News classification
Contact us
- Add: No. 9, North Fourth Ring Road, Haidian District, Beijing. It mainly includes face recognition, living detection, ID card recognition, bank card recognition, business card recognition, license plate recognition, OCR recognition, and intelligent recognition technology.
- Tel: 13146317170 廖经理
- Fax:
- Email: 398017534@qq.com
Artificial intelligence: the advent of AlphaZero 8 hours after blasting chess, chess, Shogi
Artificial intelligence: the advent of AlphaZero 8 hours after blasting chess, chess, Shogi
It is understood that the AlphaZero algorithm can beat Li Shishi version of the AlphaGo in 8 hours and 4 hours of training; training to beat the world's top chess program of Stockfish; then 2 hours of training to beat the world's top chess program Elmo. This is another new algorithm that the DeepMind team has brought to us after the advent of AlphaGo Zero research, which is "a more general version".
In addition, we see that there are several differences between AlphaZero and AlphaGo Zero. First of all, AlphaGo Zero estimates and optimizes the winning probability under the assumption that the result is win / lose two yuan. And AlphaZero will take the level or other potential results into consideration, and estimate and optimize the results. Second, AlphaGo and AlphaGo Zero will change the chessboard position for data enhancement, while AlphaZero does not. Third, AlphaZero only maintains a single neural network. This neural network is constantly updated instead of waiting for iteration. In four, AlphaZero, all games are repeated using the same super parameters, so there is no need to adjust for a specific game.
Statistics show that Alfa AlphaGo is the first artificial intelligence program to defeat the professional go players and the first champion to win the go world. It is developed by the team led by Google (Google)'s DeepMind company, led by Demes Hassabis. The main principle is "deep learning".
In early March 2016, Alfa and Weiqi chess world champion, nine players Li Shishi occupation chess machine war, with a total score of 4 to 1 victory; at the end of 2016 at the beginning of 2017, the program in the China chess website to "master" (Master) is a registered account number ten and the three chess master of chess showdown, continuous 60 no losses; in May 2017, China Wuzhen go summit, and it ranked first in the world chess champion of the world kija war, with a total score of 3 to 0. Weiqi is widely recognized as the top player of Alfa Weiqi. The ranking of the world professional go ranking announced by GoRatings website has surpassed that of chess player, the first ranking player in the world.
In May 27, 2017, after the man-machine war of kje and Alfa, the Alfa go team announced that Alfa would no longer participate in the game of weiqi.
In October 18, 2017, the DeepMind team published the strongest version of AlphaGo, the code name AlphaGo Zero.
This time, in just two months, the strongest version of AlphaGo Zero has evolved and turned into AlphaZero.
In addition, we see that there are several differences between AlphaZero and AlphaGo Zero. First of all, AlphaGo Zero estimates and optimizes the winning probability under the assumption that the result is win / lose two yuan. And AlphaZero will take the level or other potential results into consideration, and estimate and optimize the results. Second, AlphaGo and AlphaGo Zero will change the chessboard position for data enhancement, while AlphaZero does not. Third, AlphaZero only maintains a single neural network. This neural network is constantly updated instead of waiting for iteration. In four, AlphaZero, all games are repeated using the same super parameters, so there is no need to adjust for a specific game.
Statistics show that Alfa AlphaGo is the first artificial intelligence program to defeat the professional go players and the first champion to win the go world. It is developed by the team led by Google (Google)'s DeepMind company, led by Demes Hassabis. The main principle is "deep learning".
In early March 2016, Alfa and Weiqi chess world champion, nine players Li Shishi occupation chess machine war, with a total score of 4 to 1 victory; at the end of 2016 at the beginning of 2017, the program in the China chess website to "master" (Master) is a registered account number ten and the three chess master of chess showdown, continuous 60 no losses; in May 2017, China Wuzhen go summit, and it ranked first in the world chess champion of the world kija war, with a total score of 3 to 0. Weiqi is widely recognized as the top player of Alfa Weiqi. The ranking of the world professional go ranking announced by GoRatings website has surpassed that of chess player, the first ranking player in the world.
In May 27, 2017, after the man-machine war of kje and Alfa, the Alfa go team announced that Alfa would no longer participate in the game of weiqi.
In October 18, 2017, the DeepMind team published the strongest version of AlphaGo, the code name AlphaGo Zero.
This time, in just two months, the strongest version of AlphaGo Zero has evolved and turned into AlphaZero.