Go is an ancient Chinese strategy game that has been a bulwark of human intelligence when comparing ourselves to artificial intelligence. This week, however, Google’s DeepMind AlphaGo defeated the world’s top Go player, Ke Jie. by half a point, which is the closest possible margin for a game to end with.
Though the game was not allowed to be streamed live in china, the broadcast garnered more than 200-million viewers.
What is Go?
Go is an ancient Chinese strategy board game. Players make use of white or black stones to occupy spaces on a 19-by-19 square. The player who controls the most territory at the end of the game is considered the winner. The other win condition is eliminating all of the opponent’s pieces from the board.
What does this mean for advancements in AI?
Many people expected artificial intelligence to be decades away from conquering the world’s top Go players, but on May 23, 2017, the number one Chinese Go player was beaten by half of a point. This signals that a game based on intuition with a nearly unlimited number of possible moves has fallen to artificial intelligence.
While this breakthrough in and of itself may seem significant only on a minor level, it clearly demonstrates the ability of machines to utilize intuition or a similar factor to overcome even the most skilled players.
How does this impact the future of AI?
With proof that one of the most complex games in existence has seen its current top player conquered by artificial intelligence, it is becoming clear that artificial intelligence progress is accelerating at an unexpectedly quick pace. Researchers expected it to take several more decades for AI to surpass Go champions.
Go, one of the last bulwark games representing human superiority in terms of intelligence and intuition saw its number one Chinese player, Ke Jie fall to AlphaGo in a broadcast that was banned from airing live in China. The victory came at the closest margin possible, one half of a point. The close victory was a victory nonetheless.
The victory by AlphaGo puts artificial intelligence more than a decade ahead of even the most conservative estimates. Last year’s victory in South Korea came with a handicap in the form of four cornerstones for the AI competitor. The May 23 victory, however, was opened on equal footing, a clear board.
Artificial intelligence has advanced decades in a single year according to the estimates of many AI researchers and expectations of many Go players at the highest level. If Moore’s law applies t artificial intelligence with respect to exponential progression per two years, the question is, where will we find ourselves within the next ten years? Let us know what you expect in the comments section below!