| |
    
|
by Darlene Gorrill
After watching a computer playing chess at the University of Alberta, Murray Campbell decided to combine a passion for chess with his newfound interest in computing.
In an unusual twist of fate, those interests took Campbell to the top of the chess world. A research scientist at the IBM T.J. Watson Research Center in New York, Campbell participated in the development of Deep Blue, the first computer to defeat a world chess champion. In 1997, Deep Blue beat Garry Kasparov, avenging an earlier loss to Kasparov. The 1997 chess match fulfilled a longstanding challenge in computer science -- and heightened the debate over the role of artificial intelligence.
Campbell will talk about the development of Deep Blue and the implications of the computer's victory at the 1998 Honeywell-Sweatt Lecture: "Did Deep Blue's Win Over Chessmaster Kasparov Signal a Deep-Sixing of Humankind as Well?" The lecture begins at 7 p.m., May 7, followed by a reception at 8 p.m., in Room 3-210 of the Electrical Engineering and Computer Science Building at the University of Minnesota.
While Deep Blue enjoys status as a champion, the computer also raises questions about artificial intelligence. If intelligent computers such as Deep Blue can beat the best that humans have to offer, what does that mean for society? Are machines already smarter than humans? Campbell will explore those issues in his lecture.
Many factors contributed to Deep Blue's success, including a single-chip chess accelerator, a large-scale parallel system, selective search algorithms, and a complex evaluation function, says Campbell. As an expert chess player and former chess champion of Alberta, Canada, Campbell made good use of his chess expertise and focused much of his work on developing the computer's evaluation function.
"That's important," says Campbell, "because even if you can search many moves into position, search forward all the possible moves, you still have to evaluate those positions at the end of the sequence of moves that you've looked at. If you evaluate them incorrectly, you're going to play poor chess." Campbell worked to help Deep Blue evaluate positions continuously.
While completing his Ph.D. in computer science at Carnegie Mellon University, Campbell met doctoral student Fenghsiung Hsu, who had developed a single-chip move generator. The two collaborated in 1986 to construct a chess-playing computer, Chiptest, and both joined IBM in 1989.
Now Campbell's new computer, Deep Blue, has become the first computer to defeat a grandmaster in tournament play.
What's the next step for the battle between Kasparov and the computer? Can Kasparov rebound and beat new and improved versions of chess-playing computers?
He sure will try. He will come back well prepared and armed with new ideas for playing against the computer, predicts Campbell. But, with nearly 50 years of computer chess research contributing to its win, Deep Blue is no flash-in-the-pan success story. Indeed, the chess computer's time to compete at the highest level has come. "I did think that some day it would happen, that we'd be competing with a world champion, and I, in fact, thought it would be about now."
