On August 27, the Digital Economy Summit, the concurrent conference of Shenzhen (International) Artificial Intelligence Exhibition, opened in Hall 7 of Shenzhen Convention and Exhibition Center. The summit was held under the theme of "Intelligence Changes the Future, Innovation Leads the World", with Zhang Xinhong, chief information officer of the State Information Center and director of the Sharing Economy Research Center, Chen Xiaoping, director of the Robotics Laboratory of the University of Science and Technology of China and other senior experts in the artificial intelligence sector, as well as entrepreneurs from Tencent, Xiaomi, Saiwei Intelligence, Huina Technology and other high-tech enterprises in the artificial intelligence industry. Entrepreneurs from Tencent, Xiaomi, Seiwei Intelligence, Huina Technology and other high-tech enterprises in the AI industry conducted in-depth analysis and exchanges on the dynamics of industry, academia and research, discussed how to empower AI into the business and social economy from an international perspective, focused on the direction of AI industry and foresaw the development trend of AI.

The summit mainly consists of two modules: guest speakers and roundtable forums, divided into a morning session and an afternoon session. The event was characterized by a constant clash of ideas, a warm atmosphere and continuous applause.

The morning session focused on the discussion of "how to implement AI technology and integrate with the real economy".

Zhang Xinhong: "Artificial Intelligence for Real Economy Development

Zhang Xinhong, Chief Information Officer of the State Information Center and Director of the Sharing Economy Research Center, firstly shared his views on the topic of "Artificial Intelligence for Real Economy Development".

(Zhang Xinhong shared "Artificial Intelligence for Real Economy Development")

Zhang Xinhong talked about: what trends will the future development of the digital economy actually show? From the perspective of technology application and development, there will be four trends of digitalization, networking, data and intelligence. From the perspective of application scenarios and grips, there will be platform, ecology, personalization and sharing. If the eight trends can be mastered, it is likely that we can take advantage of these opportunities to promote the real economy to obtain new economic development.

First, digitalization. Digitization is the process of transforming our real world into zero and one, why into zero and one, because the computer will know two numbers, one is zero, one is one, when the world can be transformed into zero and one, the machine can identify it, calculate it, and can guide humans to change it. The digital economy that we see today has triggered all the changes related to zero and one.

The most powerful thing about the original car is its engine. The most powerful thing about Tesla is not the engine, the advantage of the engine is zero here, but the most powerful thing is its driverless system. Digitalization brings real change, is the digital redefinition of everything.

Second, networking. The network redefines everything. The network spoken of here is the Internet, and when the Internet connects everything together, the world has changed very much. When the Internet developed to the second level is called the mobile Internet era, the connection is the cell phone, through the cell phone to connect all the people together.

When the Internet develops to the second level, it is called the Internet of Things, when all objects are connected together, when all means of transportation, when all parts are connected together, it is called the Internet of Things, and when the Internet of Things era comes, we will find that data is increasing in large quantities, and objects can circulate and talk freely with each other, and they can decide what to do next.

Internet can be a channel, can be infrastructure, it is also a technology, tools, it is also a culture, or a kind of thinking, Internet thinking and so on, the Internet is becoming more and more functional, when the Internet + various factors, will find that there are countless new combinations, infinite number of combinations means infinite number of innovations.

Third, data. In the future, all actions, all decisions are data-driven, so called data-driven everything.

Fourth, intelligence. Artificial intelligence, machine intelligence will increasingly replace humans, to create more value for us, but the artificial intelligence we are currently talking about, artificial intelligence components are mostly, intelligent components are less, the future to move from the current weak intelligence to strong intelligence, super intelligence, the future world and now more different.

Fifth, the platform. No platform is not economic, large enterprises must find ways to build a platform, small and medium-sized enterprises, small and micro enterprises, if you do not have the ability to build a platform, then you should at least use the existing platform, through the platform for their own empowerment.

Sixth, ecological. The development of platform without boundaries will definitely lead to ecologization, and a large number of demands and supplies are gathered on the platform, and new supplies and demands will be generated before the demands and supplies, so the future ecologization strategy must be the goal-oriented of enterprise development.

Seventh, personalization. Your information has been widely grasped, your needs are also widely grasped, each person is unique, the future production is also personalized customization, the service is also personalized.

