Page 50 - profile2014.indd

特聘研究員
許聞廉 Wen-Lian Hsu

Distinguished Research Fellow
Ph.D., Operations Research, Cornell University
Tel: +886-2-2788-3799 ext. 1804 Fax: +886-2-2782-4814
Email: hsu@iis.sinica.edu.tw
http://iasl.iis.sinica.edu.tw/hsu/

● B.S., mathematics, National Research Description
Taiwan University, (1973)
● Ph.D., Operations Research, My main research interests include the following: graph algorithms; natural language un-
Cornell University, (1980) derstanding; and bioinformatics.
● Intelligent Chinese Input In the area of graph algorithms, we have performed extensive and ground-breaking work
Software -- Top ten Most Dis- on two fundamental classes of special graphs, namely planar graphs and interval graphs.
tinguished Chinese computer We introduced a new data structure, called PC-tree, which greatly simpli es the recognition

products（十大傑出中文資訊產
品獎）in Taiwan (1993) of these two classes of graphs. PC-tree is a natural representation for planar graph embed-
ding. Our new planarity test based on PC-trees is simple, elegant, and yields a linear time
● Outstanding Research
Awards（國科會傑出研究獎）by algorithm for identifying maximal planar subgraphs.
NSC (1991-92, 94-95, 96-97) In the eld of natural language processing, we have designed a Chinese input system called

● K.T. Lee Research Breakthrough GOING, which automatically translates a phonetic (or Pinyin) sequence into characters with
Award（李國鼎穿石獎）(1999) an accuracy rate close to 96% based on context. It received the Distinguished Chinese In-
● NSC Designated Research formation Product Award（中文傑出資訊產品獎）in 1993. It is currently being used by more
Fellow（國科會特約研究員獎） than a million people in Taiwan. Recently, we developed a search typing software, déjà vu,
(1999) which can be applied to any language to speed up typing and search, especially on mobile
● NSC Appointed Outstanding devices.
Research Award（國科會傑出
特約研究員獎）(2005) We have worked on several basic modules in Chinese natural language processing. Our

research on Chinese word segmentation won rst place in the 2006 SIGHAN CityU closed
● Academia Sinica Investigator track competition. Identifying named entities, such as persons, locations, and organiza-
Award（中央研究院深耕獎）
(2005) tion names, is very important for natural language understanding. In the past, we have
developed a mature, machine-learning-based named entity recognition (NER) system,
● IEEE Fellow (2006)
which won second place in the 2006 SIGHAN NER competition. Our main focus is on natu-
● Teco Award（東元獎）(2008) ral language understanding (NLU). While building fundamental component modules, we
● Pan Wen Yuan Distinguished successfully developed a knowledge annotation and inference kernel, InfoMap, for the se-
Research Award (潘文淵研究傑 mantic analysis of natural language, which can be applied to a wide variety of application
出獎) ( 2010). systems in natural language processing and biological knowledge bases. Utilizing InfoMap,

● ITCM Medal of Honor 資訊學會榮 we won rst place in the Chinese Question Answering contest in NTCIR held in Tokyo, Ja-

譽獎章 (2013) pan, in 2005 and 2007. We also won rst place in NTCIR-10 Recognizing Inference in Text
Task in 2013. Our recent pattern summarization technique has created a new direction for
machine learning, which would eliminate the weakness of both rule base and statistical ap-
proaches. Our natural language processing techniques have also been applied to biological
literature mining. We have successfully produced biological entity and relation recognition
algorithms integrating biological knowledge and machine learning techniques. In 2009, we
won rst place in the BioCreative II.5 gene normalization task. In addition, we have worked

on gene-disease relation extraction for hypertension, obesity, and diabetes; this work will
be continually extended to other diseases.
For bioinformatics, we have developed a suite of software for protein quantitation (Multi-Q,

50 特聘講座/特聘研究員 Distinguished Chair and Distinguished Research Fellows