From the new trend of AI innovation and integration, the development of AI technology and soaring, China’s position in global AI and how AI reshapes various industries, it elaborates on the status quo and future of global AI development.
Editor’s note: This article is from WeChat public account “Smart things” (ID :zhidxcom). The original title “Deloitte Global AI Development White Paper Released – Eight New Trends, Three Key Technologies Examined”
At present, AI has achieved technology in many fields such as finance, medical care, security, etc., and the application scenarios are becoming more and more abundant. It is realizing all-round commercialization, which has triggered profound changes in various industries, which accelerates the digitization of enterprises. It has played an active role in improving the structure of the industrial chain and improving the efficiency of information utilization. At the same time, AI has also entered the era of machine learning. The future development of AI will be the combination of key technologies and industries. However, as the understanding of AI in the investment and business sectors has gradually deepened, the AI investment and financing market has become more rational, and although the investment amount has continued to increase, the frequency of investment and financing has declined. Especially after the round of survival of the fittest in the industry, the underlying technology startups and entrepreneurial projects such as medical, education and unmanned driving industries continue to be favored by leading institutions of artificial intelligence.
In this issue of intelligent internal reference, we recommend Deloitte’s research report “Global Artificial Intelligence Development White Paper”, from the new trend of AI innovation integration, the development of AI technology and soaring, China’s position in global AI and how AI reshapes each The four major aspects of the industry have elaborated on the current status and future of global AI development. If you want to collect the report of this article (global artificial intelligence development white paper), you can get the keyword “nc407” in the smart thing (public number: zhidxcom).
A new trend of AI innovation integration
1, artificial intelligence is fully commercialized
The current artificial intelligence technology has entered the stage of full-scale commercialization, and has different degrees of influence on the participants in the traditional industry, which has changed the ecology of various industries. This change is mainly reflected in three levels.
The first layer is enterprise transformation: artificial intelligence technology participates in enterprise management processes and production processes. The digitalization trend of enterprises is increasingly obvious, and some enterprises have realized more mature and intelligent applications. Such enterpriseIt has been able to collect and utilize multi-dimensional user information through various technical means, and provide targeted products and services to consumers. At the same time, through the optimization of data to understand the development trend and meet the potential needs of consumers.
The second layer is industry change: The changes brought about by artificial intelligence technology have caused fundamental changes in the relationship between the upstream and downstream of the traditional industry chain. The involvement of artificial intelligence has led to an increase in the types of upstream product providers, and users may also change due to changes in product attributes, from individual consumers to corporate consumers, or both.
The third layer is the human transformation. The application of new technologies such as artificial intelligence will improve information utilization efficiency and reduce the number of employees. In addition, the widespread use of robots will replace the labor force engaged in process work, resulting in an increase in the proportion of technology and management personnel, and changes in the human resources structure of enterprises.
▲All-round changes brought about by artificial intelligence technology
2, AI fully enters the era of machine learning
As technology advances and develops, the way humans learn to learn evolves from evolution, experience, and heritage to spread and store through computers and the Internet. With the advent of computers, human access to knowledge has become more efficient and convenient. In the near future, the vast majority of knowledge will be extracted and stored by machines. Powerful computer algorithms will gradually gain the ability of humanoids, including vision, ability to speak and sense of direction.
In the many branches of artificial intelligence, Machine Learning is one of the core research areas of artificial intelligence. Including 89% of artificial intelligence patent applications and 40% of artificial intelligence related patents are in the field of machine learning. The original research motivation was to make the computer system have human learning ability to achieve artificial intelligence. The machine finds vacancies in the existing knowledge, and then the machine follows the human brain and simulates evolution, systematically reduces uncertainty, identifies the same points of new and old knowledge, and completes learning.
▲Artificial Intelligence Levels
The artificial intelligence core is the algorithm. As the underlying logic of artificial intelligence, algorithms are a direct tool for generating artificial intelligence. From the perspective of historical process, artificial intelligence has experienced three stages since its introduction in 1956. These three stages are also the process of algorithm and research method change: The first stage is the 1960s and 1970s, artificial intelligence welcomes In the golden age, the research method led by logic has become the mainstream. Artificial intelligence uses computer to realize the logical reasoning of machine, but it is difficult to achieve. The second stage was in the 1970s and 1990s. Among them, between 1974 and 1980, the immature and prestigious reputation of artificial intelligence technology made it into the “artificial intelligence winter”, and artificial intelligence research and investment were greatly reduced.
