, Malay dialects presents many interesting problems in linguistic. In this research, we look at speech synthesis for Malay dialect, especially for the rapid development and deployment of speech synthesis.

Spoken Document Retrieval

Current spoken document retrieval is limited because it relies on the the manual tagging of content by users. Next generation spoken document retrieval system applied automatic speech recognition to allow more accurate retrieval of information in spoken document and multimedia documents.

Malay-English Machine Translation

In this research, we study statistical machine translation in Malaysia context.