Man vs machine chess match ends in stalemate

Gary Kasparov chose to draw the deciding game of his match with the computer program Deep Junior on Friday rather than push for a win and risk defeat.

Kasparov was booed by some members of the audience for accepting the draw while having a far stronger position on the board and a better chance of winning.

But the world number one said he feared even the smallest mistake could have lost him the match. "Of course I wanted to win, but the top priority on my agenda today was not to lose," Kasparov said at a press conference held in New York after the final game.

Kasparov went into the match keen to avenge his 1997 loss to IBM's supercomputer Deep Blue. He won the first game of the match with apparent ease, after spotting a weakness in Deep Junior's opening repertoire of moves. But Deep Junior's tight defense and unpredictable style of attack frustrated the former world champion for the rest of the contest. Deep Junior also won one game and the remaining four were drawn. One point is awarded for a win and a half point for a draw.

Unpleasant opponent

Kasparov described as "unpleasant" the experience of playing against an opponent that does not feel fatigue and emotion. "If you make one mistake, you are out of business," Kasparov said.

When he lost to Deep Blue in 1997, Kasparov accused the computer's designers of altering their program's style of play before each game, making Deep Blue an entirely different player each time. Kasparov said the latest match was much fairer. "It's a first match that was a purely scientific match, because we had fair conditions for both the human player and for the machine," he said.

The match was the first between a computer and a human to be recognized by the International Chess Federation.

Kasparov is still the highest ranked chess player in the world despite losing his world title to Vladimir Kramnik in 2000. Kramnik only managed to draw against another computer program, Deep Fritz, in an eight game contest played in Bahrain in October 2002.

Text 4

Computer chip noise may betray code

The noise emitted by computer chips could help code breakers decipher encrypted messages, according to preliminary research carried out at the Weizmann Institute in Israel.

Adi Shamir and Eran Tromer sampled the high-frequency audio produced by computer central processing units (CPU) — the highly complicated devices that perform the majority of calculations inside computers - in a recording studio.

They discovered that they could distinguish between different cryptographic keys being processed by the chip, according to the frequency of the sound emitted.

They also found they could determine the length of a string of characters by measuring the duration of certain sounds. This is because these correspond to the amount of time taken to process the key.

These details could, in theory, make it substantially simpler for an assailant to break the code used to protect valuable data on a computer, «Our preliminary analysis of acoustic emanations from personal computers shows them to be a surprisingly rich source of information on CPU activity," the researchers write on a web page outlining their work. "This looks like a very exciting finding," says Markus Kuhn, at Cambridge University in the UK. "It's another indication that there are many types of compromising emanations still waiting to be discovered."

Electromagnetic signals

Kuhn notes that others have examined ways of monitoring the power-supply fluctuations exhibited by chips to remotely gather information about cryptographic activity.

Peter Honeyman at the University of Michigan is also impressed by the technique and says it seems especially promising because it is not disrupted by additional noise from a computer.

The method detects only high-frequency CPU noise and filters out lower frequency sounds like the whirr of the fan inside a computer. Computer scientists have already demonstrated a variety of different ways of using inadvertent emanations to eavesdrop on computers. It is well known that the electromagnetic emanations from a monitor can be used to reconstruct information displayed on the screen. It is for this reason that some banks install computers within "Faraday cages", which block out electromagnetic signals. In March 2002, Kuhn demonstrated that indirect light from a computer monitor, reflected on awall for example, could be used to reconstruct a monitor image. This is possible because a monitor actually creates the screen a pixel at a time at very high speed.

Text 5