From 1980 to 1987, expert system research methods became popular in artificial intelligence research, capital and research enthusiasm rekindled; from 1987 to 1993, computer capabilities have made considerable progress over the previous decades, when trying The problem was solved by establishing a computer-based expert system, but because of the small amount of data and too much limited to empirical knowledge and rules, it was difficult to construct an effective system, capital and government support were withdrawn again, and artificial intelligence ushered in the second “cold winter”.
The third stage is after the 1990s. From 1993 to 2011, with the increase of computing power and data volume, artificial intelligence technology has been further optimized. So far, the amount of data and computing power have been greatly improved. Help artificial intelligence to make great breakthroughs in machine learning, especially in the field of deep learning dominated by neural networks. Based on the development of deep neural network technology, it has gradually entered a period of rapid development.
▲The history of artificial intelligence technology
In addition, data is an indispensable supporting element in the underlying logic of artificial intelligence, and data processing without artificial intelligence will not be possible. With data mining methods such as clearing, integrating, and reducing data, artificial intelligence can have enough data to learn. With the iterative update of artificial intelligence technology, it will be replaced by machines from data production, acquisition, storage, calculation, propagation to application.
▲Development stage of data processing
3, the market returns to rationality for investment
From the scientific and academic aspects to technology entrepreneurship, artificial intelligence took only a few years. Such a change not only benefits people’s desire to liberate the productivity of new technologies and the support of policies, but also the boost of artificial intelligence in the capital market. With the deepening of the awareness of artificial intelligence in the capital market, the investment in artificial intelligence in the investment market is becoming more mature and rational. In the past five years, China’s investment in artificial intelligence has seen rapid growth. In the first year of artificial intelligence in 2015, the total investment reached 45 billion yuan, and continued to increase in 2016 and 2017. In the first half of 2019, China’s artificial intelligence field received more than 47.8 billion yuan of financing, and achieved good results.
▲Changes in China’s artificial intelligence investment and financing
Analyze the investment trend of artificial intelligence, which is mainly divided into the following points:
The easy-to-reach artificial intelligence application scenario is sought after by investors. In recent years, investment data shows that enterprise services, robotics, medical health, industry solutions, basic components, and financial sectors are higher in investment frequency and financing amount than other industries. At the company level, the world’s top teams, financial strength and technology genes are more popular with investors in the secondary market. From an industry perspective, new retail, unmanned, medical and smart adaptation education that is easy to reach indicates more opportunities, so companies in these areas have more opportunities to invest.
▲China’s artificial intelligence industry investment and financing frequency distribution
The investment market is beginning to favor the underlying technology startups. Different from the previous investment preference for applied artificial intelligence companies, the investment market began to pay more attention to the startups of artificial intelligence underlying technology. The underlying technology is more sought after, and such companies are more competitive in the market due to the high ceiling. Since the development of artificial intelligence underlying technology in China still lags behind that of the United States, and the underlying technology is an important support for the development of artificial intelligence, with the further development of artificial intelligence in China, the investment of the underlying technology will continue to grow.
The proportion of companies that have been invested in A and B is still the highest, and strategic investment is gradually increasing. At present, more than 1,300 artificial intelligence companies across the country have obtained venture capital investments. Among them, the proportion of the previous frequency of the A round has gradually narrowed, and the investors still have a high enthusiasm for the A round. Currently, it is the highest frequency of investment. Strategic investment began to break out in 2017. With the gradual maturity of the artificial intelligence market segment, leading companies dominated by Internet giants have turned their attention to strategic investments seeking long-term cooperative development. This also indicates that the strategic cooperation between the artificial intelligence industry and the industry at the capital level has begun to increase.
