8 things you must know about AlphaGo vs. Ke Jie_1
8 things you must know about AlphaGo vs. Ke Jie
Artificial intelligence and human wisdom are confronting each other again. Google AlphaGO has handed Go World Champion Ke Jie a book of war. For this man-machine war, these 8 things you should know.Click to watch the game: Wuzhen Alphago and Ke Jie man-machine battle live broadcast address AlphaGo vs Ke Jie 8 things you must know 1. The time and place of the duel. From May 23 to 27, in China’s most prestigious water town and the permanent site of the World Internet Conference-Wuzhen, Tongxiang, Zhejiang, the Chinese Go Association and the Zhejiang Provincial Sports Bureau will work with Google to jointly host the Chinese Wuzhen Go Summit.The duel between artificial intelligence and human players will be unveiled as an important part of the summit, and Google AlphaGo will face off against Go Jie, the world champion Chinese chess player Ke Jie.2. How to set the competition system? Starting from the 23rd, five matches will be held around human-machine battles, all using Chinese rules. AlphaGo and Ke Jie’s 1vs1 three-match chess system, regardless of the winning and losing sides must play a full three games, each side will have 3 hours, after the end of the time will have 5 1 minute countdowns. In the man-machine match game, two Chinese players will team up with AlphaGo for 1 hour each side, and there will be a 1-minute countdown once the time is over. Players to play: Lian Lian Ba Duan, Gu Li Jiu Duan Team Match Five Chinese players will form a team, playing against AlphaGo, 2.5 hours.After the elapsed time, there will be 3 countdowns in 1 minute. Chess players: Zhi Yuting 9th, Shiyue 9th, Tang Weixing 9th, Zhou Ruiyang 9th, Chen Yaoye 9th.3. Arrangement of event bonus. According to the organizer’s disclosure, if Ke Jie wins, he will receive US$1.5 million and the appearance fee is US$300,000.4. In addition to man-machine battles, what else does Wuzhen Weiqi Summit have? Shi Bomeng, president of Google China, revealed that in addition to the manual duel between AlphaGo and Chinese players, on May 24, Google will host two artificial intelligence forums.The experts of Google DeepMind will discuss the development of AI and human-computer collaboration together with Chinese artificial intelligence experts.5. The preparation situation of both sides of the man-machine war. In terms of Google AlphaGo, Shi Bomeng revealed that the AlphaGo version has been continuously improved since the duel with Li Shishi, and is still under development. The version against Ke Jie will be the latest upgraded version and will not be provided to Ke before the start of the event.Jie Zuo warm-up training. As for Ke Jie, the head of the Chinese Go Association revealed that what he can do now is to create a quiet environment for Ke Jie to prepare for the game. As for the specific preparation details, it cannot be disclosed.6. Ke Jie was full of confidence when he confronted Li Shishi with AlphaGo. Is Ke Jie still confident? After AlphaGo swept Li Shishi with a score of 4:1 last March, recently, the upgraded version of AlphaGo, Master/Magister, swept the undefeated record in 60 online fast chess matches with the world’s top Go players.For the human-machine Go game, the strength of artificial intelligence is fully demonstrated.Ke Jie said in a media interview at the press conference that he used to think that AI chess is more computationally powerful, but through actual observations, it was found that AI is very powerful in setting up macro chess positions.I used to think that Go was a festival of human wisdom, and no machine could beat human chess players, but now AlphaGo has changed the entire ecological environment of chess players. Standing on the node of historical change, Ke Jie said that he would play his chess well without compromising. With a win-win attitude, he must defeat AlphaGo.7. In addition to playing chess, what does Google AlphaGo do? AlphaGo is actually a novel machine learning technology developed by Google, which combines the advantages of supervised learning and reinforcement learning.Through training, a strategy network is formed, the situation on the chessboard is used as input information, and a probability distribution is generated for all feasible positions.Then train a value network to predict self-games, predicting the results of all feasible positions with a standard of -1 (absolute victory of the opponent) to 1 (absolute victory of AlphaGo). For machine learning technology, in addition to Go’s attempts at AlphaGo, Shi Bomeng, president of Google China, revealed that the next step will be to try medical treatment, and AlphaGo will play a positive role in the treatment and research of diabetes, cancer and other diseases.8. China also has a ten-stage artificial intelligence chess player stunt, will it face off with AlphaGo? Shi Bomeng, president of Google China, clearly answered: No!Although the Go artificial intelligence AI genius developed by Tencent already has ten strengths, Google does not value the previous battle between machines and machines. The confrontation between machines and human intelligence can promote the development of artificial intelligence and human society.I am afraid that the desire to see the Go game is going to fail.