Across-the-Board Humanity Lesson

When world chess champion Garry Kasparov entered a windowless 35th-floor room this afternoon to begin a six-game competition against one of the world's most powerful supercomputers, far more was at stake than the $700,000 winner's prize. Up for grabs were some very basic human perceptions about computers.

With its relatively simple rules but complex play, chess has been viewed as one of the ultimate expressions of human intelligence.

While computer programs have been able to beat recreational players since the 1970s, grandmasters such as Kasparov traditionally have been able to use brain cells to defeat electronic circuitry.

Enter Deep Blue, a 1.4 ton supercomputer designed by International Business Machines Corp. that can compute 200 million possible moves on a chess board every second. In February last year, a less sophisticated version of the machine stunned the chess world with a first-game victory over Kasparov, but the grandmaster rallied to win the match.

Today's match, won by Kasparov after the computer resigned on the 45th move, featured what some chess experts called aggressive playing by Deep Blue. The match took nearly four hours. The second game is set for Sunday.

The new computer has been dramatically juiced up with faster chips, sharper programming and advice from a former U.S. chess champion, leading some observers to question whether Kasparov can notch another victory.

"This is a very important milestone," said Oliver B. Strimpel, executive director of the Computer Museum in Boston. "Chess playing is a domain that humans have viewed as sacrosanct, something that is quintessentially possessed by human intelligence. This could change all that."

Yale University computer science professor David Gelertner predicted that a Deep Blue victory would be a seminal cultural event, "like Lindbergh flying the Atlantic."

Whether that will happen won't become clear until later this week.

It has long been accepted as fact that machines and other species can be stronger and faster than humans. Computers have been seen as more adept at rote tasks that can be reduced to numerical calculations, such as solving equations or storing driver license information. They've stumbled, however, when faced with challenges that involve more than numbers -- recognizing shapes, understanding spoken speech, and playing high-level chess.

With the Deep Blue project, said Chung-Jen Tan, its manager, "we are demonstrating we can use a computer to solve problems that are not just numerical calculations. Suddenly this computer can be used to solve problems we have always solved with human intelligence."

Win or lose, the result, Tan said, will be staggering. "This is not just about a chess match," he said. "This is really about the future -- how we will be using computers in the future."

At the same time, the computer is not yet what researchers would call "intelligent," though its software draws upon years of so-called artificial intelligence research. Deep Blue doesn't invent moves; rather, it relies on reams of complex, equation-filled computer code that tells it what to do in particular situations. Most important, it can't learn from its mistakes.

"Computers themselves have no more intelligence than a wooden pencil," said Ben Shneiderman, director of the human-computer interaction laboratory at the University of Maryland.

That's a point not lost out on Kasparov, who contends that his unique advantage over Deep Blue is his intuition. Unlike the computer, which is evaluating billions of possible positions as many as 15 turns ahead to determine which piece to move, Kasparov is relying on such non-numerical considerations as his experience and gut instinct, winnowing down myriad possibilities into a few scenarios that he'll play out in his head.

"The computer has generic weaknesses," Kasparov, 34, an ultra-serious Russian who has been the world chess champion since 1985, said in an interview before the match. "I believe they always will be beatable."

The grandmaster, who said he has spent the past eight weeks preparing for the match by trying out some of his moves against a personal computer, has complained that he has not been allowed to see any of Deep Blue's practice games, while IBM scientists reportedly studied dozens of Kasparov's past moves as they programmed the machine.

As a result, Kasparov said he planned to play a very different sort of chess over the next nine days, making nontraditional moves designed to throw the computer off balance. "This is about competition and I'm going to do whatever it takes to win," he said.

The match is taking place in a small, television studio-like room, complete with bookshelves, a large globe, a Persian carpet and a wooden table decorated with Russian and U.S. flags.

Kasparov's opponent isn't actually sitting across from him. Because it can't cool itself with sweat like Kasparov, the computer has been placed in its own specially chilled room down the hall from the competition area. A sleek black monitor on the game table instructs Tan and other IBM researchers where to move Deep Blue's pieces.

Compared with the original Deep Blue, which garnered worldwide fame despite eventually losing 4-2 to Kasparov in Philadelphia last year, the new machine has been souped up in several ways:

It's about twice as fast. The IBM RS/6000 SP machine has 32 super-fast microprocessors working in parallel, meaning it can handle multiple complex calculation simultaneously.

It's more flexible. New tools allow Tan and his team to make adjustments to the computer's playing style after each game, much like mechanics tuning up a race car.

It's savvier. The new Deep Blue has been tutored by former grandmaster Joel Benjamin. Its software is better able to determine which are the strongest moves of the billions it considers.

"The computer is going to be competing at a level we've never seen before," said Tan, who unabashedly has been predicting that Deep Blue will win a majority of the games. Other computer researchers and many chess experts aren't so sure.

"This is an amazing computer but I don't think the computer's explicit commands will be a match for Kasparov's strategic ability," said Tony Marsland, a professor at the University of Alberta and president of the International Computer Chess Association.

Getting a computer to play high-level chess successfully has been a life objective for some software researchers, largely because they thought that writing chess programs would lead to breakthroughs in other areas of artificial intelligence research. To a large degree, industry specialists say, that hasn't happened.

Nevertheless, IBM plans to use the complex logic equations written for Deep Blue for a host of commercial applications that require sophisticated probability analysis, such as research on new drugs and weather forecasting, Tan said.

The company declined to specify how much it has spent directly on developing Deep Blue.

The match also is great publicity for the company. More than 300 journalists from all over the globe are covering the event, creating a bigger media spectacle than the Bobby Fischer-Boris Spassky world championship match in 1972, chess experts said. In addition to selling out a 500-seat auditorium in the basement of a building where the game is being played, IBM is providing live coverage on the World Wide Web at http://www.chess.ibm.com.

Although some have billed the event, which ends a week from today, as a man vs. machine battle, one well-regarded researcher said a Deep Blue victory wouldn't be such a bad thing for humankind.

"Remember where hardware and software came from," said Randall Davis, the associate director of the artificial intelligence lab at the Massachusetts Institute of Technology. "We built it. This contest should be viewed as a triumph of our ingenuity. Either way, we as humans win."