▲In the first half of 2013-2019, the artificial intelligence investment round
The giant investment artificial intelligence layout is upstream and downstream in the business related industry. In the craze of artificial intelligence development, the Internet-savvy Internet giant has also begun its own strategic layout. The technology investment funds supported by the Ministry of Science and Technology, the National Science and Technology Institute of the Chinese Academy of Sciences, the Local Finance Bureau and the Economic and Information Committee, as well as the Internet giants led by Alibaba, Tencent, Baidu and Jingdong have already infiltrated the investment into various sectors of artificial intelligence.
From the perspective of the field, the projects selected by various investment institutions are in the upstream and downstream of their future industrial strategic layout, and these projects have also promoted the development of the national artificial intelligence development strategy. For example, Alibaba’s investment focus is mainly on security and basic components, and representative companies that have been invested include Shang Tang, contempt and Cambrian technology. The focus of Tencent’s investment is mainly in the fields of smart health, education, smart cars, etc. Representative companies include Weilai Auto and Carbon Cloud Wisdom. The focus of Baidu’s investment is mainly in the fields of automobile, retail and smart home. Jingdong Investment focuses on the fields of automobile, finance and smart home.
The National Department of the Chinese Academy of Sciences is involved in artificial intelligence technologies and applications in the fields of chips, medical care, and education. With digitization in each lineIn the transformation and integration of the industry, artificial intelligence will become a must for giants in many fields such as driverlessness, medical health, education, finance, and intelligent manufacturing.
▲Main investment areas of AI leading companies
As a new industry of the future, artificial intelligence companies are characterized by high growth. Based on incomplete public information and the growth rate of the artificial intelligence companies in the Deloitte Technology Fast 500 list, we screened 50 high-growth companies.
4, the city has gradually become the main battlefield of AI innovation and fusion application
The city is a comprehensive carrier that carries the fusion application of AI technology innovation, and it is also a concentrated experience where human and AI technologies generate comprehensive awareness. In the past few years, major cities around the world have played a different role in the development of AI technology, built their own ecosystems, and achieved initial results in the application of enabling industries and the development of regional economic development. Thinking, cognition and action of a round of industrial revolution. As AI applications have landed at the city level, the city has gradually become the main battlefield for AI innovation integration applications.
Although the key success factors of AI technology vary from place to place, in general, an ecological development system is conducive to the integration of technology and city. We have more than 50 AI technology segmentation application industries, more than 100 AI technology-related universities and research institutions, more than 200 head companies, more than 500 investment institutions, 7,000 AI companies, and 100,000 core talents in the AI field. Continue to follow and observe the characteristics, framework and development path of AI technology and industrial ecosystem with city as the main body. After comprehensive consideration, we believe that the degree of integration of a city’s AI technology innovation can be mainly through the following five aspects:
Top design: AI industry support policy, special legislation, data open policy and openness;
Algorithm breakthrough: the core of R&D of artificial intelligence core software and hardware such as AI chip;
Quality of the element: AI leader, capital support, scientist compensation level, industry conference influence, etc.;
Converged quality: frontier subject connectivity (AI: +Cloud, +Blockchain, +IoT, +5G, +Quantum Computing and other cutting-edge technologies), innovative subject diversity (head business, academic institutions, etc.), cultural diversity, etc.;
Application quality: financial, education, medical, digital government, medical, driverless, retail, manufacturing, integrated carrier development, etc.
According to the assessment of global cities in the above five indicators, Deloitte selected the most representative of the three categories of 20 global AI innovation integration application cities:
▲20 global AI innovation integration application cities in 2019
5, AI support system continues to exert strength
As a “troika” to advance the advancement of artificial intelligence technology, algorithms, data and computing power have been innovating over the past 5-10 years. In terms of algorithms, humans have made breakthroughs in machine learning algorithms, especially in visual and speech technology. In terms of data, the advent of the mobile Internet era has brought explosive growth to the volume of data.
The artificial intelligence algorithm model has been developed for a long time and currently covers multiple research sub-areas. Taking machine learning as an example, its core algorithms include least squares, K nearest neighbors, K-means, and PCA analysis. The core models include linear regression, logistic regression, decision trees, clustering, support vector machines, and so on.
