In recent years, there have been two hot areas in the scientific and technological world: one is 5G and the other is AI. Both are disruptive technologies that can change the times. At present, the 5G network standard has been completed, and the 5G industry is in the commercial sprint stage; in AI, the industry generally believes that the fourth scientific and technological revolution is another technological revolution following the steam technology revolution, the power technology revolution, and the computer and information technology revolution. AI is an important driving force for the fourth industrial revolution.

5G commercial timetable

As we all know, the two revolutions of AI and 5G will each establish their own huge industrial circles, bringing huge economic and social effects. In the future of data-driven, artificial intelligence, autonomous driving, and 5G are the most disruptive opportunities. They are accelerating breakthroughs and landings, solving many major challenges facing human society, and paving the way to a smart world. Technology is changing our lives faster, and with the development of 5G and AI technology, more changes are coming. Looking at 5G or AI technology alone, their development is facing many challenges. We may wish to open our minds: When the 5G and AI two subversive technologies are deeply integrated, what kind of changes will the human world make?

From AI+5G to AI¡Á5G

Earlier, in order to explore the fusion relationship between 5G and AI, Intel gathered experts and scholars in the industry to preliminarily conduct an expert symposium on the future of 5G and AI. The symposium, called "THE NEXT Afternoon Tea," aims to explore cutting-edge trends such as artificial intelligence, autonomous driving, and 5G, providing ideas and practical guidance for the future of the industry.

5G and AI, which are already an irreversible future, also represent two revolutions, as evidenced by the strong bets of giants including Intel. In a sense, 5G is the foundation of the Internet of Everything, and AI is the tool to achieve the intelligence of all things. In Intel China Research Institute, 5G and AI are already in the process of evolution. 5G network as an infrastructure has changed from serving people to becoming a service. AI is not a simple solution, but a tool.

How to combine these two core technologies is a thing that can exert a lot of imagination space, and the convergence of the two can form an alternating upward trend. For example, a drone called ¡°airborne fireworks¡± in a light show is a good example. Thanks to powerful computing power and a highly reliable network, drones can build the power of group intelligence, not just Creative performances can also be applied to search and rescue, express delivery and other fields.

"The two revolutions of AI and 5G are not simply to liberate it, but to hold it up. It is best to form an alternate way to rise." Intel China Research Institute said that AI and 5G are not just an addition. , but multiplication, and maybe even a power relationship. Because when AI can play the role of ¡°self-energy¡± in the subtleties, and then use 5G to open the cloud connection, it will definitely open the brain, create unlimited possibilities, and will be superimposed step by step to form continuous innovation.

The State Key Laboratory of Mobile Communications of Southeast University believes that only AIx5G is needed to achieve true network self-energy. This is to consider how to use AI technology to 5G to solve problems that could not be solved in the past.

So the question is, how to evolve from the current AI+5G to AIx5G?

This is not only the introduction of AI in the existing system architecture, but also to consider the real design of the system architecture to have the characteristics of AI, to achieve self-energy. A truly self-powered network should be distributed, neglected like a human brain, without "best effort", but enough.

For how to achieve the self-power of 5G networks, Intel concludes in three phases:

Operators need network transformation, adopt software-defined infrastructure, let AI's algorithm work, and automatically adjust the network architecture according to the external environment;

Intelligent, reflecting the terminal aspects, different network equipment, training through some data, and generating recommendations for decision-making;

The stage of true self-energy.

5G wants AI not only because of the goal of self-energy, but also because of the increasing complexity of the operator network, the increase of operation and maintenance and network construction costs, and the explosive growth of data and user traffic. Network equipment can't meet the user's data explosion, and AI can be deployed to different ends of the 5G network to achieve different functions of local local user requirements and network operation and maintenance resource management. With AI for data analysis, it will be used for 5G applications. It is infinitely possible.

Will AIx5G produce "new species"?

Qualcomm once said that with the advent of the 5G era, there will be many multi-domain intelligent terminals that rely on 5G technology. At the same time, AI will connect to more devices, which will bring more convenience to our daily lives. The combination of 5G and AI will lead to the intelligentization of network edge terminals. ¡°In short, the development of AI artificial intelligence on the terminal side requires the 5G bridge to communicate with the cloud big data.¡± ¡°From the perspective of wireless terminals, smart phones, Internet of Things, and automobiles can all use AI artificial intelligence technology. ¡±

Security terminal smart shines

As one of the earliest industries in China's Internet of Things application, intelligent security is committed to realize the intelligent integration of terminal products, enabling them to connect wireless sensors and transmit alarm information via the Internet. The end-side artificial intelligence algorithm has many advantages such as instant response, improved reliability, enhanced privacy protection, and efficient use of network bandwidth. The application of AI¡Á5G technology on the security terminal side is becoming more and more integrated, and has gradually become the ¡°engine¡± that drives the new generation of technological revolution.

In the future, all security-related sensors will be interconnected. The current wisdom community, smart home or other new wisdom concepts, although still a shallow connection, must rely on 5G technology in the application process. Connected.

Taking the smart home security camera as an example, the smart camera can localize the video content on the terminal side without waiting for the data to be transmitted back and forth between the network and the cloud, and the response speed will be faster; and the addition of the 5G network will make it and other The system works better, and the speed of transfer of useful data to the cloud will also speed up. Some small and medium-sized enterprises, shop owners and home users can use real-time browsing of high-definition images of home or shop terminal cameras with smart terminals, and can also realize remote control and storage recording. Capture images and keep track of the dynamics of the area of interest.

