The Google DeepMind team has completed testing an algorithm designed to play a popular game of go. AlphaGo algorithm testing began last year, and the first game with a professional player took place in the spring of 2016. Then the program played with Lee Sedol, a famous player in go. The result of the series of matches was the victory of the algorithm, which was able to win a man four matches out of five. Not bad, given that the development of DeepMind before this did not play with players of this level. But still the program had something to strive for. And AlphaGo dealt a decisive blow to the self-esteem of professionals, proving that it can play better people.
Having defeated Sedol, the algorithm hid, about it for a long time there was almost nothing to hear, but on December 29 last year, a mysterious player with the nickname Master, who in just a few days won 60 wins, visited the popular online playgrounds for go players, not losing. In a series of matches, one draw was recorded, counted as due to the fact that the opponent of the Master could not stand the pressure and left the game. Among the people who fought with the Master, there were eminent and authoritative players having weight in the community of fans of the game.
Excited to share an update on !!
- Demis Hassabis (@demishassabis)
Yesterday, developers from DeepMind confessed that under the name Master played an improved version of AlphaGo. Among the defeated players there was also the Chinese champion on go who praised the algorithm and noted that the game is still a big field for research.
Go game was invented in Ancient China, but, despite the fact that it is already several thousand years old, until recently it was widely known only in East Asia. World recognition it received only in the twentieth century.
Go is a desktop logical and strategic game, the purpose of which is to beat the opponent, taking up most of the playing field.
The article is based on materials .
Comments
Post a Comment