“Can machines think?” In 1950, Alan Turing asked questions at the beginning of his famous paper computer and intelligence.

This question not only won Turing the pride of “father of artificial intelligence”, but also awakened the development of artificial intelligence in the next 60 years.

Artificial intelligence (AI) is a technology science that researches and develops the theory, method, technology and application system for simulating, extending and expanding human intelligence.

Just like the wheel and iron in BC, the railway and electric power in the 19th century, and the automobile, computer and Internet in the 20th century, artificial intelligence is a new general technology, and its development will bring fundamental changes to human society.

It is generally believed that AI has experienced two generations so far. The first generation of AI is knowledge driven, and the overall progress is limited; The second generation of artificial intelligence is data-driven, that is, the hot big data, deep learning and so on, which has become a strategic technology highland for many scientific and technological powers.

It is understood that all the major developed countries in the world regard the development of artificial intelligence as a major strategy to enhance national competitiveness and safeguard national security, step up the introduction of plans and policies, and strengthen the deployment around core technology, top talents, standards and norms, so as to grasp the leading power of development in the new round of scientific and technological revolution. For example, Japan issued the AI technology strategy in 2017, the EU issued the AI strategy in 2018, the US launched the AI initiative in 2019, and South Korea announced the AI national strategy in 2019.

“Artificial intelligence is the commanding point of science and technology. Whoever can master it will grasp the great advantages of economic and social development. Therefore, China must make a breakthrough in artificial intelligence as soon as possible and strive for our position in the world.” Yao Qizhi, academician of the Chinese Academy of Sciences and President of the cross Information Research Institute of Tsinghua University, said.

China leads the world in application landing

Artificial intelligence started in 1978 in China.

After years of continuous R & D layout, especially since the promulgation of the new generation of artificial intelligence development plan in 2017, artificial intelligence has become a national strategy. China’s artificial intelligence has entered a new stage of rapid development, and has made important achievements in many fields. Breakthroughs have been made in key core technologies in some fields, which has global influence.

In terms of basic theory, China has made breakthroughs in the emerging deep learning theory and reasoning algorithm, brain like computing, brain computer interface and other basic frontier fields, and made significant achievements in some key technology fields such as intelligent chip. Huawei’s “shengteng”, deep learning processor chip “Cambrian”, Tsinghua University’s reconfigurable chip have reached the world’s advanced level.

In terms of key technologies, China keeps up with the world’s leading edge in machine translation, autonomous driving, intelligent robot and other technologies, achieves breakthroughs in some key technologies, and takes the lead in the fields of face recognition, speech recognition and generation.

At the same time, China’s artificial intelligence is accelerating the integration and development with various industries and fields, and artificial intelligence technology is gradually infiltrating from the Internet application to the real economy and people’s livelihood. Based on the data perception ability of the Internet of things and the intelligent computing ability from cloud to terminal, the application of artificial intelligence industry in China has been blooming, and has achieved remarkable results in intelligent medical, intelligent city, intelligent logistics, intelligent transportation and intelligent environmental protection.

“At present, the number of artificial intelligence enterprises in China is the second in the world, the financing scale is the largest in the world, and the number of patent applications is the first in the world, especially in the application landing. It can be said that China has become one of the important leading countries of artificial intelligence in the world. ” Zhao Zhiyun, Secretary of the Party committee of the China Institute of science and technology information, commented.

Zhao Zhiyun believes that with the rapid development of global AI, China’s advantages in strategy, policy, data and market application have created important conditions for the leapfrog development of AI.

First, strong strategic guidance and policy support. Zhao Zhiyun said that the promulgation of the new generation of artificial intelligence development plan means that the strategic deployment for the development of artificial intelligence in China has taken shape. Since then, various departments and localities have actively promoted the implementation of the plan. Many provinces and cities, such as Beijing, Shanghai, Tianjin and Chongqing, have issued corresponding artificial intelligence plans and action plans, increased R & D investment, established R & D institutions, and set up research and development projects Supporting policies such as talent introduction and tax incentives should be formulated to drive enterprises to speed up the pace of intelligent transformation, and the pattern of promoting the development of artificial intelligence through the cooperation of government, industry, University, research and application is taking shape.

Second, massive data resources. China has the largest number of 989 million Internet users and 986 million mobile phone users in the world, with 853 million mobile payment users. Especially in the specific application field, the data scale is huge. For example, the total number of medical outpatients in China is more than 8 billion every year“ Such a large amount of data is the only one in the world, which also provides rich resources for the development of artificial intelligence technology in China. ” Zhao Zhiyun said.

Third, rich application scenarios. China has the largest and more mature Internet market in the world, and the application space of artificial intelligence in the field of Internet is very broad. China has the most complete industrial chain in the world, and each subdivision field is facing the demand of transformation and upgrading, which has a huge demand for artificial intelligence empowerment. With the acceleration of new urbanization and the continuous expansion of urban scale, there is great potential to use artificial intelligence to improve urban infrastructure and improve the level of urban governance.

Fourth, the rapid growth and accumulation of young talents. China’s key colleges should speed up the layout of AI colleges and expand the scale of undergraduate and graduate education. Among the top international conferences and academic journals related to artificial intelligence, Chinese young scholars have become one of the most active groups. The number of AI start-up enterprises and unicorn enterprises led by young people is increasing, and the number of young scholars returning from overseas is increasing significantly.

