Google's AlphaGo AI Beats Human Go Champion
- Author: Toni Ryan May 24, 2017,
May 24, 2017, 5:31
Gu Li, a national Go champion who provided commentary on the AlphaGo-Lee match, said that he gives Ke only a 10 percent chance of winning even one of the three rounds.
Players take turns putting white or black stones on a rectangular grid with 361 intersections, trying to surround larger areas of the board while also capturing each other's pieces. During the competition, Ke looked as if he was feeling a mounting pressure, with strong expressions and fidgeting, when pondering his next moves against AlphaGo.
"We'll release some details of the architecture, of the games that AlphaGo plays against itself, later this week", said the DeepMind CEO. Hassabis said: "We want to use AlphaGo as a tool that the Go community can use to improve their games".
Google hopes to showcase the evolution of machine intelligence this week in Wuzhen, Demis Hassabis, chief executive officer and and co-founder of Google DeepMind, wrote in a recent blog post.
"I'm thinking that If Ke Jie were wining then they would be live streaming it".
DeepMind has further developed the architecture of the program since it defeated Lee a year ago.
The program defeated South Korean Go master Lee Se-dol 4-1 in March 2016.
The AI won by just half a point - the smallest possible margin of victory - in a match that lasted four hours and fifteen minutes.
"In the course of winning, AlphaGo somehow taught the world completely new knowledge about perhaps the most studied and contemplated game in history."
Following last week's I/O conference, where Google reaffirmed its position as an artificial intelligence company, the internet giant is once again drawing a spotlight on its AI-capability with AlphaGo now competing against 19-year-old world champion Ke Jie.
The ceremonial game - the second time AlphaGo has gone head-to-head with a master Go player in a public showdown - represents a major bridge-building exercise for Google in China, following a charm offensive in recent years.
This machine learning technique was not just an innovative way to teach AlphaGo how to play Go well, but it was quite necessary for AlphaGo's successful learning. Friday will have a match between two human players pitted against each other, both paired with AlphGo, and a team match where five players go against Google's AI. Ke mentioned that the gap between humans and A.I. was widening greatly, and AlphaGo was beginning to move away from using human teachers as inspiration, and starting to develop its own methods for playing the game. He noted that the number of possible board states in Go is "more than the number of atoms in the universe".More news: Macron wins Merkel backing for bid to shake up Europe