​AlphaGo returns upgraded, wins 60 straight games online

DeepMind's AI has been playing as 'Master' online on a Chinese gaming platform and won 60 straight games by defeating a litany of world champions.
Written by Cho Mu-Hyun, Contributing Writer

A mysterious AI player who has been defeating world top Go players for over 60 straight games online has been revealed as an upgraded AlphaGo.

Demis Hassabis, CEO of DeepMind which created the AI, confirmed the identity of "Master" who won straight victories on Chinese gaming platforms Tygem and FoxGo.

Hassabis said games were played to test the new prototype version of AlphaGo. "Now that our unofficial testing is complete, we're looking forward to playing some official, full-length games there this year," he said.

AlphaGo has been playing fast-paced games online since December 29 and ended its "test" matches in January 4.

During that time, it defeated China's top player Ke Jie, South Korean Park Jung-hwan, and Iyama Yuta of Japan, as well as a litany of world-renowned champions.

Chinese champions reacted strongly. Ke Jie said human players needed to learn computer patterns to improve, while world number one Gu Li offered 100,000 yuan to whoever can beat AlphaGo to protect the "cultural treasure" invented by ancestors.

AlphaGo gained global notice in March when it defeated South Korean Go champion Lee Se-dol 4-1 in a live match in Seoul.

The victory has been compared to IBM's Deep Blue victory against chess champion Garry Kasparov in 1997.

Go is an ancient Chinese board game where two players with white and black pieces try to gain more territory than the other. More pieces within game time results in victory. It has substantially more moves compared to chess and was considered difficult for computers to adopt.

Editorial standards