By Bhavya Sharma
On Feb. 10, 1996, College of Computing & Informatics Professor
Jeffrey Popyack, PhD witnessed one of the first forms of artificial intelligence win a chess game against a world champion under regular time controls.
A 6-foot-5-inch super computer, known as “Deep Blue,” developed by a team of IBM scientists defeated former World Chess Champion Garry Kasparov with 40 moves in two hours.
The Atlantic recently
published an account of the event with quotes from Popyack on his experience witnessing the “anguish” that Kasparov exuded during the game. The article recounts a short history of chess-playing computers and also mentions many recent developments in artificial intelligence; such as the IBM super computer called “Watson” that defeated Ken Jennings on
Jeopardy!
These advancements in technology were astonishing at the time, however, Popyack points out that, even today, no computers exist that can program other computers, which renders the question of artificial intelligence replacing humans highly doubtful.
In his
personal account of the game, Popyack mentions that “to beat a computer at chess, you need to be able to see further ahead than it does.” Popyack believes that computers suffer from the “horizon effect,” in that computers can only look ahead to a certain number of moves (due to limitations of time and space). Thus, many of the moves that computers generate can be worthless, and detract from the computer’s ability to focus on only the “good” moves. Nevertheless, all moves must be evaluated, as Popyack notes.
Popyack also thought that Kasparov should have considered how far the computer could see, and then set a trap that Deep Blue might not expect. Kasparov, who prior to his first loss, scoffed at the idea of a chess-playing computer, now took his competitor more seriously. He won three games in a six-game match, to be defeated again by an upgraded Deep Blue in 1997.
###