"The car of the future should be a wheeled robot that can learn." In the face of the continuously heating up of autonomous vehicles, at the first China International Intelligent Industry Expo held recently, Li Deyi, an academician of the Chinese Academy of Engineering, said, "It should be able to pass Edge computing and the learning of'driving super brain' make autonomous driving safer than human driving."
From a certain perspective, the self-driving car that has attracted a lot of attention and investment is an AI game, and the advancement of AI will be reflected in edge computing first, and will help edge computing to solve the data avalanche caused by the Internet of Things.
Although various research institutions have different predictions for the explosive growth of global data, it is an industry consensus to move calculations and data closer to users. In the eyes of the various technology gods who have entered the game, this new value of 100 billion yuan The arena will first become a must for the consumer market and smart factories.
While the relationship between edge computing and cloud is being heatedly discussed, the "What Is Edge Computing" recently written by CB Insights, a well-known venture capital research institution, has received widespread attention. The article detailed the development and application prospects of edge computing. The article stated that cloud computing is no longer sufficient for real-time processing and analysis of data generated or about to be generated by IoT devices, connected cars and other digital platforms, and edge computing will come in handy.
But what exactly is edge computing? Many people are still at a loss. In answering this question, Yan Lida, president of Huawei's enterprise business, said that it is similar to an octopus. Among invertebrates, octopus has the highest IQ. It has a huge number of neurons, 60% of which are distributed on eight legs (carpals and feet), and only 40% of the brain. However, it seems that octopuses that use "legs" to think and solve problems have excellent coordination between their arms and legs during hunting, and they never get tangled and knotted. This is due to the fact that they are similar to the "multiple cerebellum + one brain" of distributed computing ".
Current artificial intelligence applications rely more on the cloud, and edge computing pushes intelligence from the cloud to the edge. In the words of Zhang Yongqian, general manager of Horizon Intelligent Solutions and Chips Business Unit: "Computing is moving from the center to the edge." In other words, not all applications need to be placed in the cloud. For example, when an object moves in the house, the smart home The security system detection can rely on the terminal equipment resources to detect whether the dog is moving, or the thief who broke in, and make precise intelligent control based on this.
In the future, if there is no support for edge computing, many applications may be a picture. For example, the hot driverless cars have very demanding requirements for real-time information interaction, data transmission, and interaction delay indicators. Once the system responds slowly, traffic accidents will occur. , The experience of unmanned driving is greatly reduced.
It needs to be emphasized that although we will push more and more basic tasks to the edge of the device, as Philip DesAutels, senior director of the Internet of Things of the Linux Foundation, emphasized, this only means that the closer the device is to the edge, the smarter the device will become. , It cannot be said that it has nothing to do with the cloud. The cloud will also become smarter and smarter because of the edge, and its more important task in the future will be the central coordinator.
Edge Computing + AI: Coming to Solve Pain PointsAs Intel said, the world is falling into an avalanche of data torrents, and AI needs to consume a lot of computing resources and storage resources when performing analysis and processing. If you can use edge computing like an octopus, massive data can be processed nearby, and a large number of devices can work together efficiently, and many problems will be solved.
Wangsu Technology Chairman Liu Chengyan said that with the arrival of the intelligent society, the network capacity and demand for computing brought about by the explosion of the Internet of Things will be far greater than the demand brought about by the interaction between humans and the Internet. The data capacity generated by the terminal will also be far greater than that of human beings. It is expected that 50% of the computing power will be placed on the edge in the future.
Hong Xiaowen, senior vice president of Microsoft, chairman of Microsoft Asia Pacific R&D Group and dean of Microsoft Asia Research Institute, told reporters: “Edge computing can make local decisions and judgments, and can ensure uninterrupted operation in an offline state. It is suitable for smart factories and unmanned vehicles. Other scenarios that emphasize continuity and safety."
Dr. Yu Zhang, Chief Technology Officer of Intel’s China Internet of Things Division, said: “How to achieve intelligence at the edge of the network is one of the key links in controlling the data torrent and an important trend in the future development of the Internet of Things.â€
The so-called intelligent marginalization requires the rapid migration of models on the cloud to offline, transforming cloud intelligence into lightweight intelligence available at the edge, adapting to the edge software and hardware environment and usage scenarios, while the cloud native model needs to be used at the edge Nodes need to do model conversion, compression, and tuning. Mr. Liu, a well-known strategic researcher from a major factory, told a reporter from the Science and Technology Daily: “It seems that they are all technical problems, but they are not so easy to implement. There are many benefits of a unified AI architecture in the online and offline technology ecosystem, but without industry collaboration, combined with application scenarios, comprehensively building a three-dimensional ecosystem, artificial intelligence to help edge computing can only be empty talk."
Mr. Liu emphasized that artificial intelligence is not only mathematics and algorithms, but also a comprehensive application of technology. Promoting productivity and solving efficiency is the top priority of AI, not those supernatural capabilities. Looking at artificial intelligence in edge computing, you can understand this best. From an industry perspective, intelligent edge is the hardware form of artificial intelligence. Single-point breakthroughs are meaningless. The three-dimensional application breakthrough of artificial intelligence + edge computing is valuable. Can really solve the pain point problem.
Security risk: a challenge that cannot be ignored in edge computingYu Haibin, director of the Shenyang Institute of Automation, Chinese Academy of Sciences, said: “No matter what kind of combination the edge computing technology will form in the future, it is nothing more than the integration of industry, manufacturing, sensing, control, computing, storage and networking. Can edge computing be explored? It will be very important to develop a new industrial model that faces the future."
As mentioned earlier, it is not simple to push artificial intelligence to the edge. Among the various challenges and limitations that continue to emerge or have always existed, security risks are an issue that the industry needs to pay attention to urgently.
Edge computing is like the bridge between the physical world and the digital world and the first entrance to data. With the proliferation of smart edge devices (including mobile phones and IoT sensors), emerging attack vectors will continue to emerge, with real-time, deterministic, and diverse data. Many challenges such as sex are already in sight.
Internet industry analyst Chen Jinyu said: “Although edge computing can bring security components closer to the source of attacks, launch more efficient security applications and increase the number of layers to defend against core violations and risks, it is undeniable that artificial intelligence In this era, all links such as data collection and use are already facing new risks."
The fact is also true. In the data collection process, large-scale machines automatically collect thousands of user data and can fully track users. In the data use link, with the widespread use of big data analysis technology, specific individuals are easily locked in, consumption habits, whereabouts and other information are increasingly used in "precision marketing", and the scale of use has always been Grey area.
Chen Jinyu said: "Not only that, in the data life cycle, due to hacker attacks, system security vulnerabilities, etc., personal data will always be leaked security risks. At present, how will this risk be extended to artificial edge computing? The smart consumer and industrial sectors are still difficult to predict."
Directional Antenna WiFi,Directional Antenna TV,Directional Antenna LTE,Directional Antenna Indoor,Directional Antenna Outerdoor
Yetnorson Antenna Co., Ltd. , https://www.yetnorson.com