Microchip challenger proves too good for Kasparov

Click to follow
The Independent Online
WILLIAM HARTSTON

Chess Correspondent

Move over mankind - the age of artificial intelligence has finally arrived. Early yesterday morning, the world's strongest human chess player, Garry Kasparov, was soundly thrashed by one of the world's biggest computers, an IBM giant called Deep Blue.

It was the first of six games between the pair that form a sideshow at a conference in Philadelphia to celebrate 50 years of computing.

There could hardly have been a better way to celebrate. For almost half a century, computer programmers have been predicting the demise of human dominance at chess, yet their efforts met with little success until recently. In 1988, a machine called Deep Thought, which originated at Massachusetts Institute of Technology, became the first to win a tournament ahead of human grandmasters.

Next, the Deep Thought team was contracted by IBM for a five-year project aimed at defeating the world champion. Purpose-built hardware and advances in programming technique then led to massive increases in computing speed.

Deep Thought, which could calculate some 4 million chess positions every second, turned into Deep Thought II at more than twice the speed. Then it became Deep Purple, and finally Deep Blue, whose main thinking unit comprises 256 parallel processors, each with the power of the original Deep Thought. There are considerably more possible chess games than there are atoms in the universe, but at a billion moves a second, Deep Blue can begin to make an impression on the vastness of the game.

It certainly made an impression on Kasparov. It was not the first time he had lost to a machine - he was eliminated from a quit-play tournament in London 18 months ago by a floppy disk - but it was the first defeat of a world champion at the normal tournament rate of 40 moves in two hours.

By heading for an open position with dangerous tactical possibilities, Kasparov may have been trying to prove that he could beat the machine at its own game. If so, it was a grave error. With a clever 17th move, Deep Blue was on the attack, and by move 25 had reached just the sort of wild position in which it is a considerable advantage to calculate accurately a billion moves every second. Against a human opponent, Kasparov's counter-attack might have been dangerous, but here it had no hope.

In the five remaining games, Kasparov may be expected to play more conservatively and head for the type of closed, slow-manoeuvring game in which human judgement and instinct has, in the past, consistently out-performed even the fastest machines. Kasparov has always taken great pride in seeing himself as a final bastion, defending human intelligence against the advancing tides of silicon. Yesterday, however, he was made to look like a King Canute of the chess board, holding up his hand to order the opposing pawns backwards.

The moves, page 5

Leading article, page 18

Comments