AI was just science fiction just about 30 years ago, but now they are more than just a reality. Their realization came into being with us, achieving specific milestones at different points in history.
And one among the most notable ones was the game between the supercomputer Deep Blue and the world-renowned chess champion Garry Kasparov.
The reason why it is so important is that Deep Blue did something awe-inspiring. It triumphed over Garry Kasparov, and this was back in 1997.
Even then, nobody called it an AI, but this was the start of training machines to become better than humans, not just in physical work, but also in calculation and planning.
What is Deep Blue?
Deep Blue is a supercomputer created by IBM. It all started when a group of students at Carnegie Mellon University wanted to make a computer that could play chess.
The machine, which was named ChipTest, was designed to beat the best in chess and it sparred against Kasparov in 1989. The result was a heavy blow to the students as Kasparov defeated the machine with relative ease.
The dream of a machine defeating a man in chess was thought to be unattainable.
But IBM was closely monitoring the developments in the field. They were impressed by the development that a group of students achieved.
Deep Blue – the machine
Deep Blue was based on the design of RISC System/6000 (RS/6000), which was a family of supercomputers made by IBM. These were the first supercomputers from IBM that utilized POWER and PowerPC based microprocessors.
The Deep Blue supercomputer that was used in 1997 had several modifications when compared to its base design. The modifications were all aimed at improving the processing power in chess.
It had 30 nodes with 480 special-purpose VLSI chess accelerator chips per each node and 120 Mhz P2SC microprocessor per each node.
The hardware improvements made Deep Blue a hybrid. It had normal processors and chess accelerator chips.
The normal chess moves and planning were handled by the supercomputer, while the complex tasks were computed by the chess accelerator chips.
The first meet - 1996
The modification of 1997 came as a result of the duel between the Deep Blue and Kasparov in 1996. In the first match which consisted of 6 games, Kasparov won 3 times, and Deep Blue won once.
Two matches ended in a draw. This meant that Kasparov ended up as the victor.
But even when Deep Blue lost the overall match, it wrote a new page in history. This was the first time in history where a chess-playing computer managed to defeat the world reigning chess grandmaster once.
Kasparov himself was stunned by the machine. He quoted, “I had played a lot of computers but had never experienced anything like this. I could feel — I could smell — a new kind of intelligence across the table.”
Since the computer showed that it could indeed perform, this gave the researchers at IBM hope that they could win the game again by making some adjustments to the Deep Blue.
The second coming
Between 1996 and 1997, Deep Blue was tweaked in several ways. The speed of the whole system was doubled with new hardware choices.
The chess accelerator chips were improved by helping them understand the chess positions in more detail.
With the improvements, the computer was able to evaluate the chess positions and strategies very easily. The newfound processing power gave the machine the ability to sift through 100 million to 200 million positions per second.
One of the significant elements that boosted the whole team was the inclusion of grandmasters who were involved in this development, especially Joel Benjamin.
The grandmasters helped the team to perfect their opening library. The opening moves in chess have great importance because it lays the foundation for the strategy. Deep Blue would then spar against the grandmasters, which helped the whole team pinpoint the kinks in the system.
The match of 1997
In 1997, Deep Blue and Kasparov met again at New York City. Out of the six games, two games were won by Deep Blue and one by Kasparov.
Three games ended in a draw. This meant that the final scores where Deep Blue - 3½ and Kasparov - 2½. You can review the moves here.
This was the first-ever chess match won by a machine against a human.
Deep Blue instantly became famous as newspapers and other media raved about how a machine caught up to a grandmaster. The final move from Deep Blue made Kasparov visibly distressed. The match ended with Kasparov walking away, forfeiting the game.
And the world of chess was in shock as the machine rose above a human. Kasparov was hurt by the defeat, and he claimed IBM cheated.
Kasparov said that the sacrificial move made by the computer in the first or second game wasn’t at all machine-like, but felt as if it had human aid.
Was it a bug that led to Deep Blue’s victory?
A 2012 interview with Murray Campbell brought up some surprising information as he stated that the sacrificial move wasn’t a calculated decision, but a bug where Deep Blue didn’t know which move to make, so it made one randomly.
The researchers soon tried to fix the error between matches, but Kasparov didn’t know that it was a bug. He thought it was a different strategy, and it messed up his moves.
IBM denied Kasparov’s later requests for a rematch.
So has the machine triumphed by mistake? Or if the bug wasn’t present, would the outcome change?
We might never know. With modern tech, AI is undoubtedly getting powerful, and their best advantage is that they do not get tired.
We humans can only stay sharp for a brief amount of time till fatigue kicks in. But these matches show how a machine can rise up to the levels of human planning and foresight.
Are these machines capable of taking over the world? Kasparov says that instead of fearing these intelligent machines, we should work with them if we want to get the most out of technology.