While working in the snowy Midwest this week, staff at the RedState Department of History were left with little to do in the evenings. With activities cancelled most of them decided to take in a relaxing game of chess.
As is generally the case in circumstances such as these, this choice coincided with an anniversary, and today’s was rather remarkable. On this date in 1996, an IBM computer known as “Deep Blue” became the first machine to defeat an active world chess champion when it defeated Garry Kasparov.
Chess players are known for meticulous attention to the small things and the ability to juggle a very large number of potential combinations at the same time. The best players can take on a room full of opponents at the same time and come out victorious.
Yet in this particular case, the new field of artificial intelligence was just coming into prominence, and IBM Research was actively looking to create the first supercomputer that could defeat a world chess champion.
Their first attempt was known as “Deep Thought”, which began as a project of a team of graduate students from Carnegie-Mellon but which came under IBM’s umbrella soon afterward.
The machine got a shot at Kasparov in 1989 in New York City and the 26-year old native of Azerbaijan defeated the computer twice, claiming he was protecting the human race in so doing. The first game of that match can be recreated here.
Fast forward seven years — and IBM was ready to try again. Now known as “Deep Blue”, the modified computer contained memory from 700,000 grandmaster games and shocked the world by taking the first game of a six-game series with Kasparov on this date in 1996.
The reverberations were enormous – a computer could defeat a human being. That said, Kasparov rallied to win the six-game series 4-2, restoring human superiority for the time being.
The following year, IBM tried again. Its computer had been massively upgraded, to the point where it could evaluate 200,000,000 positions per second. This computer, known to some as Deeper Blue, did what its predecessors couldn’t – it defeated Kasparov 3 1/2 to 2 1/2 in a six-game series to become the first computer to defeat a world champion in an entire series.
To the proud Kasparov, that meant the computer must have cheated. He accused the machine’s human handlers of reprogramming it during matches, which would have been a violation of the rules.
Before the match, he also asked IBM to provide him with a summary of the computer’s previous games so he could get to learn his opponent. IBM refused. After it was over, Kasparov demanded a rematch. Not only did IBM decline, it retired Deep Blue.
Its creators described Deep Blue as an example of “brute force” computing rather than artificial intelligence. Simply put, the computer had been programmed with a massive amount of power and memory which enabled it to overwhelm Kasparov, whose human error in the sixth match of the second series cost him the event.
To see a full-length documentary on Kasparov and Deep Blue, click here. Since it’s snowing again here in the Midwest, we’re off to watch a little YouTube, but you all enjoy your Sunday and today’s open thread!