For more than twenty years, tech experts and major digital intelligence programs have been striving to teach a computer program to play the ancient game of Go. Teaching a program to play a strategy game may seem frivolous, but it’s actually a major milestone, and it’s a big deal that Google’s team managed to pull it off.
Go has long been considered to be the most difficult game for an artificial intelligence system to learn how to play successfully. Nobody’s managed to do it in almost two decades. So it’s significant that Google successfully taught an A.I. to play the game. It’s even more significant that they weren’t the only ones in the race to reach that technological milestone.
Just one day before Google announced that they’d done it, Facebook CEO Mark Zuckerberg had posted online that the company’s tech team was “getting close” to having an artificial intelligence be able to play Go. Here’s his full post about the ongoing project, released before the news of Google’s achievement was made public:
“The ancient Chinese game of Go is one of the last games where the best human players can still beat the best artificial intelligence players. Last year, the Facebook AI Research team started creating an AI that can learn to play Go. Scientists have been trying to teach computers to win at Go for 20 years. We’re getting close, and in the past six months we’ve built an AI that can make moves in as fast as 0.1 seconds and still be as good as previous systems that took years to build. Our AI combines a search-based approach that models every possible move as the game progresses along with a pattern matching system built by our computer vision team.The researcher who works on this, Yuandong Tian, sits about 20 feet from my desk. I love having our AI team right near me so I can learn from what they’re working on.You can learn more about this research here: http://arxiv.org/abs/1511.06410”
Posted by Mark Zuckerberg on Tuesday, January 26, 2016
It was a close tech race, but Google pulled out ahead.
Go is an ancient Chinese strategy game, dating back approximately 2,500 years. The game takes place on a 19 x 19 grid. The two players place black or white stones on the board with the goal of surrounding the opponent’s pieces, which are then considered as “captured.” Ultimately, the player needs to control at least 50% of the board in order to win.
Before you start to compare it to chess (which artificial intelligence programs have been able to beat humans at since their infancy) you should know that chess has 10 to the power of 60 possible scenarios for the computer to calculate. Now compare that to Go, which has 10 to the power of 700 potential scenarios. It’s difficult enough for humans to play, who have the advantage of years of practice. So for an artificial intelligence to be able to play and beat a human at Go is an impressive feat.
Google’s accomplishment opens up a new expanse of technological possibilities. Being able to play Go is a demonstration that their artificial intelligence is now able to process and search through a greater number of potential scenarios and series of actions than ever before. In short, the A.I. is now more capable of working through a more complex number of logical steps and is able to reason close to on par with the humans that developed it.