Netease smart news May 24 news, Go game summit will enter the next day, at today's artificial intelligence summit, DeepMind founder, Demis Hassabis (Demis Hassabis) delivered a speech, introduced the development process of AlphaGo , and what AlphaGo means.
AlphaGo is an artificial intelligence system created by DeepMind. DeepMind was founded in London in 2010 and joined Google in 2014. According to Hasabis, DeepMind participated in Google's artificial intelligence "Apollo Project", and also created a new way to develop science.
"DeepMind's vision is to study what is AI, and then use smart to solve all problems, that is how we put forward effective suggestions to solve the problem, and we ultimately hope to establish general artificial intelligence." Hasabis said.
According to Hasabis, the general-purpose learning machine built by AlphaGo has two features. One is “learningâ€, which is a non-programming pre-set and can learn raw materials autonomously. The other is versatility, that is, the same system can perform multiple tasks. "A series of algorithms and systems can do a series of tasks, these tasks may be unprecedented." Hasabis said that the general strong artificial intelligence, and now weak artificial intelligence is not the same, the current weak artificial intelligence are Presupposedly, in fact, the chess program designed by IBM in the 1990s is also a preset default artificial intelligence. “It is through brute force search. The machine passively accepts this program and cannot learn from itself.†Hasabis Speaking.
DeepMind hopes to create a reinforcement learning framework, and Hasabis said that strengthening learning framework is to let “agents†in a real/virtual environment to accomplish a specific goal. In order to accomplish this task, the agent can observe the environment, which includes visual, auditory, and tactile observations. After this, set an idea in your mind, and then take action and practice the plan. Can act in real-time observation. "If we can solve this problem, in fact, strong artificial intelligence has already been realized, which is the same way as the human agent observes and learns," said Hasabis.
Hasabis mentioned the Atari Agent, which was an AI program developed by DeepMind before AlphaGo (three years ago). It passed the 2600 test stand and tested more than a hundred eight-legged eighty-eighth Atari games. The Atari agent of the time only input the original pixel (~30), the goal is to use deep reinforcement learning and play good Atari games.
AlphaGo is currently the latest DeepMind artificial intelligence system. Hasabis said that while AlphaGo is playing Go, it can also do other things.
Why is it so difficult to go to the computer? This is because its complexity makes it difficult to solve exhaustive search. This problem includes two aspects. One is that it is impossible to write assessment procedures to determine winning or losing, and the other is that the search space is too large.
In Hassabes's view, it is even more difficult for Go not to rely on calculations, such as chess, but to rely on intuition. “There is no hierarchy concept in Go, and all pieces are the same. Go is a defense game, so you need to plan for the future. In the process of playing chess, the board is in the heart and you have to predict the future. A small piece can move the whole world, Launching the whole body. Gossip's wonderful hand is inspired by the apocalypse." Hasabis thus explained the intuition in Go.
Technically, AlphaGo uses two kinds of networks: a strategic network and a valuation network. These technologies were published in Nature last year and inspired many researchers to design their own artificial intelligence systems.
Hasabis recalled that we have conducted tests since then. In 2016, we let AlphaGo play against Li Shishi. In the end, AlphaGo defeated Li Shishi 4:1. “We waited 10 years at this moment and it was really a decade. Li Shishi filled with emotion.
AlphaGo attracted 280 million viewers, 35,000 reports, and board sales increased 10 times in the West. Hasabis recalled the highlights (AlphaGo's game against Li Shishi). They were the amazing 37th move in the second game. It made us ponder: For thousands of years, you humans have underestimated the 5th line. the value of. In addition, there was a 78th goal in the fourth game. He won a set. Li Shishi said, "I think this has introduced new ideas for Go. I feel that I have found a new reason to continue playing Go."
Hasabis said that many arts are subjective. AlphaGo regards Go as an objective art and analyzes the impact of each step. "Thus, my definition of intuition is that I can obtain preliminary perceptions through experience and I cannot express myself. I can confirm the existence and correctness through behavior." Hassabis stated that AlphaGo has been able to imitate human intuition and has creative ability to pass. The ability to combine existing knowledge or unique ideas. So AlphaGo already has intuition and creativity, but these capabilities are currently limited to Go.
After this, DeepMind hopes to make up for the gap in AlphaGo's knowledge. Later, it released a new version of "Master," and played the game on the Internet. It also won a great victory. Ke Jie sighed after playing with Master. Humans have been practicing for thousands of years. The computer tells us that human beings are all wrong. Hasabis said, “In the 1930s and 40s, Wu Qingyuan brought revolutionary power to go. I believe that AlphaGo can also open a new era of Go. Chess programs speak tactics and AlphaGo speaks strategies.â€
"How far we are from the best, how is the perfect game? 3000 years of playing is not enough to find the best game. AlphaGo allows us to explore these mysteries." Hasabis said.
Beyond Go, Hasabis hopes to apply artificial intelligence to various fields. Hasabis said, “Human-machine cooperation can achieve 1+1>2 results, human intelligence will be amplified by artificial intelligence. Artificial intelligence and AlphaGo are tools, just like the Hubble telescope, can advance the progress of human civilization. "Hazabis said that numerous other areas will also be bombarded by combos. Strong artificial intelligence is also the best tool for our exploration, such as the use of AI in materials design, new drug development, and real-life applications such as Medical, smart phones, education, etc.
Currently, DeepMind has been very successful in using AlphaGo in data center optimization. The results show that it can save 50% of the power.
Finally, Hasabis concluded that information overload and system redundancy are huge challenges, and we hope to use AI to find meta solutions. "Our goal is to implement artificial intelligence science or artificial intelligence assisted science. Of course, AI must have ethical and responsible constraints. In short, artificial intelligence technology can help us better explore the mysteries of the human brain."
燑br>
Second, lithium is much lighter than other metals used in batteries, such as lead, which is important for small objects such as phones but also for cars that require many batteries. Third, lithium-ion batteries are rechargeable, because lithium ions and electrons move easily back into negative electrodes.
The new Streamlight Stinger 2020 Flashlight is powered by its SL-B26 Battery Pack to ensure a powerful performance. The two packs are protected lithium ion batteries with a 2600mAh capacity, on-board safety control circuits, and built-in USB charging ports. These batteries can be recharged up to 500 times, saving you time and money.
Power your world with GLIDA Battery Products- Lithium Ion Battery. ... The latest trend in lithium battery used in consumer electronic products and IT gadgets. It is flexible in footprint and ... $0.00. 3.7V 6000mAh 18650 (1S2P) Li-ion Battery Pack.
Lithium Ion Battery,Battery Lithium Ion,Li Ion Battery,Rechargeable Li-Ion Battery
Shenzhen Glida Electronics Co., Ltd. , https://www.szglida.com