Facebook has been demonstrating how intelligent its AI engine is becoming, putting the system up against humans for the Chinese strategy game of Go and almost beating them every time.
The research, developed with Cornell University, is based upon predicting the next move of the human player and applying this knowledge to the AI engine.
"Go's high branching factor makes traditional search techniques ineffective, even on leading-edge hardware, and Go's evaluation function could change drastically with one stone change," the computer science project's abstract explains.
Although some methods have been developed to beat the humans in the game of Go, this new research sets out to use AI called Darkforest to better apply itself, it explains.
"Darkforest substantially improves the win rate for pattern-matching approaches against MCTS-based approaches, even with looser search budgets."
The demonstration of th artificial intelligence appeared on Facebook, where CEO Mark Zuckerberg posted a video of the robot working in real time against a human player.
"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," Zuckerberg in a blog.
"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."
It is being developed by two members of Facebook's team - Yuandong Tian and Yan Zhu who apparently sit right next to Zuckerberg in Facebook's HQ.
Google is also working on a solution to beat human players of Go and rumours suggest the AI engine is closer than Facebook's attempts.
