Top Go player beaten by Google's latest DeepMind AI

AI researchers at Google DeepMind are making waves this week with claims they've developed an artificial intelligence, AlphaGo, which can beat expert Go players at their own (ancient board) game.

Anyone with an interest in video game AI may appreciate this news, as DeepMind researchers claim -- in a paper published in the science journal Nature this week -- to have designed AlphaGo to mimic expert players using "deep neural networks" and to improve by learning from games it plays against itself, ultimately enabling it to beat European Go champion Fan Hui 5-0. 

In March AlphaGo will play a 5-game match against the world's top-ranked Go player, Lee Sedol, but Google is already celebrating its victory over Hui as one of the most significant breakthroughs in AIs beating humans at games since IBM's Deep Blue computer beat world champion Garry Kasparov at chess in the late '90s. 

"Since then, the...holy grail, if you like, has been Go," DeepMind CEO Demis Hassibis stated in a video {embedded above} published today by Google. "Go is a much more intuitive game, while chess is a much more logic-based game. AlphaGo is our program to actually try and crack Go." 

Of course, longtime game developers may recall that before he was building Go-bots, Hassibis was a game developer -- one who co-designed (with Peter Molyneux) and programmed Bullfrog's Theme Park when he was 17 years old.

Hassibis went on to found Elixir Studios in 1998, where he helped create games like Evil Genius before eventually leaving the industry. In 2012 he founded DeepMind Technologies, which Google bought in 2014 for an estimated $400 million.