Eighth, sharing. Sharing is the development of the economy, so all that can be shared will be shared, and you should learn to maximize the value of the advantageous resources you have with the sharing economy thinking.

Chen Xiaoping: "Three key issues for the landing of artificial intelligence industry in 2020-2035

Chen Xiaoping, Director of Robotics Laboratory, University of Science and Technology of China, shared the theme of "Three key issues for the landing of AI industry in 2020-2035".

(Chen Xiaoping shared the theme of "Three key issues for the implementation of artificial intelligence industry in 2020-2035")

Chen Xiaoping mentioned several typical paths of AI industry landing: completely enclosed, semi-enclosed, partition method/flexibility, and acceptance view. About the first landing path: completely enclosed, for example, the automobile production line is completely enclosed, this technology is not only in the automobile production line, in many other industries are beginning to apply. The second landing path: semi-enclosed, high-speed rail is an example, high-speed rail route is completely closed, there are guardrails, but the station platform, the marquee this is not enclosed, this is semi-enclosed, this example is very inspiring to us, we have not seen examples of large-scale industrial applications with semi-enclosed, this is a very good opportunity. About the third landing path: the partition method / flexibility. Flexibility is a very complex production process, or workflow, where part of the unit is manual and the other units are intelligent.

Chen Xiaoping said, looking ahead to 2020-2035, artificial intelligence industry landing need to solve the key problem is closed and acceptance view. Closure - so that AI technology better into the industry; acceptance view - so that artificial intelligence better able to approach the people.

Xiao Jing: "Artificial Intelligence Empowering Finance and Healthcare

Xiao Jing, chief scientist of Ping An Group and rotating president of Shenzhen Artificial Intelligence Industry Association, brought the theme sharing of "Artificial Intelligence Empowering Finance and Healthcare".

(Xiao Jing shared "Artificial Intelligence Empowered Finance and Healthcare")

About artificial intelligence technology + finance

Xiao Jing said that risk control is the most core aspect of finance, risk control in the financial field is particularly diverse, a single document may have fraud, each document looks normal, converge to a person, this person has fraud, or this enterprise has problems, to the terminal industry may have risk, to the macro market has international risk and so on, so the risk is very diverse, very complicated, also need to have a large number of technology to help us do better, to control the risk well, if only rely on human resources is difficult to do.

In addition, from the market environment, the financial industry is also facing many changes, such as increasingly stringent regulation, more and more diverse financial models, more and more open, and so on.

In this market environment, we need to make a lot of changes to our business model, and we need to operate in a mixed industry. In addition, we need to be asset-light, to refine our management model, to enhance our experience, and to flatten our organizational structure more and more. For example, refinement of management requires agile development; flat structure requires comprehensive talents, cross talents, who know both technology and finance; strengthening customer experience requires data analysis ability, through data analysis to better understand customer needs and better provide high-quality services. From these aspects, we can see that technology is the core of our response to market changes and pain points in the financial industry, which are very important.

Artificial intelligence technology is actually not easy to apply, artificial intelligence technology has many challenges, first of all, requires data, algorithms, arithmetic, in addition to the three elements of technology is not enough, but also a very good understanding of the scene, business, to be able to really implant the technology inside the business process, the business process to do transformation, so there should be business experts involved, the cooperation of business experts, and finally form a complete business solution. In addition, the program also needs to iterate in the real scene, only in this way, can really realize the so-called "three mentions and two reductions", mention the efficiency, mention the effect, mention the user experience, reduce the risk, reduce the risk.

About artificial intelligence technology + medical

Just now, we mainly talked about the financial field, and the following is about the medical field. There are many pain points from prediction, prevention and control, auxiliary screening, diagnosis, treatment and rehabilitation to patient management.

Ping An has created a whole process business solution, from predictive prevention and control to screening and diagnosis, auxiliary diagnosis to treatment and rehabilitation and patient management. During the epidemic, there are actually three platforms used in many places, including the medical imaging system for automatic detection of coronary pneumonia; the telephone system for telephone screening and robotic screening without human screening; and epidemic forecasting, including some psychological counseling. In Wuhan and Zhejiang, we have used robotic screening systems in many places, and robots do screening for community workers. This is an example of medical imaging. Through image recognition, medical image analysis technology can automatically do the detection of coronary pneumonia, and can track changes in the condition of pneumonia, and can do follow-up assessment, detection and analysis, and can produce results and reports in ten seconds, with an accuracy of 97-98% or more.

