Robotics Reference
In-Depth Information
Through a combination of more advanced chip design and faster chip
speeds, the IBM group created a machine for the 1997 match employing
480 Chess chips that could analyse a total of some 200 million moves
per second. This raw computing power, combined with a big increase in
Chess knowledge from its grandmaster “advisors”, most notably a former
U.S. Champion, Joel Benjamin, enabled Deep Blue to play remarkable
strategic Chess. When the program outplayed Kasparov in the second
game of the match, slowly but steadily manoeuvering its pieces onto bet-
ter and better squares, the price of IBM stock immediately rose by $4.
Kasparov's morale was shot to pieces and he was unable to play at his best
for the remainder of the match. After five games the score was still level at
2
1
/
/
2
but Kasparov was in no fit psychological state to play the sixth
and final game (see
Figure 24)
, with the result that he was forced to re-
sign after no more than one hour's play. AI had made history. Feng Hsu
and his colleagues, Murray Campbell and Jospeh Hoane, had achieved
one of the original goals of Artificial Intelligence as defined during the
Dartmouth workshop of 1956.
Despite Kasparov's loss to Deep Blue in 1997, it appears that the
very best human Chess players, Kasparov included, are still (in 2005) a
little stronger than the best programs, but not sufficiently so to be able
to outperform a top program under match conditions. One reason why
2
-2
1
Figure 24.
Garry Kasparov in play against Deep Blue, operated by Feng-hsiung Hsu, the
sixth and decisive game, New York, 1997 (Courtesy Murray Campbell)
