AlphaGo won the first game from Lee Sedol

Google’s AI was stronger than one of the best go players.

Artificial intelligence can demonstrate its superiority over man in the form of winning a logical game. Such games are divided into those in which the AI ​​is already playing stronger than the best of the people, and those that have not yet succumbed. Today, as a result of the first of five games, the Asian game of go has become one step closer to falling into the first category.

DeepMind, a UK-based artificial intelligence company, was acquired by Google in 2014. DeepMind has already attracted media attention by developing the DQN system, which has independently learned how to play 49 old Atari games. For the past two years, a group of company employees has been working on AlphaGo. It is an artificial intelligence system for the game of go, which shows excellent results. It achieves this with a combination of the Monte Carlo method and neural networks of policy and value. Neural networks of the program were taught on 160 thousand batches from the KGS server.

The rules of go may vary, but their basis is about the same. Two players place stones of two colors on a board of a certain size, the standard field is 19 × 19 lines. The goal of the game is simple: you need to fence off the territory of a larger size than your opponent’s size on the board with stones of your own color.

Go is considered difficult for computer systems to play. So it turns out because of the huge search space and the complexity of the choice of moves. There are more than one googol (10 ¹⁰⁰ ) times more possible positions of stones on a standard board than in chess. The number of possible positions is greater than the atoms in the universe. Existing programs can play go at amateur level. For a time, it was believed that before the appearance of systems, it was stronger than man for another decade.
')
AlphaGo has demonstrated that this is not the case. The program easily beats not only other products, but also a person. Three-time European champion Fan Hui played against AI. The result of five games hit: in all of them the second professional dan player lost to the computer system . Later, Google's choice was criticized : the level of go ownership in Europe was not very high, and the mistakes of a representative of a team of people raised questions. But Google immediately said that AlphaGo will play against legendary Korean go-pro Lee Sedol, the best go player in the last decade, the second in the number of titles. Obviously, his level of play must be higher.


Today in Seoul was the first of five games. She attracted attention, which is not surprising for Korea, where tens of millions play gogo. The game was visited by two members of the Korean parliament and the former head of Google, Eric Schmidt. AlphaGo operator Aja Huan is located opposite Lee Sedol. By the way, the researchers did not sit on the spot after the match against Fan Hui in October last year and improved AlphaGo. For the past five months, the AI ​​has been studying in batches against itself.


The man played black and went first, and the AI ​​white. The played start did not differ much, it was even a bit like Fan played in Britain. Sadol demonstrated his typical style and made aggressive moves. But AlphaGo was not holding back - the system began to attack in about 12 moves. According to commentators, AlphaGo plays like a human: the system searches for and creates vulnerable groups of stones to make strong moves. Although the AI ​​made some mistakes, blacks had good reasons for concern. The game of 186 moves for 3.5 hours ended with the ninth dan master admitting defeat.

here's the amazing victory over lee se-dol https://t.co/16QOqOL467 pic.twitter.com/4x31ppMDge

- sam byford (@ 345triangle) 9 March 2016

It’s still early to talk about the historical moment. The last fifth game will take place on March 15th. Then the fate of the prize million dollars will be determined - either it will go to charity, or a human player will receive it. Next Tuesday it will be clear if the AI ​​has conquered the new game or it’s too early to talk about it. The saddle is full of determination: he believes that the opening moves better will allow to achieve greater chances of victory.

All games are broadcast on YouTube .