UDS-SJTU Joint Research Lab for Language Technology

recently update: 2011-01-13

———————————————————————————————————————

SJTU Department of Computer Science Sino-Germany Joint Lab for Language Technology

UDS-SJTU Joint research lab for language technology has been established in March 2005 and located in the Minhang campus of SJTU. It is the result of the long-term cooperation between Shanghai Jiaotong University (SJTU) and University of the Saarland(UDS) , German research center for artificial intelligence. University of the Saarland, a renowned university in Germany, has solid strength and high reputation in the area of computer science. SJTU is one of the oldest and best universities in China. Both research teams from both universities have made a lot of achievements after many years successful scientific cooperation including joint application and research on projects, research staff exchange, joint workshops holding and so on.

The language technology has developed rapidly with the tremendous expansion of the Internet. Language processing is embedded in many software applications assisting their users in exploiting the wealth of digital content without drowning in the growing flood of information. Human language is the fabric of the multi-medial WWW, because language is the medium in which society and organizations express their precious knowledge. Language technology can make computers understand human languages, realize intelligent processing and provide more service for human beings. As a new platform for the research and cooperation of the two universities, the joint lab will join forces to develop powerful multilingual applications and to shorten their way from the research lab to commercialization.

There are about 20 undergraduate students, graduate students, Ph.D students and research staffs in the joint lab.

Current Research:

1. Project name: Threads and topics detection for news events
   Time Duration: Jan,2009 ~ Dec.2011
   Introduction:
   Categorizing news reports and list of news items can not satisfy the requirements that people need to know not only what is going on, but also how the event develops and evolves. How to organize the news event automatically and how to represent its development as the time elapse, have become one of the problems need to be solved in this information overloading world. The project will focus on three points: 1) the topic model of news event based on LDA in order to establish a description model of different news events 2) related model of news events, to capture the relationship among a seed event and related events, especially the causation and elaboration 3) detecting threads and themes of a news event to reflect different aspects of a news event and its developments in order to organize the event automatically. By exploring the internal structure of a news event, we hope to show the evolvements of hot topics, such as SARS or other news events automatically.
   Group leader: Fang Li,     Team Leader: Chu Keming,     Group Members: Wang Quanjian, Ding xiaoshan,Letian Wang, Zheyi Qian, Shun Wang, Bin Shan
2. Project name: User Interest analysis based on user behavior and page contents
   Introduction:
   Based on the previous work on Intel project, we will focus on the following tasks:(1). improve the clustering method (2) map the clustering result(key words) on to the interest category. (3). finding the interest trends along with the time.
   Previous members：Li Feng, Zhou Kai, Li Yihong, Wang Xinguang, Wu yanchen
   Current Team leader: Zheng Donghui
   Group member: Li Yihong, Wang Xinguang, Wu Yanchen, Stephan
   Group leader: Li Fang
3. Opinion mining techniques and applications for Chinese texts (proposed by our research group): mine opinions of free texts regarding automobiles on web, and then make statistics of opinion categorization for evaluated objects, especially polarity and strength of topics (attributes), finally, give visualization results depending on the above statistics.
   Group leader: Linlin Li. Group menbers: Ke Chen, Hang Yin, Xiaokai Zhang

Completed Project:

1. An Intelligent platform for Information Retrieval (2004-2006 Leader: Fang Li)
   Group leader: Ying Han, Group Members: Kebin Liu, Lei Liu, Li Feng, Kai Zhou, Feng Li
2. Teenage interests analysis based on web access log (Intel fund   Leader: Fang Li):
   We use language technologies to analyze the access log and the content of Web pages and generate teenage’s interest based on clustering result.
   Group leader: Feng Li. Group Members: Yanchen Wu, Xinguang Wang, Yihong Li

Contact

• Address： Electronic,Information and Electrical Engineering Building 3, Room 501 502, Shanghai Jiaotong University, 800 Road Dongchuan, Minhang District, Shanghai.
• Zip Code：200240
• E-mail：fli@sjtu.edu.cn

Leave a Reply

You must be logged in to post a comment.