MIT Releases 2017 Global Top Ten Breakthrough Technologies: Most Life Sciences Technology Selected
Recently, Beijing has a heavy snowfall. As the exclusive operator of the MIT Technology Review in mainland China, DeepTech is jointly affiliated with IBM China Research Institute, Netease Technology, People's Posts and Telecommunications Press, Yunxiangke, and Digital Home in Guomao III. The China Ballroom hosted the 2017 MIT Technology Review, the world's top ten breakthrough technology list. Keda Xunfei, Baidu, Beneficial Technology, LeTV, Horizon, Zhongke Chuangxing, Huachuang Capital, and Mi Capital, Horizon, Epro Payment, Tsinghua University, University of Science and Technology of China, Chinese Academy of Sciences, Zhongke Jingyun, Hesheng Gene, The guests of dozens of institutions including Yuanji Gene, Jinzhun Gene, Huaxing Capital and Asao Venture Capital participated in the conference. As one of the world's most famous technology lists, MIT Technology Review has the world's top 10 breakthrough technologies with great global influence and authority. It has been held for more than 16 years. Some of the annual rankings have already been applied in reality, and some still take time, but their importance is undoubtedly ruthless, destined to have a major impact on our economic and political life in the future, and even completely change the whole The cultural face of society. DT Jun's US partner MIT Technology Review publisher and editor-in-chief made a special speech on the top ten breakthrough technology China, and made a brief introduction to the list. Reinforcement learning Reinforcement Learning Technological breakthrough: Reinforcement Learning (RL) is an artificial intelligence method that enables computers to learn independently as humans without explicit guidance. Important: If machines are not able to autonomously hone their skills through environmental experience, the speed of progress in autonomous vehicles and other automation areas will be greatly limited. main researcher: - DeepMind - HKUST - Mobileye - Alibaba - OpenAI - Microsoft Research Asia - Google - Chinese Academy of Sciences - Uber - Baidu Maturity: 1-2 years Reinforced learning technology is the key to AlphaGo's ability to master complex Go games and beat the world's strongest professional players. Today, reinforcement learning is rapidly evolving and gradually infiltrating artificial intelligence into all areas besides games. In addition to improving the performance of autonomous vehicles, the technology also allows robots to comprehend and master skills that have never been trained before. In essence, reinforcement learning technology is a basic rule of learning from nature. Psychologist Edward Thorndike also noticed this more than 100 years ago. In the most famous mezzanine experiment, Sondike put the cat in a murder box, and the cat can only escape by pressing a lever. Observations show that after a long period of back and forth, the animal will eventually accidentally step on the lever and escape. Some of the earliest artificial intelligence researchers believe that the process of the lost box experiment is likely to be effectively reproduced in the machine. As early as 1951, Marvin Minsky created the world's first learning machine, using a simple form of reinforcement learning to simulate how a mouse learns to get out of the maze. However, there have been few gratifying achievements in this field in the following decades. In 1992, IBM researcher Gerald Tesauro demonstrated a program for playing backgammon using artificial intelligence. Soon, this app is very skilled and is enough to compete with the best human players. This is a milestone in the history of artificial intelligence development. Reinforced learning techniques work because the researchers figure out how to let computer programs calculate the enhancement values ​​that should be assigned in each state. In the case of the box-box experiment, in the process of getting out of the labyrinth, the computer program will make a prize or punishment evaluation every time the "simulated mouse" makes a "turn left" or "turn right" action. Also, all assigned enhancement values ​​are stored in a large table, and the calculation program gradually updates the data as the learning progresses. But for large and complex tasks, this method is computationally impractical. However, in recent years, deep learning techniques have proven to be an extremely efficient way to identify data patterns, whether the data here refers to a turn in a maze, a locus on a chess board, or a screen on a computer game. The pixels, or the complex road conditions that are faced when driving automatically. In China, taking Keda Xunfei as an example, this company has already carried out research and application in several directions for reinforcement learning, including human-machine dialogue system, intelligent customer service system, machine-assisted driving, robot control, etc. Applied research. Take a multi-round human-computer interaction system such as the dialogue system as an example. It is a very typical case of reinforcement learning application. In the traditional task-complete dialogue system, users need to describe their needs clearly in an interaction process. Such interaction is not natural. In the AIUI interactive system framework of Xunfei, the idea of ​​multiple rounds of interaction is introduced. A deep reinforcement learning (Markov decision process) model is used to guide the user to input requirements, so that the user tasks can be completed quickly and naturally. Greenhouse Film Manual Roll Up The Greenhouse Film Electric Roll-Up Unit replacing traditional manual roll-up unit to reach high speed automation in revolving & releasing films and greenhouse environment, it is the new generation of greenhouse ventilation and temperature regulation facilities. Portability, output torque, output lock, wide range and precise adjustment, waterproof, dew-proof and power-saving are the key characteristics for Electric Roll-Up Unit, which is the best choice for modern daylight greenhouse applied in different climates. Greenhouse Film Manual Roll Up,Greenhouse Manual Winch,Greenhouse Side Film Roll Up JIANGSU SKYPLAN GREENHOUSE TECHNOLOGY CO.,LTD , https://www.alibabagreenhouse.com
