Public Release: 

Illinois To Study Biological, Artificial Intelligence Under New NSF Program

University of Illinois at Urbana-Champaign

CHAMPAIGN, Ill. -- A team of University of Illinois researchers has received a $775,000 grant from the National Science Foundation as part of a nationwide $22.5 million initiative announced Oct. 2 to probe how learning works in humans, animals and artificial systems.

Twenty-eight grants were awarded to 20 institutions in the first stage of a three-part NSF program in knowledge and distributed intelligence. The initial grants are for interdisciplinary research in learning and intelligent systems, which covers a broad range of studies designed to promote rapid advances in how humans learn and create.

The U. of I. team will pursue an integrated view of concept learning in humans and machines. The project will involve psychological experiments and artificial-intelligence computational modeling, said Brian H. Ross, professor of psychology. Researchers will investigate and integrate a variety of concept-learning issues that typically had been studied individually, and they will draw upon the resources of different fields.

"People have the ability to conceptualize and file their life's experiences into categories that can be recalled and applied when they are confronted with a new experience," Ross said. "As a psychologist, I am interested in how people can do this. Likewise, in artificial-intelligence systems, it is important to incorporate a way to make use of previous experiences in solving new challenges in new situations."

Working with Ross will be psychology professors Gregory Murphy and Karl Rosengren and computer scientists Gerald DeJong and Leonard Pitt. The five scientists are afffiliates of the U. of I. Beckman Institute for Advanced Science and Technology, where much of the research will be conducted.

The collaborative effort, Ross said, will combine biological and artificial intelligence.

"The computer scientists approach this work with different goals. By working back and forth and seeing the results from each other's viewpoints, we will bring the biological and computer knowledge closer together, allowing us to get farther as a team than we could on our own," Ross said. "A long-term goal is to devise a computer system that builds on previous knowledge so that the system is more like a person."

Nationally, researchers from engineering, computer and information sciences, mathematical and physical sciences, biological sciences, social and behavioral sciences, and education will take part in this NSF inititative. They will try to understand and formulate solutions to such questions as: What kinds of knowledge or skills can actually be learned? How do humans learn? How do other living beings learn? Can artificial systems learn? What kinds of knowledge to they produce?

Such a broad approach to research is exactly what the Beckman Institute is about, said Jiri Jonas, director. "We are very pleased with the NSF award for work on "Learning and Intelligent Systems" as it illustrates the multidisciplinary approach the Beckman faculty is taking in studying and solving complex problems.

"This specific project has faculty from two of our three main research themes: human-computer intelligent interaction and biological intelligence," Jonas said. "The award for this project not only represents yet another example of the strength of multidisciplinary research programs at the Beckman Institute, but also illustrates the complementarity of our main research themes."

Other Illinois institutions receiving grants under the initiative are the U. of I. at Chicago, one grant for $250,000; Northwestern University, two grants totaling $2.16 million for collaboration with the University of Michigan; and the University of Chicago, one grant for $881,000 for work with Temple and Northwestern universities.

-jb-

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.