DeepMind’s AlphaGo program will test its artificial intelligence capabilities in May against top Go player Ke Jie.
The match of three games in Wuzhen, China between AlphaGo and the Chinese player comes about a year after the computer program beat by 4-1 a key player, the South Korean Lee Se-dol, in a game that is regarded as involving more complex strategy than even chess.
During the game, players take turns to place black or white pieces, called “stones,” on the 19-by-19 line grid. The aim is to capture the opponent's stones by surrounding them and encircling more empty space as territory. AlphaGo has been seen as a major contender because of its ability to learn from its experience, sometimes resulting in far-from-human but nevertheless successful moves.
After Lee’s defeat there were some game analysts who held that AlphaGo would find it far tougher to beat Ke.
Besides playing Ke, AlphaGo will also contest in games that involve both humans and artificial intelligence programs. In “Pair Go,” one Chinese professional player will play against another but they will both have their own AlphaGo teammate, alternating moves, “to take the concept of ‘learning together’ quite literally,” wrote DeepMind CEO Demis Hassabis in a blog post on Monday.
In an indication that DeepMind is now more confident about the capabilities of AlphaGo, the program will also take on in “Team Go” a five-player team consisting of China’s top pro players. The human players will be “working together to test AlphaGo’s creativity and adaptability to their combined style,” Hassabis said.
AlphaGo follows in the footsteps of the chess-playing Deep Blue computer that beat Garry Kasparov in 1997. Another IBM computer, Watson, won in 2011 in the Jeopardy quiz show.
Watson went on to become a technology that IBM has commercialized to offer natural language processing and machine learning for the analysis of unstructured data.
DeepMind, acquired by Google in 2014, is now part of the Alphabet group. It described AlphaGo at the South Korea match with Lee as still a prototype, requiring a lot of testing and training of the platform. The games that the program will be playing in Wuzhen are evidently aimed at providing AlphaGo with that exposure.
AlphaGo has already played Ke and other top players in unofficial online games, according to reports. "Now that our unofficial testing is complete, we are looking forward to playing official, full-length games later this year in collaboration with Go organisations and experts...," Hassabis wrote in a Twitter message in January.
Google has used machine learning methods behind AlphaGo to tackle problems such as reducing energy use. It has also used machine learning in many of its products, including Google Photos and Google Translate.
Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.