Google DeepMind’s AI system, AlphaGo, has received the primary of three matches it’s taking part in towards the world’s primary Go participant, Ke Jie.
It follows its historic win towards Lee Se-dol final yr, described by consultants as a breakthrough second for AI.
The AI received by simply half a degree in its newest match.
Ke Jie described the AI as “like a god of Go gamers”, whereas DeepMind co-founder Demis Hassabis thanked him for a hard-fought match.
“It was a such shut recreation, an thrilling recreation and confirmed how a lot work Ke Jie put into getting ready for the match,” mentioned Mr Hassabis in a post-match press convention.
“It was fascinating for us to see him utilizing strikes from AlphaGo’s earlier video games, and we have been intrigued to see how AlphaGo offers with its personal methods used – large respect to Ke Jie for pushing AlphaGo to its limits.”
He added that the final word plan for AlphaGo was a wider deployment “in areas of drugs and science”.
Of the match, Ke Jie mentioned: “There have been some sudden strikes and I used to be deeply impressed.
“I used to be fairly shocked as there was a transfer that will by no means occur in a human-to-human Go match.”
He added that AlphaGo was a “fantastic participant” and “utterly totally different” from final yr.
AlphaGo now has large affect on the Go circuit, and was star flip on the Way forward for Go Summit, organised by Google, in China this week.
The second recreation will probably be happen on Thursday, with the ultimate match on Saturday.
In Go, gamers take turns putting stones on a 19-by-19 grid, competing to take management of probably the most territory.
It’s thought-about to be one of many world’s most complicated video games, and is far more difficult for computer systems than chess.
AlphaGo has constructed up its experience by finding out older matches and taking part in hundreds of video games towards itself.