Deep Blue's computer chess victory creates deep puzzles about humanity

AS THE chess pieces were being put away, and as IBM's Deep Blue super-computer was being powered down, the question of what world chess champion Garry Kasperov's stunning defeat by a computer means for human intelligence was being debated.

For Chung-Jen Tan, head of the team of programmers and chess experts at International Business Machines, Deep Blue's victory was nothing less than a monumental landmark. "One hundred years from now, people will say this day was the beginning of the information age. Historically, for mankind, this is like landing on the moon," Tan said.

For others, the Kasparov defeat meant little more than a victory of computer brawn over brain. The consensus is that it was simply an inevitable result. All it proved was that if you could build a computer that could calculate enough chess positions fast enough - in Deep Blue's case that is about 200-million per second - and you could give it a simple set of rules, you would eventually win.

The chess match was promoted by IBM from the beginning simply as a test of its programming and hardware that would lead to applications in medicine and other areas with benefits for all.

But in reality the contest was widely seen as a match of man against machine. Kasparov certainly saw it that way. In the first series of games, played last year, he spoke about defending the "dignity of humanity" in the face of the cold, calculating power of Deep Blue.

Kasparov won that first contest, but this time it was that cold, unemotional aspect of Deep Blue that, in many ways, won the contest. Kasparov made mistakes and became flustered, resigning a game he could have played to a draw.

A large part of chess is a the psychological aspect at which Kasparov is a true master - as long as he is playing against a human. To that extent, Deep Blue managed to unnerve Kasparov, making it difficult to say if it was solely IBM's technology and the skill of its programmers that won the contest. For example, Kasparov became convinced that some games were headed for a draw simply because Deep Blue had already seen so far ahead that to continue was futile.

Kasparov admitted that he was afraid of playing Deep Blue and was not sure why - showing that the machine had a psychological advantage. "Gary has been used to playing against humans for more than 25 years. He didn't stand up to the pressure of playing against a computer and he simply cracked in the end,"' says Frederick Friedel, computer adviser to Kasparov.

"We did nothing to deliberately unnerve Kasparov," says Gabriel Silberman, an IBM researcher and the chess team co-ordinator. "We did everything we could to try to make him comfortable. He unnerved himself." Silberman added that the IBM team could have played the psychological part of the game by choosing unorthodox strategies or using a "rapid fire" mode in which moves are made very quickly.

An important question has been whether Deep Blue's performance represents true artificial intelligence. Nearly 50 years ago the UK computer pioneer Alan Turing pondered the issue of what would constitute artificial intelligence. He proposed what is now known as the Turing test - that if a person could converse with a computer via a keyboard and monitor, and could not tell whether he or she were computing with a computer, then a degree of artificial intelligence had been achieved.

Although Kasparov clearly knew he was playing against a computer, he said he detected glimmers of true intelligence in Deep Blue's chess moves. If Kasparov had to guess whether he was playing a computer or a person, Deep Blue may have passed the Turing test, if you could classify chess moves as a "conversation".

Whether or not Deep Blue's victory constitutes true artificial intelligence, it does represent the slow but inexorable gain that computer-based intelligence is bound to show over the coming years.

Earlier this year, at the Association of Computing Machinery's 50th anniversary conference in Silicon Valley, experts debated just what it is that makes us human. Most predicted that within 50 years computers would match the intelligence of humans.

"I fully expect computers to become as intelligent as my dog within 20 years, and they will match the intelligence of humans within 50 years," said Nathan Myhrvold, chief technology officer at Microsoft.

This, however, leaves some intriguing questions. As computers match humans in a increasing number of activities, what then constitutes being human?

Having failed to defend the dignity of humanity, a tired, frustrated and angry Kasparov vowed that he would "tear to pieces" Deep Blue in a future contest. He suggested a 10-match contest.

IBM says it is considering a rematch but it will not sponsor it.

But at least one thing is clear: IBM scored a public relations coup, netting millions of dollars worth of free worldwide publicity. However, Deep Blue has another opponent to face. Susan Polgar, the women's world chess champion, has challenged the machine, saying she wants to pit her "woman's intuition" against the computer. - Financial Times

Source: btimes.co.za