Man vs. Machine
Ajinkya Nene SYBSc
The year was 1997. Right on the brink of the turning of the millennium, 1997 was a period of progress, a time when the inexorable, seemingly inevitable rise of technology was making itself known. Fueled by movies like Terminator, the idea that humanity’s own creations would soon eclipse them and then in short order replace or even discard them was a popular one, just like it is today so many years later.
Thus, it made sense that any “Man versus Machine” matchup would immediately catch the public eye, serving as the focal point of this deeply unsettling undercurrent. But where could such a clear matchup be found? The answer presented itself in the form of a chess match.
For a long time, chess had been the chosen medium for programmers and technology companies to test the effectiveness of their machines, a convenient yardstick to measure so-called artificial intelligence against humans. There was a long string of chess engines built and tested against human players, and indeed they had come a long way, capable of beating chess grandmasters in individual games. However, the true test would be if a chess engine could win a formal match against the reigning world champion under regular time controls. So far, the humans were winning.
Grandmaster Garry Kasparov playing against a computer. Source
The champion of the machines this time was IBM’s Deep Blue supercomputer, built for the sole purpose of playing chess. It would be playing the reigning human world champion, Garry Kasparov.
There are few people throughout sporting history who can claim to have dominated their sport with as much longevity and sheer unwavering competence, as one of the greatest chess players to ever shuffle pieces across the board: Garry Kasparov. A legendary figure in chess, Kasparov was World Number 1 for a record 255 months overall, and is the second-highest-ranked player of all time, surpassed only by Magnus Carlsen.
Before we see the famous Deep Blue v. Kasparov matches, it is worth considering what exactly makes chess so complex, and yet so alluring for computer programmers to code.
Well, each player begins with the ability to make 16 pawn moves and 4 knight moves. That means that when both players have made their first moves there are 20 (total possible first moves) x 20 (total possible first moves for the other player) – a total of 400 potential positions by the end of the first move alone! As subsequent moves are played, the number of potential positions rises exponentially, with the lower bound of total potential positions being 10120 or 10108 trillion – a truly incomprehensibly enormous number.
A computer can either attempt to work through all the possible variations manually (called brute-forcing) or attempt to develop pattern recognition somewhat akin to humans. Deep Blue used a combination of both methods, although mainly relying on brute force.
Kasparov playing against Deep Blue. Source
So it began in 1996, the much-awaited clash of mind and CPU. Deep Blue and Kasparov played a total of six games. Kasparov won three, drew two and lost one; Kasparov won 4 to 2!
It was still a landmark moment for artificial intelligence, however, for it was the first time a computer beat a current world champion in a regular game. To the Deep Blue team, Kasparov’s solitary loss was a sign that they were on the right track.
For the rematch that followed in 1997, the Deep Blue team made several advancements, including engaging grandmasters to further develop the engine’s pattern recognition skills and opening repertoire. Without going into the gory details of each game, it should suffice to say that after six hard-fought games, three of which ended in draws, Deep Blue emerged on top.
It was a highly publicised event, with one publication calling it “The Brain’s Last Stand”, while another referred to Kasparov as the “The Defender of Humanity”. Deep Blue’s victory stands as a milestone today for artificial intelligence, the first time in history that one of our own creations beat us at our own game.
After Garry’s defeat, controversial and close as it may have been, the general media and news coverage had a nervous and even slightly apocalyptic undertone to it, perhaps insinuating that chess was just the first domino to fall to advanced technology. My personal favourite example to illustrate this sentiment is this line from a news show: “In a related story, earlier today the New York Mets were defeated by a microwave oven”. Although satirical, it is not quite far off from what it must have seemed like at the time.
In the years that have followed, advances in artificial intelligence and technology in general have improved manifold. Contrary to public perception when Kasparov lost, chess in fact was not solved nor was there a significant decline in the popularity of chess just because technology had finally largely demystified the game’s inner workings. Indeed, chess has even seen a resurgence in recent years, with technology assisting growth instead of hampering it.
Chess was the guinea pig on which the newest technological techniques of the day were tried and tested. It is an expected outcome at this point that other sports will face a potentially disruptive technology, sooner or later. While it may not be a direct opponent like it was with chess, a deadly intersection of statistics and technology could be applied to pretty much any sport ( remember Moneyball?).
If the aftermath of the rise of AI in chess proved anything at all, it is that rather than ‘Man vs Machine’, ‘Man plus Machine’ is not only a viable alternative, it is a successful one!