Source: William Hsu, 785-532-6350, bhsu@cis.ksu.edu
http://www.kddresearch.org
http://www.mias.uiuc.edu
Photo available. Contact media@k-state.edu or 785-532-6415.
News release prepared by: Beth Bohn, 785-532-6415, bbohn@k-state.edu
Monday, July 30, 2007
K-STATE STUDENTS ATTEND NEW DATA SCIENCES SUMMER INSTITUTE, WORK ON PROJECTS THAT COULD ONE DAY HELP WITH HOMELAND SECURITY
MANHATTAN -- A Kansas State University faculty member and six of his students participated in a new information science institute this summer created to prepare future leaders in the field and to work on projects that may one day help with homeland security.
The first Data Sciences Summer Institute was sponsored by Multimodal Information Access and Synthesis, a research center at the University of Illinois at Urbana-Champaign. It was funded by the U.S. Department of Homeland Security and administered by the U.S. Office of Naval Research as a way to develop research that could help with the department's mission of protecting America. The three-year, $2.4 million grant for the institute is led by the University of Illinois, with K-State and the University of Texas-San Antonio as collaborating institutions. Also funded under three-year awards through the Homeland Security program were Rutgers University, $3 million; University of Southern California, $2.4 million; and University of Pittsburgh, $2.4 million.
The chief goal of the awards program is to meet a growing need for tools to analyze vast amounts of Web-based information in different representations, such as text, captions, comments and images, according to K-State's William Hsu, associate professor of computing and information sciences and director of the university's Laboratory for Knowledge Discovery in Databases. Hsu was among the faculty members who taught at the institute.
"The rapid growth of the Web means that intelligence and business analysts will need to monitor ever-increasing numbers of documents from news feeds and other sources," Hsu said. "Each summer, students participating in this institute have a chance to work with leading researchers in the field of information extraction and machine learning."
The eight-week institute included a monthlong short course on the foundations of data sciences, which presented several fundamental topics from mathematical probability, statistics and computer science.
Hsu said one of the benefits of this approach was that students received immediate and direct exposure to the latest in real-world applications such as intelligent Web search, information extraction and next-generation image and text analysis systems.
Hsu assisted with the short course and taught two tutorials, including two days of a five-day short tutorial on data mining. The second tutorial was a four-week research seminar on advanced topics in learning and reasoning with graphical models of probability.
After the short course, Hsu and his six students worked on one of three intensive research projects working with real Web data. Each student at the institute was assigned one of the projects with a faculty mentor and two graduate student project leaders. The projects included the Virtual Web, a focused Web-crawling project for combining Web search and crawling technology; Named Entity Recognition, which had the objective of "tag" names of people, places and organizations in text; and Image Annotation, which involved collecting, analyzing and marking up images from the Web.
K-State students selected to attend the institute included:
From Manhattan: Waleed Al-Jandal, graduate student in computer science; Vikas Bahirwani, graduate student in computer science; Joseph Lancaster, graduate student in computer science; and Martin Paradesi, graduate student in computer science.
Tony Clark, junior in computer engineering, Paola.
From out of state: Paul Edgington, junior in computer engineering, Fritch, Texas.