The current status of medical big data processing and mining will be developed in the future. It will provide a series of reports on typical cases, investment and financing, and industrial layout in the field of medical big data at home and abroad for reference by investors, entrepreneurs and medical institutions in the industry. AlphaGo defeated the world's Go champion Li Shishi, and the reputation of DeepMind has since spurred. Now it can not only defeat the top Go players, but also defeat professional gamers! In fact, DeepMind is an artificial intelligence laboratory based in London, UK. Their research direction is to develop a universal self-learning algorithm. It was acquired by Google Lightning in 2014. Because Facebook is also competing, the price is not cheap, 400 million. Dollar! After the acquisition, DeepMind has been operating independently, unlike the usual pre-programmed machine learning. Their goal is to combine the best techniques from machine learning and neuroscience systems to build powerful and versatile learning algorithms. The industrial field can be widely used. What is the core of Deep Mind technology? DeepMind was founded in 2012 by neuroscientist, young talented player Demis Hassabis and two partners. On DeepMind's official website (just revised), three core sections are very eye-catching: AlphaGo, DQN and Health, which can be seen from DeepMind. The focus of research. The foundation of DeepMind research is "deep learning" technology! Deep learning is an important branch of Machine Learning. Its goal is to imitate the way human neural networks perceive the external world. An emerging topic in the field of AI (Artificial Intelligence), through the use of neural networks to enhance machine learning. Capabilities, Google's speech recognition and NLP technology are relatively mature, and there are also voice and NLP-based development platforms and smart hardware products. The more mature deep learning part of the application mode is as follows: 1. computer vision: a science that studies how to make a machine look, uses a camera and a computer instead of the human eye to identify, track, and measure the machine vision of the target, and then perform image processing so that the human eye can proceed. Observed action. 2. Natural language processing (NLP): A branch of artificial intelligence and linguistics that explores how to handle and apply natural language. 3. Object Recognition: In the field of computer vision, it refers to finding an object in an image or a group of video sequences. 4. Machine translation: A field of computational linguistics that studies the translation of a passage or speech from a natural language into another natural language through a computer program. What does ordinary machine learning look like? Basically, it is a coding instruction that has been set in advance in the computer. It learns or establishes a pattern from a large amount of data, and speculates on new examples or learns more knowledge according to this model. Of course, the problem is inevitable. On the one hand, it has to spend a lot of time, and it relies on human coded programs to let the machine learn abstract concepts, so it can't surpass human wisdom. Unlike ordinary machine learning, the core of DeepMind deep learning is how to let computers discover the patterns that exist in big data. The solution is a combination of deep neural networks and reinforcement learning. In the evolution of the algorithm, it has also experienced continuous improvement of Q-learning, Reinforcement learning, Deep Q-Networks and alphago. By simulating large-scale neural networks, it enables computers to learn. Without the need for direct human intervention, computers themselves “think brains†like humans, and artificial intelligence agents interact with the environment to learn knowledge. What the computer learns with deep learning algorithms is a more abstract expression concept. The rapid accumulation of big data, the rapid development of large-scale parallel computing, and the emergence of new algorithms have jointly promoted the transformation of neural network technology and played a greater role! It sounds more complicated. For example, the most amazing thing is that Deep Q-Network (DQN) can realize the rules of self-learning games. Without prior information to any relevant games of DQN, it can still improve its performance through continuous games. According to overseas media reports, DQN actually tested 49 different rules including "Space Invaders", "Breakout" and "Pong". 29 games outperformed humans, and 43 games had internal AI. Being constantly improved! Demis Hassabis, founder of DeepMind, said that "the next phase of AI DQN will develop a system that can learn more complex 3D games. This way, if you can drive normally in a racing game, you can realize its intelligent control." The real car." At present, the DeepMind team has more than 100 professional papers and achievements on neural network technology published in professional journals such as "Nature" and "NIPS". Caladium Praetermissum,Caladium Bicolor,Caladium Humboldtii,Caladium Cultivars fanhua nursery , https://www.fanhuanursery.com.jpg)
Google DeepMind can mimic human neural networks using precision medicine to prevent disease
Next Article
Do you know a few cold tips about these colds?
Prev Article
Eggplant "Trilogy"