Tech moves fast! Stay ahead of the curve with Techopedia!
Join nearly 200,000 subscribers who receive actionable tech insights from Techopedia.
AlphaGo is a narrow AI, a computer program developed by Google DeepMind to play Go, a Chinese strategy board game for two players similar to chess. AlphaGo is the very first AI program that was able to beat a professional human player, 2-dan player Fan Hui in October 2015, on a full-sized board with no handicaps. It then beat one of the highest ranked human players in the world, 9-dan Lee Sedol, in March 2016, winning four games out of five.
The AlphaGo project was started in 2014 as a test-bed in order to see how well Google DeepMind's neural network algorithm utilizing deep learning could compete at Go. The algorithm for AlphaGo is a combination of tree search and machine learning techniques and reinforced with extensive training with both humans and other computer players. It uses the Monte Carlo tree search and is guided by a policy and value network, implemented using deep neural network technologies. The policy network is trained and helps the AI predict the next move most likely to win while the value network is trained to narrow down the search tree and determine the value of those positions, estimating the winners in each position rather than searching all the way down to the end of the game.
AlphaGo was first fed with historical match moves from human players, utilizing a database of around 30 million moves, making it mimic human plays. Once the AI reached a degree of proficiency, it was trained further by making it play against instances of itself, using reinforcement learning to improve and learn more.
In October 2015, a distributed computing version of AlphaGo played and defeated Fan Hui, a 2-dan European Go Champion, marking the first time ever that a computer program had beaten a professional player at Go. Fan Hui then helped as a consultant for the DeepMind team months after his defeat. In March of 2016, AlphaGo went up against Lee Sedol, one of the highest ranked players in the world, having achieved the top level of 9-dan. Winning four games to Lee's one, this marked a major breakthrough in AI research since this meant that the deep learning and neural networks algorithm used by DeepMind can be used for any other purpose since it was not really programmed to play Go, but rather was taught how to play Go. This opens up a whole new world for AI research.