In addition, advances in terminal side AI x 5G will be critical to the development of near-natural language voiceprint biometrics. People are accustomed to quick response and seamless conversation during the conversation. Therefore, natural voice interactions should have no delay in the user experience. With AI and 5G technology, machine terminals will be closer to natural conversations when interacting with humans.

A new breakthrough in the driverless field

Compared to traditional cars, autonomous vehicles have the elements of a new species.

When the Internet of Vehicles, autonomous driving, and information sharing technologies are mature, the AI-driven cars are incorporated into smart transportation and even the smart city as a large network platform. We have seen a new species that releases social productivity and will have a large number of human beings. The driver is freed from driving.

The car is just a smart node, it will connect with smart traffic and even smart city AI through 5G. The car's driving route planning, speed, start and stop can be managed by the AI of smart traffic. The vehicle sensor synchronizes the road condition information during the driving process with the smart traffic AI in time. When an emergency or unexpected situation occurs, the vehicle AI actively controls and reports to the smart traffic AI in real time to wait for further processing instructions. Smart Transportation AI will synchronize information with other related autonomous vehicles and generate further automatic control.

We simply do a calculation, assuming that 50 million cars are on the road every day, and that 50 million drivers drive one hour a day, which is 50 million hours. If you have automatic driving, this 50 million hours of savings can be used for entertainment and work for almost 5,000 years. If self-driving breaks through, it will spawn more unmanned systems and save more human time.

Tsinghua University experts said that 5G network connection solves the problem of data flooding in the process of automatic driving, but only with AI can make the unmanned vehicle have cognitive ability, 5G has already applied the application scene in automatic driving, such as Remote security monitoring in autopilot testing.

"The data of these end-to-end connections cannot be handled by hand, but by automated methods, which requires the full use of artificial intelligence. From data filtering, optimization, filtering to network design flexibility, software, The use of artificial intelligence is required."

Artificial intelligence has two layers of end-to-end, one layer is the input end to the output end, which requires deep learning neural network to realize detection, classification and perception functions, and another layer end-to-end is the edge end and the cloud end. From the latter perspective, it solves the connection problem. However, from the perspective of the safety of autonomous driving, the connection cannot be closed-loop, and it is necessary to achieve local autonomy at the edge. In this way, the automatic driving can realize the simultaneous energization of both the edge end and the cloud side, and the cloud end support with the data power is very strong, and the edge end support of the local autonomic calculation is supported.

"5G must solve the problem of applicability and autonomy on the basis of artificial intelligence, because it is an industrial application scenario. The basis of 5G plus artificial intelligence may become a 6G network, and automatic driving will change. It¡¯s better.¡±

"AI" makes "5G" more self-sufficient, "5G" makes AI more ubiquitous

As a basic infrastructure, the 5G network has undergone major technological breakthroughs from serving people to serving at the same time. Since AI technology is a basic enabling technology, AI applications can be found in almost every field, as well as in 5G.

Different from the past 2G to 4G era, focusing on mobility and transmission rate, 5G should not only consider enhancing broadband, but also consider multiple dimensions such as Internet of Everything, diversification of future demand, diversification of key technologies, and diversification of evolution paths. The 5G demand has also become very diverse, and the technology and evolution path are naturally many times more complicated.

The current carrier network complexity is getting higher and higher, data traffic is exploding, user traffic is exploding, and existing network devices cannot satisfy users' data explosion. As the complexity of the network increases, the cost of operation and maintenance and network construction has greatly increased. In particular, 5G is not only a connection between people and people, but also a connection and object. The existing network maintenance and management methods or manual intervention methods have no way to adapt to the needs of the 5G era network.

Therefore, 5G needs to be managed by self-energy. Autonomous connection path selection, automatic network connection health analysis, and even the ability to repair known faults themselves. Utilizing AI's self-learning, data analysis and other technical expertise, it gives 5G the ability to be "autonomous" and "self-energy".

AI gives 5G self-energy, while 5G gives AI a broader connection.

The 5G connection from the edge to the cloud is the most ideal connection to date. Through the 5G connection, the decision-making and planning parts are put into the cloud processing, and the empowerment is doubled from the edge to the cloud, so that the AI algorithm has the ability to extract the corresponding association and enhance itself; after the individual is promoted, through the 5G network and the cloud brain The ability will be quickly distributed to other individuals.

That is to say, after the combination of AI and 5G, the machine will produce the ability similar to group wisdom, bringing value to the whole society. It will also give birth to the requirements of the network's own adaptive capabilities, which is a new opportunity for mutual promotion and spiral development.

5G is the cornerstone of the Internet of Everything, and AI is the booster of the Internet of Everything. As the productivity of the new era, the two will bring about changes in the production mode of the whole society and the improvement of productivity. Adding and interacting with each other, AI will enable 5G, optimize 5G networks, and push 5G to land. A 5G network with AI attributes is a self-powered network. 5G is also an enabling technology that changes production methods, changes social life, and makes AI everywhere. As a new basic network facility, 5G not only serves people, but also serves the goods and serves the society. 5G's connectivity will drive the intelligent interconnection of everything. But AI and 5G will not just add up, the future development is more AI ¡Á 5G.