Recently a new player in the Chinese boardgame Go has been making waves across the web due to its unusual play style. Implementing strategies that set opponents off-balance with rapid, unorthodox moves, Go players began to speculate the player was a computer controlled AI. These guesses were confirmed when DeepMind CEO Demis Hassabis tweeted an update on the company’s AlphaGo AI. The tweet explained that AlphaGo had been unleashed upon Go players residing on the Tygem and FoxGo servers under the aliases “Magister” and “Master.”
The entirely computer controlled AI was able to rack up an impressive 60 wins and no losses, including the defeat of the reigning champion Ke Jie. No human player was able to best the AI, with only a handful of “Ties” forced when the human opponent went offline, resulting in the match defaulting to a tie.
Ke Jie admitted humans were no match for the AI after losing three games to AlphaGo.
This isn’t the first time AlphaGo has been tested in the field. Last year it bested the world-champion Go player Lee Sedol. Go is a Chinese boardgame that many thought couldn’t be mastered by a computer AI due to the complexities of its mechanics and strategy. The Go community was shocked last year when AlphaGo handily defeated a world-champion.
This was only a test for AlphaGo, and the company plans to further the project with publicized matches against other top-tier players in the game. The DeepMind CEO was gracious about the AI’s victories, stating “We’re excited by the results and also by what we and the Go community can learn from some of the innovative and successful moves played by the new version of AlphaGo.”
Acquired by Google in 2014 and based in London, Deep Mind is a leader in machine-intelligence research and their AlphaGo project aims to improve advanced computer-controlled AI.