4 Classic Battles Between Man (or Horse) and Machine
By Rob Lammle
Nothing could prepare Jeopardy! champ, author and regular mental_floss magazine columnist Ken Jennings for the battle he's about to face. In episodes that will air next month, Ken will take on fellow Jeopardy! genius Brad Rutter and Watson, an IBM supercomputer. Yesterday, the three challengers squared off in a trial round for the press. So far, it's not looking too good for mankind – Watson came in first. The final match-up should be one of epic proportions, but it's not the first time silicon has squared off against a biological opponent. Here are four stories of man versus machine that are sure to get your gears turning.
1. Garry Kasparov vs. Deep Thought and Deep Blue
Garry Kasparov became a Grandmaster chess player at the age of 17, held the World Championship title for nearly 22 years, and is still the highest-rated player to have ever touched a rook. To those outside the chess world, though, he is probably best remembered for playing against a computer.
The competition between Kasparov and computers had been long-standing; starting in 1989 when the Russian handily defeated an earlier IBM supercomputer, Deep Thought, in a 2-0 shutout.
However, in February 1996, IBM returned with Deep Blue, a supercomputer with custom-made chess-playing processors capable of analyzing 100 million moves per second. This time around, Deep Blue surprised everyone by defeating Kasparov in the first match of their 6-game battle. But Kasparov fought back, beating the machine 4-2. Unfortunately, his victory was short-lived.
The engineers at IBM took Deep Blue back to the lab and were able to double its processing power, bumping its analysis to 200,000,000 moves per second, up to 30 moves ahead. This proved to be too much for Kasparov in their May 1997 rematch, as the Grandmaster bowed to the machine by a score of 3 1/2 points to 2 1/2 points.
Kasparov asked for a tie-breaker match, but IBM refused. Instead, they retired Deep Blue and dismantled it. Parts of Blue now reside at the National Museum of American History in Washington D.C., and at the Computer History Museum in Mountain View, California.
Just because his most famous rival went into retirement didn’t mean Kasparov was done with computer opponents. In 2003, he played against two different chess programs - Deep Junior (named as an homage to Deep Blue, but not related) and X3D Fritz. Both matches resulted in a draw.
2. A Horse vs. Tom Thumb
Aside from having a great, flourishing name, John Hazlehurst Boneval Latrobe was a lawyer working for the B&O (Baltimore & Ohio) Railroad. As the railway was being built, Latrobe witnessed an unusual test of endurance and speed between one of America's first Iron Horses and a four-legged, flesh-and-blood competitor.
The year was 1830 and the first American-built steam locomotive, nicknamed Tom Thumb, was being tested on a 13-mile run between Baltimore and Ellicott Mills (now Ellicott City), Maryland. The locomotive ran on a track that was parallel to an existing set of tracks used by a horse-drawn cart. On August 28, the horse driver, probably feeling a bit threatened by this new technology, challenged the designer and engineer of the locomotive, Peter Cooper, to a race. While Tom Thumb could only chug along at a top speed of 18mph – about 10mph slower than a good horse could gallop – the machine could maintain that speed over the distance, whereas the horse would eventually slow down. Thinking he had a sure win, Cooper accepted and the two lined up on their respective tracks.
The horse took an early lead as the locomotive required more than a quarter of a mile to build speed. But once it had a full head of steam, the locomotive quickly caught up. For a brief stint, Latrobe says, “The race was neck and neck, nose and nose.” But soon, “The engine passed the horse and a great ‘hurrah’ heralded the victory.” Just as the horse cart was about to concede, a pulley slipped on Tom Thumb, causing the locomotive to lose steam. As the machine cruised to a stop, the horse burst back into the lead. Cooper was able to get the locomotive fixed and regain much of the lost ground, but in the end, nature won out over machine.
Despite the outcome, the railroad company was convinced of the locomotive's dominance. On July 31, 1831, less than a year after the race took place, all horses on the B&O Railroad were officially replaced with steam locomotives.
3. Piet Mondrian vs. an IBM 7094 digital computer
On the surface, the works of Piet Mondarin, one of the founding fathers of the modern art movement, appear to be simple. But once you start examining the finer details, you begin to uncover a deeper meaning and sense of purpose behind his “abstract geometry” pieces.
A. Michael Noll was a creative pioneer in his own right, as one of the first people to use computers in the creation of artistic works. Noll's early-1960s art required hours of designing complex computer programs that would be interpreted by the room-sized computers at Bell Laboratories in New Jersey, where Noll worked. The printed output of the computer's calculations formed the basis for his art.
One of Noll's 1964 experiments was an attempt to recreate the style of Mondrian's artwork. He wanted to know if a computer, using random processes, could create images that were just as stylistically interesting and appealing to the eye. To find out, Noll chose to emulate Mondrian's 1917 piece, Composition With Lines, a series of sometimes-intersecting horizontal and vertical lines, arranged to create the vague outline of an undefined circle.
Using an IBM 7094 Data Processing System control panel, Noll was able to write a program that would draw lines anywhere within a defined circle, so that every possible point would be just as likely to contain the beginning or end of a line. This made the drawing purely random, unlike the more precise and purposeful aesthetic of Mondrian's original work. With these instructions, the computer created Noll's version, Computer Composition With Lines.
Only 28% were able to correctly identify the computer-created art. And if you look at the two pieces side-by-side, it's easy to see why. The Mondrian piece has an underlying order, just as one might expect a computer to “think” when drawing. However, a surprising 59% of respondents said they actually preferred the computer version, describing it as more varied, more imaginative, and more abstract than the Mondrian.
4. Marion Tinsley vs. Chinook
Marion Tinsley is widely considered to be the best player of checkers (also called draughts) in history. He went undefeated during World Checkers Championship games, held the title multiple times, and only lost seven matches total over his 40-year career.
Tinsley's greatest (and some say only) competition came in the form of Chinook, a computer program written by Johnathan Schaeffer, a professor at the University of Alberta. Schaeffer and his team of researchers started writing Chinook in 1989. By 1990, the software had advanced enough to earn second place (behind Tinsley) in a qualifying tournament to gain a spot at the World Championship games.
The Championship was a pretty even match-up with the two competitors playing to 33 draws. But in the end, Tinsley beat Chinook with four wins, while the computer program handed Tinsley two of his seven career losses. During the battle, after Schaeffer had made a move for Chinook, Tinsley looked across the table and said, “You're going to regret that.” Sure enough, 26 moves later, Chinook had no choice but to resign the game. After the tournament was over, Schaeffer went back through his records and found that Tinsley was right - it was after that one bad move that Tinsley took control of the board and Chinook never had a chance to win.
The two opponents had a rematch in 1994, but Tinsley had to withdraw after only six games due to his failing health. After Tinsley left the competition, number two-rated player Don Lafferty took over and played Chinook to a draw. However, because Tinsley had conceded, Chinook became the first computer to become a world champion in any game with human opponents. Sadly, a week after his withdrawal from the match, Tinsley was diagnosed with pancreatic cancer and died less than a year later.
Chinook and Lafferty battled again in 1995 when Chinook defended its title with a 1-0 win (and 31 draws). After the match, Schaeffer took Chinook out of competition so he could focus on the theory behind playing checkers. To that end, in July 2007, Schaeffer and his team announced in Science journal, that they had “solved” checkers, meaning they could determine “the final result in a game with no mistakes made by either player” after only one move. It would now be pointless for any human to play Chinook, as every match would either end in a draw or defeat at the hands of the computer.
* * * * *
So what do you think - Do Ken and Brad have a chance against Watson? Or will the machine win again in the end?