BIRTH OF A CROSS-FIELD RESEARCH - NATURAL CHINESE INPUT
Before Dr. Hsu joined Northwestern University on graph theory research, he majored in mathematics and operation research. In 1989, he returned to Academia Sinica to start the research with Dr. Ke-Jien Chen on word-to-syllable Chinese input method. This method, as groundbreaking as it was, was actually riduculed by other Chinese scholars who were more interested in how to come up with a perfect Chinese character decomposition method instead. This prompted Dr. Hsu to look into the implementation of AI technique to optimize the word-to-syllable method. After two years of work, the intelligent phonetic Chinese input came into shape and, with the help from a few third-party companies on designing the interface, the first national language input method was born for PC in 1992. In 1994, Natural Input method came into light.
KNOWLEDGE MANAGEMENT AND SOFTWARE AGENT
In the summer of 1996, Dr. Hsu was on research visiting at the CSLI in Stanford University. After attending six to seven different seminars each week, he started to realize the importance of knowledge management in the role of problem solving, as well as the trend of incorporating effective KM with intelligent agent software. Therefore, upon return to Taiwan in 1997, Dr. Hsu renamed the lab as "Intelligent Agent System Lab" and initiated the development of InfoMap system in 1998. Infomap specializes in conglomerating common knowledge with linguistic and domain-specific knowledge so to make proper inference with such information to facilitate the conversion between words and syllables. The system provided a platform for computer scientists and knowledge managers to work complementarily and simultaneously.
NATURAL LANGUAGE PROCESSING V.S. BIOINFORMATICS
At a random occasion in 1994, Dr. Hsu had a discussion with Dr. Lan-yang Chang about representing DNA and protein sequences as a language pattern for bio analysis. Since then Dr. Hsu started exposure to simple genomics knowledge. Due to Dr. Hsu's effort and the proximity of such concept to graph theory, he completed the first publication on processing bio-sequences under the influence of experimental noise in 1998. Unfortunately at the time, most people were more attracted to the internet boom than devoting in bioinformatics; it was only until 2001 when Dr. Hsu systematically hosted summer bioinformatics workshop did students begin to pay attention to such work and join in related research of algorithms and natural language processing technique on the usage in bioinformatics.