The afternoon session mainly focused on the topics of AIoT, AI+ business application, AI+ smart city, AI+ finance and other fields.

Cui Baoqiu: "AI+IoT, create a new era of intelligence together

In the first session of the afternoon, Cui Baoqiu, Vice President of Xiaomi Group and Chairman of the Group Technology Committee, firstly brought the theme sharing "AI+IoT, Creating a New Era of Intelligence".

(Cui Baoqiu shared the theme of "AI+IoT, Create a New Era of Intelligence")

Cui Baoqiu talked about: Alot is AI+IoT, the combination of artificial intelligence and the Internet of Things.

From IoT to Alot, we are actually experiencing a qualitative change, not simply a quantitative change, but a qualitative change. The first is the qualitative change in the interaction between people and devices; the second is the qualitative change from the individual to the whole, especially for Xiaomi users today he bought a set of devices, a series of devices, users buy a whole set of services, buy the overall intelligence, rather than individual intelligence. The third qualitative change is that in the Alot era is not only interconnection but human-centered intelligent services. So the intelligent ability is no longer a single device provider, but the cloud, edge and end as a whole.

The cell phone and Alot, called the dual engine, dual engine is a unique advantage of Xiaomi, the cell phone is the largest application platform for AI, many functions on the phone and AI are related. AI is everywhere on the phone, and then multiplied by Alot, so AI in Xiaomi everywhere has two layers of meaning: first, all AI technology in Xiaomi software, hardware, Internet, triathlon have a place. Second, all of Xiaomi's products need AI empowerment. The latest earnings figures show that Xiaomi's Alot platform is today the world's largest consumer lot platform, the number of devices connected to the world also has 270 million units, this figure does not include Xiaomi smartphones, nor Xiaomi notebooks, if you add these smart devices, Xiaomi has hundreds of millions of units of the end under its control. Today there are nearly 56 million households are using Xiaomi smart devices Xiaomi's smart device category is spreading rapidly.

There is a soul in the middle of the phone and Alot, the soul is called Xiaoxiaoyou, Xiaoxiaoyou is the control center of Xiaomi Alot, is its soul. So around Xiao Ai students to create a strong AI integrated capabilities, Xiao Ai students left to link a variety of smart devices, the right to link the content and services. In the cell phone multiplied by Alot strategy, AI is increasingly important, which is the experience Xiaomi has accumulated over the past 8 years or so, from cloud computing to big data, to AI has accumulated a whole set of technology stack.

Zhang Hongjun: "AI+X - The application of artificial intelligence on the ground in various scenarios

Ltd. and Zhang Hongjun, Chairman of Zhangjiang Artificial Intelligence Chamber of Commerce of Shanghai Pudong District Federation of Industry and Commerce, shared the theme of "AI+X - the application of artificial intelligence in various scenarios".

(Zhang Hongjun shared the theme of "AI+X - the application of artificial intelligence on the ground in various scenarios")

Zhang Hongjun: The most important application that Huina Technology has done in the past few years is called Huikei Cloud, which is a cloud platform that mainly has real-time traffic data from more than 2,000 offline shopping centers, and through the machine vision method, some sensors are installed to aggregate real-time traffic data to the platform, which basically counts more than 10 billion people per year. Of course, there is also a combination of other data, such as the basic information base of the country's major shopping malls, and brand tag library, and more than a million stores, store information, the data collected together, the aggregated platform is called Huikei cloud, through a large amount of data, the collection of offline data we combine with the commercial scene to do some data services, commercial scenes, mainly shopping centers, retail stores, some investment, site selection, etc. to do data modeling and analysis.

Huikei Cloud is mainly two parts, one part is called data collection. The second part is the data application, data collection is through machine vision, the installation of some sensors, all algorithms are in the end, here are a few scenes, now see this kind of we are in the edge of the computing, in the super each door, shopping malls each store door are equipped with sensors, it can real-time statistics location in and out of the number of people. We use artificial intelligence technology to do some data collection, the scene of these data collection and then do it analysis, mining, and then applied to the scene.

The biggest value of AI may lie in the openness, in the openness of the whole ecology, so all these AI scenes really produce data we think data connection, data flow, data sharing will produce more value. So what we have just seen is not the real Huikei cloud, Huikei cloud is through these little data collection, through each small scene eventually all the current national offline commercial traffic data called between people and the field, between people and commercial field data convergence into a platform. With such a platform and such a data connection, that is really what Huikei Cloud really wants to do. This time during the epidemic, we also did a national shopping center traffic rebound test.

