Yesterday, DeepMind CEO Demis Hassabis posted a statement on Twitter: They are working hard to improve the intelligence of AlphaGo, and the new version of AlphaGo will come back in 2017. DeepMind will post more information in the near future.
Fan Wei also posted on Weibo that AlphaGo's chess ability has been greatly improved.
Fan Wei is currently the head coach of the French Go team. He has won the European Championship three times. He is the first professional player to play against AlphaGo. In October last year, he lost to the machine 0-5 in a closed game. Fan Wei then joined the DeepMind team as a full-time sparring.
After Li Shishi and AlphaGo fought in March this year, Ke Jie was looking forward to challenging AlphaGo. Yang Junan, secretary of the Party Committee of the National Sports General Administration's chess and card sports management center, disclosed in public that Ke Jie will challenge AlphaGo, and the news was denied by DeepMind. According to industry insiders, the Chinese chess house did reach a battle agreement with DeepMind, but due to special reasons, it was delayed. Therefore, Ke Jie’s chances of playing AlphaGo next year are very high.
What is AlphaGo doing during the one-year “closed practice�Previously, Tian Yuandong was so "What is AlphaGo so powerful?" 》 mentioned in the article:
Compared to the previous Go system, AlphaGo relies less on the domain knowledge of Go, but it is far from the extent of the general system. Professional players can understand their opponent's style and adopt corresponding strategies after watching a few games. A senior game player can also get started after playing a new game a few times, but so far, the artificial intelligence system has to reach Human level still needs a lot of sample training.
According to industry insiders, during the retiring period of AlphaGo, it may conduct a large number of sample training, and apply reinforcement learning to constantly “move and play with each otherâ€: simulate it on the computer and generate a new game, so that the collected experience and The sample has become more and gradually enhances its ability.
At the same time, Tian Yuandong also explained that sample training is important, but the dynamic combat experience may play a greater role.
In AlphaGo, the role of Reinforcement Learning is not as big as it seems. Ideally, we hope that artificial intelligence systems can dynamically adapt to the environment and opponents' moves in the game and find ways to counter it, but enhanced learning in AlphaGo is more used to provide more quality samples. , Supervised Learning to train a better model.
For this reason, DeepMind recruited the top players such as Fan Wei to give AlphaGo a sparring session, specializing in dynamic combat training. As for the effect, it is still unknown. Zhuang Zhuang, an IBM senior engineer and amateur 4th player in the game city, revealed to Lei Feng that:
AlphaGo is against the Li Shishi version of the V18 version, now the V20 version, and it is not surprising that the V21 version will be officially released early next year. On the surface, at least three versions of the change can be seen that AlphaGo's upgrade speed is relatively fast, and the strength should be improved.
Ke Jie and AlphaGo who are so powerful have been placed on hot topics, so what should human players pay attention to when playing against AI?Zhuang Zhuang pointed out:
The mentality of working with the machine is very important, which is different from playing against real people. Professional players can study their opponent's chess preferences when preparing for major competitions. They can prepare some layout routines, and they can speculate on opponents' choices to some extent, but these are not useful for AlphaGo. What is the chess style of AlphaGo? In fact, it is not so easy to say clearly, but one thing that should be recognized is that value judgments outweigh human chess players in most cases. Its choices have no feelings, not based on chess style, but based on winning percentage and value. When a player faces a strong opponent like AlphaGo, the best strategy is to constantly pursue the best and most direct one, in order to maximize the winning percentage. During this period, the players should be more engaged with the cold machine system to form a unique mentality to deal with the machine players.
In addition to AlphaGo, Zen Go AI has now been upgraded to V13, with a good level of intelligence. It is foreseeable that in the future we will not only see the contest between AI and people, but also the confrontation between the AI ​​systems of the companies that transcend humans.
Magnetic Buzzer Self-drive Type
The magnetic buzzers (Self-drive Type) offer optimal
sound and performance for all types of audible alert and identification. Our
magnetic Buzzer solutions are offered with various mounting options. We also
provide you with a washable version for your preferred soldering method. Our magnetic buzzers, also known
as indicators, are designed with an internal drive circuit for easy application
integration. During operation, current is driven through a voice coil to
produce a magnetic field. When a voltage is applied, the coil generates a
magnetic field and then allows the diaphragm to vibrate and produce sound. This
buzzer type has a low operating voltage ranging from 1.5 – 12V+. Our magnetic
buzzers are desirable for applications requiring a lower sound pressure level
(SPL) and frequency.
Passive Buzzer,Dc Magnetic Buzzer,Electro Magnetic Buzzer,Magnetic Buzzer Self Drive Type
Jiangsu Huawha Electronices Co.,Ltd , https://www.hnbuzzer.com