[an error occurred while processing this directive] [an error occurred while processing this directive]
[an error occurred while processing this directive]
  1. K-State Home >
  2. Media Relations >
  3. January news releases
Print This Article  

Source: William Hsu, 785-532-6350, bhsu@k-state.edu
Pronouncer: Hsu sounds like "shoe"
News release prepared by: Erinn Barcomb-Peterson, 785-532-6415, ebarcomb@k-state.edu

Tuesday, Jan. 13, 2009

K-STATE COMPUTER ENGINEERS WORKING ON SYSTEM THAT CAN USE THE INTERNET TO TRACK HUMAN AND ANIMAL DISEASES BY EXTRACTING INFORMATION FROM WEB SITES

MANHATTAN -- Information that could help scientists and public health officials mitigate the spread of human and animal diseases may be as close as the nearest Internet connection.

However, making sense of all of the information available is no easy task for human beings. That's why Kansas State University engineers are getting computers to do the heavy lifting.

William Hsu, an associate professor of computing and information sciences, is working with Marty Vanier, director of operations for K-State's National Agricultural Biosecurity Center, to create a system that would extract information from the Web that offers clues about disease outbreaks. With a grant from the U.S. Department of Defense, K-State is the lead institution and is working with Elder Research Inc. in Charlottesville, Va., Radiance Technologies Inc. in Huntsville, Ala., and the University of California-Davis.

Hsu said the objective is to scour news stories and other online public sources that could offer clues about where diseases have started, how many animals were involved and how the diseases spread. With this information, they also could plot outbreaks and governmental responses. By developing the proper tools, Hsu said that such information could be presented in the form of timelines or maps.

Hsu said the challenge comes from the fact that online sources may have valuable bits of information, such as where disease outbreaks occurred, but these online nuggets often are imbedded in text like news stories rather than in a database format.

"Extracting text from Web pages is harder than it sounds because Web pages have text such as banner ads, and systems don't know exactly where text begins and ends," Hsu said. "Our system includes software that will clip text more accurately."

Research on the system was presented in fall 2008, including in November at the International Conference on Artificial Neural Networks in Engineering in St. Louis and in September at the International Conference on Database and Expert Systems Applications in Turin, Italy.

Hsu said the project has been a source of research for both graduate and undergraduate students. Current K-State students involved in the project include:

From Greater Kansas City: Andrew Walters, senior in computer science, Merriam; Danny Jones, senior in computer science, Overland Park; and John Drouhard, sophomore in computer engineering, Prairie Village.

Daniel Greene, senior in computer science, Junction City.

From Manhattan: Waleed Aljandal, doctoral student in computer science; Abhijit Erande, master's student in computer science; Wesam Elshamy, doctoral student in computer science; Jincheng Gao, master's student in software engineering; Shuang Hao, master's student in computer science; James A. Louis, master's student in software engineering; Sowjanya Karumuri, master's student in software engineering; and Svitlana Volkova, master's student in software engineering; Jing Xia, doctoral student in computer science.

Landon Fowles, sophomore in computer engineering, Norwich; Imran Hameed, senior in mathematics and computer science, Olathe; Sam Henke, senior in computer science and mathematics, Salina; and Caitlin Luttjohann, sophomore in mechanical and nuclear engineering, Topeka.

The project also employs a research associate, Tim Weninger, who graduated in December 2008 from K-State with a master's in computer science.

Hsu also is working on other information extraction projects, including a bioinformatics project with Doina Caragea, assistant professor of computing and information science. Since 2007, the professors have been using public data from wet labs at Stanford and Carnegie Mellon universities to create networks that link proteins to one another and genes to one another in an organism.

"This can be used to understand gene function, which is one of the big challenge problems in systems biology," Hsu said. "This information could be used for research on anti-cancer drugs, for instance."

The research will appear in several upcoming publications and will be presented in March at the Institute of Electrical and Electronics Engineers Symposium on Computational Intelligence in Bioinformatics and Computational Biology in Nashville, Tenn.

 

 

[an error occurred while processing this directive]