Zhou Qiru: "Under the new situation, how AI can empower new scenes and new applications

Zhou Qiru, Director, Executive Vice President and General Manager of Artificial Intelligence Division of Shenzhen SaiWei Intelligence Co., Ltd. gave a keynote speech on "AI Empowers New Applications in New Scenarios". Zhou Qiru said, Sai as intelligent from the traditional integrator transformation of artificial intelligence enterprises, is to take the second time the determination and courage to fight, during the epidemic, Sai as intelligent artificial intelligence products also have new application areas, our inspection, temperature measurement plus face recognition linkage, etc., has been used in the Shenzhen Third People's Hospital, it is said that the feedback effect is good, the recognition rate is very high.

(Zhou Qi Ru made a "new situation, AI how to empower new scenes and new applications" theme sharing)

Since the country launched the "new infrastructure" policy, Saiwei Intelligence has responded positively, using 5G and IoT technology synergy, hoping to empower more traditional industries. In July this year, Sai for China Telecom provided a data room 5G unmanned inspection robot, it can effectively reduce the cost of late data room manual inspection in the case of 5G, after reducing the cost of operation and maintenance, but also timely response and firefighting to do linkage.

We believe that talent technology, innovation, industry, integration and application are very important, and we are very open to artificial intelligence technology and environment.

In the era of "new infrastructure", where is the future of artificial intelligence products, Saiwei Intelligence believes that in the short term, technology may face some ceilings, such as face recognition 99.999%, and face recognition rate of 99.998%, there is not much difference, but the 0.001 technology is very difficult to improve, so We believe that artificial intelligence will be mainly to empower the industry, in the form of auxiliary existence, and Saiwei intelligence has been hoping to artificial intelligence technology and product empowerment to the traditional industry. We believe that the future of artificial intelligence is reflected in "AI + technology", based on the core basic technology and R & D innovation, through the integration of multi-technology dispersion, resulting in a number of breakthroughs.

AI + scene, multi-scene empowerment, is the core direction of the future development of artificial intelligence industry, through the specific scene landing, implementation and application, the technology, products from core cities to small and medium-sized cities, from large cities to rural areas. In terms of market, we insist on the main domestic large cycle, forming a new development pattern of domestic and international double cycle and benign interaction.

Zhiyong Wang: "The Development and Application of Ant Cognitive Intelligence

Wang Zhiyong, Technical Director of Ant Group Artificial Intelligence Department, gave a keynote sharing on "The Development and Application of Ant Cognitive Intelligence".

(Wang Zhiyong shared the theme of "The Development and Application of Ant Cognitive Intelligence")

Wang Zhiyong talked about: artificial intelligence ant development for so many years, we think the whole development is actually a trend, that is, we from the original pure three-mode perception intelligence gradually to the direction of multimodal cognitive development. For example, artificial intelligence is like the old Chinese doctor, the old Chinese doctor may have the ability to look and smell, in fact, look and smell is a perception ability, you can look at your tongue, measure your body temperature, measure your heartbeat, which are perception ability, for intelligent body image recognition ability, language recognition ability, video understanding ability, natural language understanding are perception ability. It is not enough to have perception alone, at this time a doctor's good or bad depends on his specific medical experience, his clinical experience, so knowledge is one of the most important modalities, so there are knowledge mapping products to help him finally make decisions.

In addition, in addition to static perception, dynamic perception will be added, the doctor will also ask you, how is the recent situation, how many days has the flu, so will do modeling in the user behavior distinction, to help him make overall predictions, of course, will also use search technology, the whole technology application trend towards the integration of a variety of capabilities, which can not be separated from the support of the cognitive service platform AIDesk and knowledge mapping platform.

AIDesk architecture mainly includes two parts: model production capability and model service capability. The entire Ant Knowledge Graph platform is not quite the same as Baidu or Google general knowledge graph, but more of domain knowledge graph, including knowledge modeling, knowledge extraction and knowledge fusion. After constructing a good knowledge map, we can do knowledge inference and serve specific applications, such as enterprise knowledge map, wealth knowledge map, medical and health, etc. The knowledge map serves different business fields of Ant.