The mainstream algorithm model library enables the efficient implementation of common algorithm models: Caffe framework, CNTK framework, etc. are collected and integrated for different algorithm models, and have high practicability in the development and utilization of algorithms. With the continuous improvement of big data technology, the cost of tag data that artificial intelligence relies on is reduced, and the processing speed of data is greatly improved. The efficiency of broadband is increasing. The continuous iteration of IoT and telecommunications technologies provides the infrastructure for the development of artificial intelligence technology. In 2020, the number of devices connected to the Internet of Things will increase to 50 billion units. The development of 5G, which represents the development miles of telecommunications, will provide the fastest 1Gbps information transmission speed for the development of artificial intelligence.
In terms of computing power, the parallelism of chip processing power and hardware price reduction has greatly increased the computing power. Up to now, the computing power of global artificial intelligence is mainly based on GPU chips. However, with the continuous iteration of technology, the computing unit category such as ASIC and FPGA will become the underlying technology to support the development of artificial intelligence technology.
▲ China’s artificial intelligence chip market size and growth forecast (2016-2020)
6, top-level policy tilting strength continues to increase
The impact of artificial intelligence on social and economic development has become increasingly prominent. Governments have also introduced policies on the development of artificial intelligence and raised them to the height of national strategy. Up to now, many countries and regions including the United States, China and the European Union have promulgated national-level artificial intelligence development policies.
▲Countries’ policies for artificial intelligence
As of 2019, the Chinese government continues to support the development of artificial intelligence in various forms. Previously, China formed a joint artificial intelligence promotion mechanism involving the Ministry of Science and Technology, the National Development and Reform Commission, the Central Network Information Office, the Ministry of Industry and Information Technology, and the Chinese Academy of Engineering. Since 2015, it has issued a number of policies to support the development of artificial intelligence, providing a large number of project development funds for the development of artificial intelligence technology and landing, and providing support for the introduction of artificial intelligence and enterprise innovation. While these policies provide a solid policy orientation for industry development, they also send positive signals to capital markets and industry stakeholders. In terms of promoting market applications, the Chinese government has taken the initiative to directly purchase related products for the application of domestic artificial intelligence technology, and has successively landed several smart cities and smart government projects.
From a strategic perspective, the “New Generation Artificial Intelligence Development Plan” is China’s first system deployment document in the field of artificial intelligence, specifically the overall thinking, strategic goals and tasks of China’s new artificial intelligence development in 2030, Safeguards for system planning and deployment. According to the current development status of China’s artificial intelligence market, the plan puts forward the requirements for the development of the basic layer, the technical layer and the application layer, and establishes the “three-step” development goal of China’s artificial intelligence in 2020, 2025 and 2030.
7, the global AI market exceeds 6 trillion US dollars
Artificial intelligence will increase social labor productivity, especially in effectively reducing laborDynamic costs, optimized products and services, new markets and employment have revolutionized human production and life. More and more governments and organizations around the world are recognizing the economic and strategic importance of artificial intelligence and are involved in artificial intelligence from national strategies and business activities. The global artificial intelligence market will experience phenomenal growth in the coming years. We predict that the world artificial intelligence market will exceed US$6 trillion in the next 2025, and the compound growth rate will reach 30% in 2017-2025.
▲Global Artificial Intelligence Market Size
In terms of industry, the large-scale areas of the traditional market will continue to lead. In 2030, manufacturing, communications, media and services, natural resources and materials will occupy the top three with 16%, 16% and 14% respectively. Among them, the huge manufacturing enterprises have begun to accelerate the digital transformation, and promote intelligent management, intelligent factories, intelligent logistics and other all-round intelligence, so the manufacturing industry is also the fastest growing area. At the same time, in the new field, the application of artificial intelligence technology in education has begun to penetrate into the whole process of learning, and the growth rate cannot be ignored.
▲The size of the artificial intelligence market (by industry)
The scale of China’s artificial intelligence core industry has exceeded 100 billion yuan, and it is expected to grow to 160 billion yuan by 2020, driving the scale of related industries to exceed 1 trillion yuan. Among them, the scale of artificial intelligence related industries in Beijing, Shanghai, Zhejiang, Jiangsu and Guangdong is in the forefront of all provinces and municipalities, and is expected to reach 140 billion, 130 billion, 270 billion, 100 billion and 280 billion respectively in 2020.
In Shanghai, for example, since the launch of the “Implementation Opinions on Promoting the Development of a New Generation of Artificial Intelligence in the City”, the development of the artificial intelligence industry has accelerated. In 2019, the scale of related industries can reach 120 billion yuan. Relying on the location advantage of the Yangtze River Delta, Shanghai Artificial IntelligenceEnterprises can obtain sufficient and high-quality resources in terms of talents and capital. The benefits brought by enterprise clusters are significantly improved, which is conducive to the continuous expansion of the company and industry scale.
8, Beijing-Tianjin-Hebei, Yangtze River Delta, Pearl River Delta AI Enterprise Collection
After entering the commercial application stage, artificial intelligence technology has been gradually applied in many industries. Its development prospects are widely recognized by the government, enterprises and other social parties, and it has undoubtedly become an important force affecting economic development.
In order to promote industrial upgrading and realize the transformation of new and old kinetic energy, local governments have promulgated industrial planning guidance opinions related to the artificial intelligence industry, providing tax incentives, capital subsidies, talent introduction, and optimizing government processes to optimize the business environment and attract Strong companies settle in and cultivate local artificial intelligence companies.
▲The distribution of Chinese artificial intelligence companies
Under the dual forces of policy and capital, the number of artificial intelligence enterprises has risen rapidly. According to incomplete statistics, there are more than 4,000 artificial intelligence enterprises in China. The Beijing-Tianjin-Hebei, Pearl River Delta and Yangtze River Delta are the most intensive artificial intelligence enterprises. area. At the same time, due to the large number of traditional manufacturing industries that need to use artificial intelligence technology for intelligent upgrades, coupled with government policy support, the western Sichuan-Chongqing region has also become a gathering area for artificial intelligence enterprises.
From a city perspective, Beijing, Shanghai, Shenzhen, and Hangzhou are the cities with the largest number of artificial intelligence companies, with more than 600, in the first echelon.
Investment and financing amount: Beijing, Shanghai artificial intelligence start-ups have the largest amount of financing. The active capital environment will have a positive impact on supporting artificial intelligence start-ups to upgrade technology, acquire users, and expand the market, and promote the scale effect of upstream and downstream enterprises in the artificial intelligence industry chain, thereby enhancing the strength of urban artificial intelligence industry.
▲The amount of financing for artificial intelligence start-ups in each city (2015-2019 first half)
Start-ups take on the role of pioneers in the R&D and commercialization of new technologies. The amount of financing that start-ups receive represents to some extent the development prospects of new technologies in the region. Artificial intelligence technology has entered the commercial stage, and its application range has been extended to finance, transportation, medical care, manufacturing and many other aspects. The more financing of start-ups means that more funds will push artificial intelligence to penetrate more industries.
In terms of the amount of financing obtained by start-ups, since 2015, the amount of financing for artificial intelligence start-ups in Beijing and Shanghai has exceeded 50 billion yuan, respectively, to 155.9 billion yuan and 58.2 billion yuan. This is because Beijing and Shanghai gather most of China’s artificial intelligence start-ups. The company has strong technical strength, and customers have higher acceptance of new technologies, so they have a broader application market.
The difference between research institutes and institutions is obvious: Beijing is strong, Shanghai depends on universities, Shenzhen depends on enterprises, and Hangzhou is relatively single. Research institutes and institutions are important places for the development of artificial intelligence technology. The number of artificial intelligence papers in China surpassed that of the United States in 2014 and far surpassed that of other countries. This is inseparable from the rapid development of artificial intelligence research institutions and institutions. At the same time, research institutions and institutions are also the main force of artificial intelligence patent applications. Therefore, the analysis of artificial intelligence research institutions and institutions in various cities can help to understand the technical strength of the city.
▲Characteristics of artificial intelligence research institutions and institutions in each city
Artificial intelligence talent: agglomerate economically developed regions. Artificial intelligence competition is ultimately the competition of talents. The distribution of artificial intelligence talents in China is uneven, mainly concentrated in the Beijing-Tianjin-Hebei, Yangtze River Delta and Pearl River Delta regions. In addition, the central and western regions have also formed certain talent pools, mainly distributed along the Yangtze River. From the perspective of the proportion of artificial intelligence talents in various cities, Beijing has the most advantage, accounting for nearly 28%, which is twice the second place in Shanghai (12.1%). Shenzhen and Hangzhou accounted for less than 10%, ranking second.
▲The proportion of artificial intelligence talents in each city
▲The number of international papers published by various universities in artificial intelligence
City Intelligent Management: Affected by the policy, Shenzhen, Shanghai and Hangzhou are the first. The improvement of urban management efficiency under the framework of smart city mainly provides information support for smart city management by using information technology to realize the informationization of government affairs system, and then promote the intersection of data in various fields. Smart city management scores in Shenzhen, Shanghai and Hangzhou are higher. The informationization of these cities has started earlier, the digital divide has been greatly reduced, and departmental resource sharing, collaborative office and online approval have been widely realized. Due to the special status of Beijing, the government needs to have more considerations when implementing smart city management, so the ranking is relatively lower.
The development of artificial intelligence technology takes off
1, the key technologies of artificial intelligence are maturing
Artificial intelligence has progressed rapidly in the last decade, including machine learning, natural language processing, computer vision, and smart adaptation technology. According to data from Tsinghua University, computer vision, speech, and natural language processing are the three largest application directions in the Chinese market, accounting for 34.9%, 24.8%, and 21%, respectively.
▲Artificial Intelligence Application Technology Hotspot Ranking
Fast and mature computer vision technology: Computer vision is machine vision that replaces the human eye to identify, track, and measure objects. Computer vision has a wide range of applications, such as smart home, voice visual interaction, augmented reality technology, virtual reality technology, e-commerce search shopping, tag classification retrieval, beauty effects, smart security, live surveillance, video platform marketing, 3D analysis, etc. There are considerable advances in the aspects.
Technology giants in the fieldAnd unicorns gather, representative companies and research institutions including Baidu, Tencent, Hikvision, Tsinghua University, Chinese Academy of Sciences and so on. Baidu developed the face detection deep learning algorithm PyramidBox; Hikvision team proposed to predict the human body’s central axis instead of predicting the human body frame, to solve the problem of weak targets in pedestrian detection. Tencent Youtu and the Chinese University of Hong Kong team proposed PANet in CVPR2018. Based on MaskR-CNN, the underlying and high-level features are further aggregated. For ROI Align, the feature mesh corresponding to the candidate region is sampled at multiple feature levels. Pooling for merging operations facilitates subsequent predictions. In addition, Shanghai Cloud’s innovative computer vision companies, including Technology, Shenlan Technology and Qiu Niu, have leading technologies in computer vision.
The giant speech recognition technology: speech recognition allows the machine to automatically recognize and understand human language through signal processing and recognition techniques, and convert it into text and commands. Its application scenarios include smart TV, smart car, telephone call center, voice assistant, smart mobile terminal security, smart home appliances and so on. In terms of speech recognition technology, the recognition accuracy of mainstream platforms such as Baidu, Keda Xunfei and Sogou are all above 97%. At the same time, emerging startups, including Shanghai Yunzhisheng, have a place in the speech recognition industry. Keda Xunfei has a deep full-sequence convolutional neural network speech recognition framework, and the recognition accuracy of the input method reaches 98%. Sogou speech recognition supports dictation of up to 400 words per second. Alibaba Artificial Intelligence Lab has developed an artificial intelligence product for voiceprint shopping through speech recognition technology.
The technology of autonomous unmanned systems is on the horizon: Due to the continuous advancement of AI and machine learning, the technologies of unmanned vehicles, drones and medical robots have been significantly developed. The root cause is attributed to the algorithm of autonomous unmanned systems. support. Deep learning has proven to be an excellent ability to handle complex tasks. Modern computing devices, such as graphics processing units (GPUs) and computing frameworks such as Caffe, Theano and Tensor Flow, help designers and engineers build innovative, unmanned systems. Alibaba’s Artificial Intelligence Lab develops a bicycle intelligence system that achieves full-scale, all-weather centimeter-level positioning. Baidu’s driverless technology includes five core competencies: obstacle perception, decision planning, cloud simulation, high-precision map service, and end-to-end. Horizon has launched a deep learning processor IP for autonomous driving and its platform for autonomous driving. In terms of industrial applications, Shanghai Xijing Technology has explored in the field of unmanned cargo transportation.
Artificial intelligent adaptive learning technology is becoming more and more mature: as the most breakthrough technology in education, artificial intelligence adaptive learning (Intelligent AdaptiveLeThe arning technology (hereafter referred to as the intelligent adaptation learning) simulates the teacher’s one-to-one teaching process for students, and gives the learning system the ability to teach individually. Compared with the traditional one-person teaching method, the intelligent adaptation learning system brings the students a personalized learning experience, which improves the students’ learning input and improves the students’ learning efficiency. Smart adaptation learning technology has been used in the United States and Europe for more than a decade, and is used by a large number of users of all ages, accumulating more than 100 million users. Both product and technology are polished. Relatively speaking, the amount of data accumulated by the smart adaptation learning technology in China is slightly behind, and it is at the initial stage of development. The advantage is that China has a large population base and a fast development rate, and it is expected that the latecomers will be in the future. In China, the intelligent adaptation teaching enterprise represented by squirrel AI in genetic algorithm, neural network technology, machine learning, graph theory, probability map model, logistic regression model, knowledge space theory, information theory, Bayesian theory, knowledge tracking Theory, education data mining, and learning analysis techniques have all achieved technology accumulation.
2, the artificial intelligence open platform construction is steadily advancing
The vast industry and solution market is a major advantage in the development of artificial intelligence in China. In addition to the large number of search data, rich product lines and market advantages provided by a wide range of industries, the above-mentioned advantages are also driven by the technology giants at home and abroad to promote the open source technology community to help entrepreneurs at the artificial intelligence application level. Break through the barriers of technology and apply artificial intelligence technology directly to the research and development of terminal products. From the perspective of industry, artificial intelligence has been applied in many vertical fields such as medical, health, finance, education, and security.
With the accelerated commercialization of artificial intelligence technology, technology giants and emerging artificial intelligence startups have formed their own technological advantages. In order to make greater use of technological advantages to expand their business advantages and support the development of artificial intelligence industry, technology-leading artificial intelligence enterprises began to build their own artificial intelligence open platform.
The Artificial Intelligence Platform is a tool for building artificial intelligence applications. These tools combine intelligence, decision-making algorithms, and data to enable developers to create their own business solutions through the platform. Some artificial intelligence platforms provide pre-defined algorithms and simple frameworks. The artificial intelligence platform has the functionality of “Platform as a Service” (PaaS), which provides basic application development; some require developers to develop and program themselves. These algorithms can functionally support a range of machine learning related technologies such as picture recognition, natural language processing, speech recognition, recommendation systems, and predictive analysis.
The artificial intelligence open platform is built to create an artificial intelligence industry chain from source technology innovation to industrial technology innovation. The open platform connects the two ends of the industry chain. On the one hand it can connect developers and some research institutions. On the other hand, you can connect to many downstream companies, such as an imageThe identification-based artificial intelligence open platform can open relevant technical capabilities to entrepreneurial teams that want to open up business in the field of image recognition.
▲Open platform for technology and applications at home
In 2017, the Ministry of Science and Technology and other departments have fully investigated and demonstrated the first batch of national new-generation artificial intelligence open innovation platforms: relying on Baidu, Alibaba Cloud, Tencent, and Keda Xunfei to build autonomous driving, urban brain, Medical imaging, intelligent voice 4 national new generation artificial intelligence open innovation platform. In September 2018, the Ministry of Science and Technology relied on Shangtang to build a new generation of artificial intelligence open innovation platform for intelligent vision. From the current technology maturity point of view, education, retail government and other fields have already had an open platform for applications driven by core technologies:
Automatic driving national open platform. The “Automated Driving National Open Platform” is mainly based on the Baidu Apollo platform. It is an open, complete and secure software, hardware and service platform based on Baidu technology to help developers build a complete autonomous driving system. In August 2019, Baidu Apollo unmanned vehicle passed the Changsha test and completed the first demonstration of the high-speed scene auto-driving road between the L3 and L4 models in China. At this point, Baidu L4 level autopilot city road test mileage has officially exceeded 2 million kilometers,
Baidu’s Apollo open platform partners more than 120, covering all aspects of the industry chain, including automakers, component manufacturers, travel service providers, start-ups, communications companies, universities and local governments. Xiamen Jinlong, BMW and Daimler have cooperated with the Apollo platform. “Apollo” has been commercialized in Beijing Xiong’an, Shenzhen, Fujian Pingtan, Hubei Wuhan, and Kyoto, Japan.
The urban brain is open to innovation. Relying on Alibaba Cloud’s urban brain country artificial intelligence open innovation platform, based on the urban brain system, it will build an open and diversified ecological system for urban security, urban public services and other intelligent applications in various industries, for a new generation of artificial intelligence. Technology provides support services for innovative applications in all areas of the smart society. The algorithm system platform can optimize the large-scale visual computing platform. The full-time global traffic automatic patrol alarm system can provide real-time perception of traffic events and accidents in the city. The recognition accuracy rate is over 95%. The traffic flow forecasting system passes through the area.Historical and real-time video data, accurate and accurate prediction of traffic flow and flow of people in the entire area in real time.
The application deployment of the development platform is mainly in the transportation: urban unified data fusion engine, traffic flow forecasting system, large-scale data fusion control engine, and urban overall traffic situation detection system. At present, the project platform has provided thousands of proprietary cloud server computing resources to government customers in Hangzhou, Zhangzhou, Shanghai, Jiaxing, Macau and Kuala Lumpur, and supports real-time analysis and processing of massive multi-channel video data. The Urban Brain Algorithm team provides a variety of algorithm services such as image detection, recognition, and segmentation to public security, transportation, and municipal related customers. Taking Hangzhou City Brain as an example, Yinjiang Science and Technology cooperated with Zhejiang University to control the panoramic data of video, coil, microwave and Internet in real time, so that the traffic control experience of the traffic police and the traffic light management strategy of the city brain are superimposed in the urban area of Hangzhou. The effect in Xiaoshan District and Yuhang District is remarkable.
The medical imaging open innovation platform. Tencent’s AI image has been expanded from single disease to multi-disease, from early esophageal cancer screening to screening for lung cancer, diabetic retinopathy, breast cancer, colorectal cancer, cervical cancer, etc. AI axis The diagnosis platform can assist doctors in diagnosing and predicting more than 700 diseases, covering 90% of high-frequency diagnosis in hospital clinics.
Tencent has built a medical imaging innovation platform that is attended by medical institutions, research groups, device manufacturers, AI startups, information technology vendors, universities, and nonprofit organizations. The platform connects the core participants of innovation, entrepreneurship, industry chain cooperation, academic research, and HP Public Welfare, aiming to promote the national artificial intelligence strategy in the medical field. At present, the open platform for medical impact based on “Tencent Film” has reached cooperation with more than 100 hospitals in China, with a total of 106 million readings for hospitals and a total of 950,000 patients, suggesting 130,000 high-risk lesions. The medical record was 6.14 million copies.
Intelligent voice open innovation platform. The National Intelligent Voice Artificial Intelligence Open Innovation Platform is mainly based on the voice platform technology of the University of Science and Technology. A new artificial intelligence research center and data center were built. By the end of October 2018, the number of platform developer teams had exceeded 860,000, and more than 200 companies have settled in the platform. A complete artificial intelligence industry chain covering technology R&D, basic platform, Internet of Things, and intelligent hardware has been formed. At present, the national standards for leading and participating in six intelligent voice-related countries have been officially released, and the independent intellectual property rights and standards systems for intelligent voice technology and application fields have been established to form a sustainable innovation mechanism for production, education and research systems.
Technology’s intelligent voice core technology areas include: speech synthesis technology, speech recognition technology, machine translation technology, voice evaluation technology, and cognitive intelligence technology.