|
My name is Hao Tang and today, I would like to introduce myself. Through this month's tutorial, I'm going to introduce you to my previous work, my current research interests and my vision for ISIP. I am currently a graduate student studying toward a Ph.D. degree in the Department of Electrical and Computer Engineering at Mississippi State University. I received my B.S. and M.S. degrees, both in Electrical Engineering, from the University of Science and Technology of China, in 1998 and 2003, respectively. During my graduate study in China, I held a senior research position at iFlyTEK, a leading Chinese speech technology provider. This position provided me with plenty of exposure to the research. I have led and participated in over fifteen projects at iFlyTEK and my previous work made me win three scientific and technical awards from the governments, of which one is the National Scientific and Technical Progress Award conferred by the State Council of P. R. China. In addition to these accomplishments, I have three pending Chinese patents. My past research fell mainly under the area of speech synthesis. I have been a member of the team that developed a number of Chinese Speech Synthesis Systems. You can evaluate the performance of these speech synthesis systems, by clicking on the links given below. These links point to the three synthetic wave files, of which the first one was generated by our KBCT2.0 Cantonese TTS engine, the next one was generated by our KBCE2.0 Chinese-English Mixed-lingual TTS engine, and the last one was generated by our KD2000 Chinese TTS engine. (Note: KBCT2.0, KBCE2.0 and KD2000 are trademarks of iFlyTEK.) KBCT2.0 (Prototype) Apr. 2003 Wave file 1 KBCE2.0 (Product) Sep. 2002 Wave file 2 KD2000 (Product) Jun. 2000 Wave file 3Apart from speech synthesis, I have also worked in the area of distributed and embedded speech computing. I proposed an innovative concept Distributed Speech Synthesis aimed at developing high naturalness speech synthesis systems on resource-sensitive mobile and embedded devices. I also developed several prototype systems under the proposed architecture. Why do I want to pursue a Ph.D. and why do I want to make a transition from speech synthesis to speech recognition ? The answer to both these questions is simple. My career goal is to become a faculty and direct a research lab. In order to become a professor and conduct in-depth research, a Ph.D. degree is a must. I believe that speech recognition is one of the most fundamental and challenging field and thus deserves much more research efforts than speech synthesis. While working in ISIP, I will have ample of opportunities to get involved in speech recognition research. My experience in industry has helped me develop my own style of learning. That is, to learn and master knowledge from practical work. ISIP provides an environment that emphasizes practical work and that reflects my style of learning. My first project in ISIP is to develop a speaker verification system. Speaker verification is indeed a perfect start for me because the underlying theory of speaker verification is conceptually similar to speech recognition. I am sure the experience and knowledge I gain from this project will definitely benefit my future work on the more complicated and challenging speech recognition problems. |
