Sihai network

Koji vs. alpha dog time koji vs. alpha dog live address

4hw.org: the most popular topic recently is Corzine's battle with alpha dog. At present, the first game is over. Alpha dog wins by a quarter. Although Corzine's victory is very small, the century war of go will be remembered.

There are three games for Ke Jie to fight against alpha dog. Let's take a look at the time schedule of the match, but let's watch the video address!

On May 23, the first Bureau of man-machine war started in Tongxiang, Zhejiang Province. After guessing, Ke Jie took charge of the black first. In the end, alphago, an AI developed by deepmind team, beat Ke Jiejiu, the first Chinese player in the world in terms of level score, 1 / 4 in white, temporarily leading by 1-0. The second game of the two sides will continue at 10:30 on the 25th.

Hassabis, founder of deepmind: no matter how successful the result is, it belongs to human beings

In the speech before the game, hassabis, the founder of deepmind, said that he did not agree that it was a man-machine competition, but that people use computers to discover new knowledge. Whatever the outcome, the ultimate victory belongs to mankind.

The beauty of go, he said, is that we explore new ways of playing on the go board just as we explore the universe. Go is also a great challenge to AI, with many changes and mysteries. Last year's competition showed not only AI's creativity, but also human's creativity.

He doesn't agree that it's a man-machine competition. It's not any computer competition. It's people using computers to find new knowledge. Like the Hubble telescope, we explore the universe with it. And alphago is like Hubble, and we explore go with it.

"We believe that we can make unprecedented achievements with artificial intelligence. AI can help domain experts to achieve greater breakthroughs at a faster speed. We hope alphago will continue to develop and explore new ways of playing go in this week's competition. Whatever the outcome, the ultimate victory belongs to mankind. '

The AI technology behind alphago deserves more attention

Alphago showed great strength before the game. Last year, Li Shishi, a Korean chess player, won 4:1 in the total score for the first time. At the beginning of this year, he won dozens of games in weiqi.com and wild fox weiqi.com under the pseudonym "master" and won a number of weiqi experts. Therefore, many people had not been optimistic about her performance.

In response, Ke Jie posted a long microblog on the eve of the game, saying that whether he won or lost, it was the last three sets of competition between him and AI. He believes that although AI develops very fast in the field of go, they have no passion for go, just cold machines. He will give a good account to the friends who love go and care about the game, so that everyone can 'watch and cherish'.

Compared with winning or losing, alphago's machine learning technology deserves more attention.

Alphago is a go AI developed by deepmind acquired by Google. Deepmind is committed to creating an AI that can 'learn' how to play games and reach a high level in the same way as humans. Their AI has learned to play a variety of games including "break out" before, and has reached the level of Superman.

Different from the traditional means of violent exhaustion, there are three 'secrets' to support alphago to defeat Li Zaishi: deep neural network, supervised / enhanced learning and Monte Carlo tree search. Relying on the above three 'weapons', alphago has become the best go AI produced by human beings.

Alphago had an upgrade before the game. According to Wang Xiaochuan, CEO of Sogou company, the upgraded alphago 2.0 gave up supervised learning and Monte Carlo tree search, and strengthened enhanced learning, making alphago more like a person. And deepmind founder hassabis hopes that they can test alphago's creativity and adaptability through competition.

At present, the AI developed by deepmind is not only excellent in playing go, but also has been used in practical applications. Deepmind's AI is already working for Google, helping Google reduce the cost of its computer room cooling system by 40%. They also hope to work with the national grid to reduce energy consumption in the UK by 10% using artificial intelligence.

What's more, alphago inspires people to rethink the relationship between machines and people. It tells people that there is still a great space for human development, and there may be other life evolution paths and different results.

At present, alphago and Ke Jie have only completed the first match, looking forward to the wonderful match. Whatever the final result, we have reason to applaud human innovation.