On March 22, the Institute of Computing Technology of the Chinese Academy of Sciences released the world's first "neural network" processor chip capable of "deep learning", called "Cambrian". This achievement will be officially put into industrialization this year. In the near future, anti-fraud brush pay, image search, etc. will be more reliable and easy to use.
Not long ago, a Go program "AlphaGo" developed by Google Inc. defeated Korean player Li Shishi 4:1. The secret of "AlphaGo" is to imitate the "deep learning" of human beings through neural networks.
Deep learning refers to a multi-layered artificial neural network and a method of training it. Popular speaking means that computers use deep neural networks to simulate the mechanism of the human brain to learn, judge, and make decisions. In recent years, this method has been applied in many fields, such as face recognition, speech recognition, etc. It will be a hot research direction in the field of artificial intelligence in the near future.
"Deep learning" can be developed to the present stage, benefiting from the increased computing power of computing systems, and this improvement is the result of the explosive development of processors as technical support. Currently, the processor used by "AlphaGo" is a CPU processor that is common in other fields. In 2010, Google used 16,000 processors to run for 7 days to train a deep learning neural network that recognizes cat faces. In the game of Go, the human "AlphaGo" needs more processors. Ordinary people want to use this. Technology is impossible.
The deep learning processor is to create a chip for the computer that mimics the multi-layer artificial neural network of the human brain. In the operation of the deep learning processor, the computing power of the computing system is the key to determining the efficiency of deep learning processing. The release of the "Cambrian" processor is two orders of magnitude better than the processor used by "AlphaGo", which means it can make artificial intelligence run faster and farther.
This year, the research team and the Institute of Computing Technology of the Chinese Academy of Sciences have already incubated the Chinese Academy of Sciences, and officially started the industrialization of scientific research results. The future application aims at three high-performance servers, high-performance terminal chips and robot chips in enterprises and research institutes to realize more functions. For example, taking a photo with a mobile phone knows who the photo is, and can intelligently classify many videos. The image search will be more accurate and easy to use. Just take a tree on the side of the road and you can search for this tree. All the information, not just the current text search. The future service areas of the "Cambrian" include both social and people's livelihood as well as major national needs.
Candy consumption trend is developing towards nutrition, health and functional field in worldwide. Therefore, we develop a series of healthy sugar-free chewing gum for sub-health people, including refreshing energy gum, collagen gum, weight losing gum, probiotic gum, GABA sleep gum, VC skin-care gum, hangover cure gum, nicotine gum, oral-care gum, etc., those can not only protect the health of teeth but also supplement the nutrients needed.
Functional Chewing Gum,Swallowing Gum,Sugarless Chewing Gum,Natural Chewing Gum
Soar Biotech Co.,Ltd , https://www.xylitgum.com