Google's artificial intelligence (AI) program DeepMind will take on the best Go player on the planet for a $1 million prize. Its Alpha Go AI system will challenge world champion Lee Sedol from South Korea in a man vs. machine contest from March 9 to 15 during a live YouTube broadcast from Seoul.
DeepMind's CEO and founder Dennis Hassabis described Lee as a legend. He stated it was an honor for Google AI to take on the world's best Go player for the past decade.
Google AI was only able to beat super low-level players until it defeated three-time European champion Fan Hui recently. The victory over the French Go player was the first time an AI defeated a human professional player without playing with a handicap.
Google's AI defeated the French Go champion 5-0. The Alphabet company did not publicize the event but it was later featured in the scientific journal Nature, according to VentureBeat.
AI machines play Go differently than chess. Due to the almost unlimited number of possible moves DeepMind does not try to calculate the opponent's next plays.
Instead DeepMind's programmers taught it the strategies of Go masters. It then played several virtual games to reinforce them through machine learning, according to Engadget.
Hui's Go ranking is a 2 dan player. Dan is a ranking system first used during Japan's Edo period.
However, Lee is ranked as an ultra-high 9 dan. Lee shared that he heard DeepMind's AI is a strong player but he is confident he can beat the supercomputer.
However, Lee pointed out that he can certainly win this time. Google's AI continues to improve and experts first believed it would take another decade for computers to defeat master Go players.
In the past other supercomputers have taken on world champions of other games. IBM's Deep Blue became the first AI machine to win a chess game against reigning world champion Gary Kasparov in 1996.
Then IBM's Watson supercomputer defeated two former Jeopardy champions in 2011.During the game it was not connected to the Internet but had access to 200 million pages of content.
This video explains DeepMind's AlphaGo system: