Tuesday, March 15, 2016

Go, Google AlphaGo beats the champion Sedol 4 to 1 – PianetaCellulare.it

In the end he won the supercomputer AlphaGo Google Deepmind, who beat South Korean Lee SEDOL to GO.

The world champion of Go – an ancient but complicated game table – was defeated by the latest development in computer technology in the Google labs in four games – five in total – with the aim to help determine whether a machine can imitate the human intuition.

On March 9 in Seoul, Lee Sedol, a professional Go player to South Korea to rank 9-dan of South Korea, has surrendered during the first game “Go” against a Google AlphaGo owned computer program. The second and third challenge had the same outcome, and it is on Google in the hope that his computer could beat the champion with a nice 5-0. The fourth challenge , however, was won by Sedol sample, and when you stop at a score of 3-1 for AlphaGo. The fifth and final match between AlphaGo and Sedol which was held Tuesday, March 15, broadcast live streaming on YouTube, came out on top supercomputer AlphaGo Google Deepmind.

FINAL RESULT 4 games won for Google Deepmind machine, 1 victory for the South Korean Lee SEDOL.

Go is a strategic board game for two players known as Weiqi in Chinese, or go igo in Japanese, and baduk Korean who plays with white and black round stones on a grid of 19×19 goban called boxes. The go originated in China, where he played for at least 2500 years, and is very popular in East Asia, but has spread around the world in recent years. Here in Italy it is not very popular. Go is a complex game strategically though its rules are simple (read more ‘below as you play). There is also a Korean proverb that no game of Go has never been played twice, which is likely if we consider that there are 2,08×10 170 different positions possible on a goban.

AlphaGo vs. the Go world champion. “I never thought that AlphaGo could play so well,” said the world champion of Go game – Lee Sedol, who has 33 years – adding that he is very surprised that AlphaGo made “moves that would not have been possible to do to a human being.”. The founder of DeepMind Google David Silver Division said Sedol “brought AlphaGo extremes of its capabilities.”

The big challenge for AlphaGo creators was to establish a set of rules that tell the computer which position is winning, according Demis Hassabis, CEO of Google, which has acquired in recent months DeepMind, a company specializing in artificial intelligence whose technology helped create intelligent software that can play Go.

“Go is primarily a game of intuition rather than calculating how the game of chess,” said Hassabis at a press conference, as reported by NBC News. While Go is “the most stylish game ever invented by humans” is also “deeply complex,” he added. “There are more possible positions in Go than atoms in the universe,” said Hassabis.

Other human-computer challenges . The challenge of 9 March between the Go world champion against a program for Google AlphaGo owned computers was not the first time that computers have defied a human mind. In 1997, the Deep Blue computer manufactured by IBM specifically designed to play chess won in a series of six games against the World Champion Gary Kasparov at the time – this challenge was seen as an important milestone in the development of artificial intelligence .

AlphaGo also defeated the European champion in charge Fan Hui 5-0 in a Go tournament in October 2015. This was the first time that a computer program beat a human professional player Go. Speaking before the first game, Lee said that he felt “strange” in the thinking of playing against a computer, but has not given up the challenge. “It ‘s different to prepare for a game against a non-person,” he said. “When I prepare for a game against a person, it is important for me to read the energy of that person, which is not possible with a computer”

Google’s ultimate goal with these AI software does not it is to create computers that can beat humans at table games, but the Mountain View giant wants to fine-tune the AI ​​to be applied in areas such as robotics for health care and care for the elderly.

How to play Go . If you are intrigued by how to play Go, the gameplay involves two players who alternately place black and white pieces (called stones) on empty intersections of a “checkerboard” (called goban) equipped with a 19×19 grid boxes. The aim of the game is to control a larger area than that goban controlled by the opponent; to succeed, the players try to have their stones in positions where they can not be taken by the opponent, taking at the same time of the territories that the opponent can not ‘invade without being caught. In fact, when a stone or a group of enemy stones is completely surround with their stones, the opponent’s stones are ‘captured’. Players seek both to build a defense for its stones that create a barrier to surround the opponent’s stones and then steal them. The game ends when a player runs out of opponent’s pieces, but when players agree that neither of them can expand their territory or decrease the opponent.

Share this news
LikeTweet

No comments:

Post a Comment