Google's AlphaGo artificial intelligence defeats the world's number one chess player

Add an extremely close victory of artificial intelligence from Google before the world leading chess players.

According to Engadget, Google 's " Go " star , AlphaGo , took a 1-0 lead in the match against Ke Jie - the world' s number one go player and who has held this position since at the age of 17 - with only half the difference (minimum gap in the Go flag).

Historical match between human and artificial intelligence in China.(photo: DeepMind).

After the match, Demis Hassabis, Google CEO DeepMind explained how they programed Alpha Go: maximizing the winning rate, not the winning number (the item is the amount of land earned in the game and Whoever has the higher number of items will be the winner. This latest AlphaGo version, nicknamed Master, has 10 times better computational performance than the old version - the version that defeated Lee Sedol, a 9-player Korean chess player, ranked second. gender, and the Master operates from a computer connected to Google's cloud server.

In order to help those who do not know the rules, we would like to report the summary of the match as follows: Google AI has gained an advantage of 10-15 points right from the beginning of the game, making The opportunities for Jie to respond significantly reduced. Although Jie also has the upper hand, AlphaGo has quickly regained the advantage, making sure that the more people his opponent is at the end of the match the less chance of victory.

AlphaGo and Ke Jie match reports.

According to Jie, AlphaGo has a lot of subtle steps: AlphaGo's 24th move has shown a very high level strategy, completely changing the face of the game. Interestingly, this super player even uses some moves and tactics of AlphaGo.

After the game, the DeepMind team also shared that AlphaGo is programmed to win, but other versions of this AI can be created for many different purposes, including "expressing the difference in qualifications " - win with the maximum possible difference. Mr. Hassabis said: "We want AlphaGo to be a tool for the Go community to improve our skills. We will reveal some details about the strategy that AlphaGo used when fighting itself. We want this AI to be widely used in science and medicine to help experts quickly achieve more breakthroughs, and the team will have a lot of work ahead in the coming years. ".

AlphaGo is programmed to win.

The second game will take place this Thursday (May 25). In the event that Ke Jie can win this game, the decision is expected to take place on Saturday (May 27). Machines are tireless, Master will continue to participate in two more matches. One match, it will serve as a partner of one of the two professional Chinese players fighting each other. In the other match, it will challenge five professional players at the same time (and almost certainly to show off).