Two CU computer science professors receive NSF Early Career Awards
By Bill Steele
Cornell Chronicle (September 11, 2003)
Two Cornell researchers have been awarded Faculty Early Career Development (Career) grants from the National Science Foundation (NSF). They are Thorsten Joachims and Jayavel Shanmugasundaram, both assistant professors of computer science. Coincidentally, both are working on better ways to search the Web.
The Career program is the NSF's most prestigious award for new faculty members, designed to recognize and support the early career-development activities of those teacher-scholars "who are most likely to become the academic leaders of the 21st century." Each award carries a substantial grant to support the faculty member's research.
Before coming to Cornell in the fall of 2001, Shanmugasundaram worked for two years at the IBM Almaden Research Center in San Jose, Calif., a major center for database research. He will use his five-year, $406,750 NSF grant to develop a new data management system that can search plain text or query structured databases and, if necessary, combine the results. He hopes to make it possible to search plain text with the kinds of sophisticated queries possible in databases, such as searching for a limited range of prices or dates.
One application is to make accessible what's called the "deep Web"-- information stored in formally structured relational databases that can be reached online but ordinarily can be searched only one at a time by proprietary systems. Examples range from eBay's list of items for auction to lists of used cars maintained by thousands of individual car dealers. A prototype Web site that will allow sophisticated searches of Shakespeare's plays will go online in a few months, he said.
Shanmugasundaram received a B.E. in computer science from the Regional Engineering College, Tiruchirappalli, India, in 1995, an M.S. in computer science from the University of Massachusetts-Amherst in 1997 and a Ph.D. in computer science from the University of Wisconsin-Madison in 2001.
Joachims, who receives a $400,000 NSF grant, is working on tools that will help individual Web users zero in on the material they want out of the vast array that sometimes appears in response to an Internet search. "The top 10 results are prime real estate," he pointed out. They should be the items the user most wants to see, not necessarily the ones Google or some other search engine selects, he said.
"I want to have a system that gets better the more you use it, that learns by looking over your shoulder," he explained. He is developing software that will notice which results users click on, then rank search results in accordance with their interests. Academics, for example, may be more likely to click on links to .edu than .com sites. "And when I type in 'Michael Jordan,' I usually mean the professor at Berkeley, not the basketball player," he said.
Joachims plans to build a prototype search engine in about a year, possibly to search Cornell's online physics arXiv, a collection of scholarly papers in physics and mathematics, which offers special challenges for searches.
Joachims received his B.S. in 1997 and Ph.D. in 2001, both in computer science, from the University of Dortmund, Germany. From 1994 to 1996 he was a visiting scholar at Carnegie Mellon University. He became a postdoctoral research associate with the Knowledge Discovery Team of Fraunhofer Institute of Autonomous Intelligent Systems in Germany before coming to Cornell in the fall of 2001.