Facebook wants to use artificial intelligence to crack the game no computer has ever mastered

GoGoban1The game of Go.

There’s one surefire way researchers like to test the capabilities of their artificial intelligence systems — games.

Earlier this year, Google Deepmind created a computer that can learn how to play video games with no instructions. In 30 minutes, the computer became the best Space Invader player in the world.

And most are familiar with the time when IBM’s Deep Blue computer beat world chess champion Garry Kasparov in 1996.

But artificial intelligence has yet to tackle the game Go —  an Eastern, two-player game that has more than 300 times the number of plays as chess.

Facebook, though, is now attempting to do what other computers have thus far failed at and beat a human at Go for the first time ever, Wired first reported. 

Go is a truly difficult game to learn that requires a lot of thoughtful planning (I attempted to master it junior year of college and have yet to revisit that quest). The game is played on a 19 X 19-line grid. The players are given a set of black and white stones and must attempt to cover a larger surface area of the board than their opponent. 

Where it gets tricky is you can play in patterns that allow you to take over your opponents stones and claim territory. To be successful, you have to envision what kind of moves your opponent is trying to make, just like when playing Chess. 

This is why Facebook researchers are teaching their artificial intelligence system to recognise visual patterns.

Facebook notes on their research page that using games to train machines is a common approach, and that the difficulty of Go will help refine Facebook’s artificial intelligence so it’s capable of sophisticated pattern recognition.

Think of it this way: after the first two moves of a chess game, there are 400 possible next moves. In Go, there are close to 130,000.

“We’ve been working on our Go player for only a few months, but it’s already on par with the other AI-powered systems that have been published, and it’s already as good as a very strong human player,” Mike Schroepfer, CTO of Facebook, wrote on the research page.

Here’s Facebook’s artificial intelligence bot playing Go:


To improve the system, researchers modelled out each possible move as the game progress (which is a traditional search-based approach) and combined that with a pattern matching system built by their computer vision team.

We’ve interacted with forms of Facebook’s artificial intelligence, perhaps unknowingly. When Facebook recommends you tag a specific friend in a photo, it’s using artificial intelligence for facial recognition.

So why develop a bot capable of playing Go? Well, Facebook’s research page notes that it stems back to their development of Facebook’s personal assistant, M. Strong pattern recognition will be necessary for M to complete tasks like making purchases.

Apple is also working on different artificial intelligence projects, though the company is hush when it comes to its projects. So hush that people who are hired by Apple to work on artificial intelligence are told not to announce the job on social media.

But Apple acquired two artificial intelligence companies in October, indicating that something is brewing.

Google acquired two artificial intelligence companies in 2014 and has been working on a host of projects. Recently, the company announced that it has developed artificial intelligence capable of writing your emails for you, and is beginning to roll out the feature.

NOW WATCH: Briefing videos

Business Insider Emails & Alerts

Site highlights each day to your inbox.

Follow Business Insider Australia on Facebook, Twitter, LinkedIn, and Instagram.