Institute of Information Science Academia Sinica
Topic: Targeted Collective Natural Language Processing
Speaker: Prof. Richard Tzong-Han Tsai (Yuan Ze University)
Date: 2012-03-30 (Fri) 10:30 – 12:00
Location: Auditorium 106 at new IIS Building
Host: Ming-Tat Ko

Abstract:

In recent years, a collective approach in which related objects are associated and predicted simultaneously has become widely and successfully adopted in several data mining tasks—e.g. social network analysis and NLP. However, when analyzing a sentence, most collective NLP models set a short fixed window to limit computational complexity and only associate objects in this range, such as two consecutive tokens. This fixed window is counter to our intuition of how humans understand natural language sentences. In this talk, I will describe a novel approach, targeted collective NLP, which can greatly extend this range to the whole document by using knowledge resources and patterns to limit the analysis to highly correlated tokens. Preliminary experiments and comparisons have been conducted on three biomedical NLP tasks. Future plans will also be discussed in this talk.


BIO:

Dr. Richard Tzong-Han Tsai received his bachelor’s degree and master’s degree in Computer Science from National Taiwan University in 1997 and 1999, respectively. He received his PhD degree in Computer Science from National Taiwan University in 2006. He joined the faculty in Yuan Ze University since August, 2007. He is now an associate professor. His main research interests include natural language processing (especially in biomedical domains), cross-lingual information retrieval, text-based sentiment analysis, digital humanities, and context-based recommendation, etc.

Dr. Tsai achieved great success in several international academic completions. In 2006, his team won the first place of the 3rd Chinese word segmentation bakeoff held by SIGHAN. In 2007, his team won the first place of the NTCIR Korean-Chinese cross-language information retrieval shared task held by NII, Japan. In 2009, his team won the first place of the INT subtask in the BioCreative ii.5 workshop held by CNIO, Spain. In 2011, his team won the English-Korean named entity transliteration shared task in the NEWS workshop held by Institute for Infocomm Research, Singapore. 

Because of Dr. Tsai’s contributions in the above mentioned areas, he received the Young Scholar Research Award from Yuan Ze University in 2008, which is only the second year since he worked in this university. In 2009, he received the Outstanding Research Award from Yuan Ze University. Until now, he has published 29 papers in top SCI-indexed journals, including Bioinformatics, BMC Bioinformatics, Nucleic Acids Research, IEEE-ACM Transactions on Computational Biology and Bioinformatics, Journal of the American Society for Information Science and Technology, IEEE Transactions on Audio, Speech and Language Processing, and Decision Support Systems. His papers have been accepted by leading artificial intelligence (AI) or natural language processing (NLP) conferences such as IJCAI, COLING, and IJCNLP.

Dr. Tsai is currently a guest editor of Journal of Information Science and Engineering (JISE) for the special issue of artificial intelligence. He served as reviewers for several top journals, including Bioinformatics, BMC Bioinformatics, Information Processing and Management (IPM), Journal of the American Society for Information Science and Technology (JASIST), IEEE-ACM Transactions on Computational Biology and Bioinformatics, JISE, and IJCLCLP.