What makes language the key distinguishing ability of the human species? Why can't computers understand sentences as well as three-year children? How do human languages relate to computer languages? How did you learn English or Chinese? How does language enable intelligence?

A few years ago I introduced some deceptively simple hypotheses concerning questions like these, initiating a long-term scientific debate. Now, during the past year, intriguing new large-scale results supporting these "ITG hypotheses" have been published by several international research groups.

The ITG hypotheses established today's dominant paradigm in machine translation, namely statistical tree-based models. But the ITG hypotheses impact a wider range of fundamental AI issues like these: What machine learning models are really important? Which of the many models are actually too expensive and unnecessary? How can we simplify learning models to be more elegant, effective, and efficient? How are machine learning models affected by language representation, and how can they take advantage of it? Such questions lie at the core of many of the most important growth areas of computer technology this decade, including Intelligent Search (Google), the Semantic Web, Knowledge and Document Management, Unstructured Data Mining, Spoken Language Understanding, and many other diverse areas.

Biography

Prof. Wu received his PhD in Computer Science from the University of California at Berkeley, and was a postdoctoral fellow at the University of Toronto (Ontario, Canada) prior to joining HKUST in 1992. He received his Executive MBA from Kellogg and HKUST in 2002, and a BS in Computer Engineering from the University of California at San Diego (Revelle College departmental award, cum laude, Phi Beta Kappa) in 1984. He has been a visiting researcher at Columbia University in 1995-96, Bell Laboratories in 1995, and the Technische Universität München (Munich, Germany) during 1986-87. Prof. Wu serves as Associate Editor of ACM Transactions on Speech and Language Processing, Machine Translation, Journal of Natural Language Engineering, and Communications of COLIPS, and is Co-Chair for EMNLP-2004. He has also served on the Editorial Board of Computational Linguistics, the Organizing Committee of ACL-2000 and WVLC-5 (SIGDAT 1997), and the Executive Committee of the Association for Computational Linguistics (ACL).