Weak foundation affects strategic prospect

There are also weak links in the development of artificial intelligence in China.

First, there is a big gap between the basic theory of artificial intelligence and the original algorithm. Although the number of high-quality papers in China has increased significantly in recent years, the top papers and major theoretical innovations are still dominated by the United States, the United Kingdom, Canada and other countries.

This means that there are less basic innovations from 0 to 1 and more application innovations from 1 to N in the field of artificial intelligence in China“ Although it has branches and leaves, its roots are not in China. ” Zhao Zhiyun said.

Zhao Zhiyun believes that the weakness of core algorithms and open source systems is one of the most prominent technical bottlenecks in the field of artificial intelligence in China, which leads to the lack of new major achievements and original theoretical contributions such as deep learning model and generation of confrontation network in China, and the insufficient layout of general open source algorithm platforms such as machine learning, The development of the industry mainly depends on the open source code and system framework of international giants.

According to Yao Qizhi, the biggest pressure for the development of artificial intelligence comes from basic research“ We all know that the application of artificial intelligence in China can definitely catch up with many places in the world, or even lead the world, but our basic research on artificial intelligence is still in a relatively short stage. So we must cultivate talents, give them a good environment and encourage them to engage in basic research. This is the only choice to promote the future breakthrough of artificial intelligence. “

Chen Yunji, a researcher at the Institute of computing technology of the Chinese Academy of Sciences, pointed out in his article “intelligent computing system: a systematic course for the major of artificial intelligence”: “the more research on the upper layer of artificial intelligence (algorithm layer and application layer), the more contributions Chinese researchers make to the world; The more the bottom layer (system layer, chip layer), the less contribution of Chinese researchers. In various Imagenet competitions, the algorithm models of many institutions in China have shown a trend of “dominating the list”, which can be said to represent the world’s leading level. But most of these algorithm models are developed on the basis of CUDA Programming Language, tensorflow programming framework and GPU. Among these “hard technologies” at the bottom, Chinese researchers have made relatively little contribution to the world. The lack of research capability at the bottom not only delays the basic research of artificial intelligence in China, but also makes China’s intelligent industry become a castle in the air, and embarks on the old road of information industry restricted by core chips and operating systems. “

Second, the foundation of high-end chips, key components and high-precision sensors is weak. It is understood that NVIDIA, Qualcomm, Intel and other international giants still monopolize the global high-end chip business, especially the mergers and acquisitions among major manufacturers in 2020, which further makes the initiative mastered by western developed countries. These factors lead to the weak foundation of China’s key equipment, high-end chips, major products and systems, basic materials, components, software and interfaces, and the monopoly of hardware technologies such as graphics processor, application specific integrated circuit and field programmable gate array in Europe and the United States.

Third, it failed to form an AI innovation ecology with international influence. Zhao Zhiyun said that international giants have led the construction of innovation ecology by establishing an open platform of artificial intelligence and opening up the hardware system industry chain. Although China’s national AI open innovation platform for specific fields has achieved initial results, it is still insufficient in terms of general open source algorithm platform for machine learning, and its driving force and international influence on the industrial chain need to be further improved.

Behind these three weaknesses, the biggest bottleneck is the lack of high-level talents. According to the AI development report 2011-2020 jointly released by the AI research institute of Tsinghua University and the joint research center for knowledge intelligence of Tsinghua Chinese Academy of engineering, the number of high-level AI scholars in the United States is the largest, with 1244 people, accounting for 62.2%. China ranks second after the United States, but only 196 people, 8%.

“We still have a gap in the super first-class scientific research team.” Cao Heping, Professor of School of economics of Peking University and President of Shenzhen Bay District digital economy and Technology Research Institute, said, “we can’t solve problems only when there are problems. Instead, we need to prepare a group of strategic and innovative scientists. They take the problems that have appeared and will appear, and form their thoughts in advance. Then they concretize these thoughts into problems, form intellectual patents and set out in the laboratory. Then it will connect with incubators and accelerators in industrial parks in the big national economy system to form industrial capacity. “

Technology leading and application driving

In view of this, experts suggest that in the future, China should adhere to the strategic guidance of science and technology leading and application driving, strive to enhance the ability of scientific and technological innovation, comprehensively promote the application of artificial intelligence, and realize the goal of making up for the shortcomings in theory, being self-control in technology, and occupying the high point in industry as soon as possible through the two-way development of science and technology leading and application driving.

The first is to enhance China’s AI technology innovation capability as a whole. We should increase the steady and continuous support for basic research in the field of artificial intelligence, promote the cross integration of basic disciplines such as artificial intelligence and mathematics, and support the original and non consensus exploratory research. We should concentrate our efforts on tackling key and core technologies, guide and organize superior forces to make great efforts to solve the “neck sticking” problem. Accelerate the construction of national strategic scientific and technological strength in the field of artificial intelligence, and strengthen the integration and layout of relevant innovation bases such as national artificial intelligence laboratory and State Key Laboratory. We should grasp the major window of opportunity for the transformation of artificial intelligence technology in time, accelerate the implementation of major projects such as scientific and technological innovation 2030 – “new generation of artificial intelligence”, and solve the major technological needs for the intelligent upgrading of China’s economy and society.

Second, promote the application of artificial intelligence scene on a large scale. Strengthen the dominant position of enterprise innovation, deepen the cooperation between industry, University and Research Institute, improve the adaptability of artificial intelligence technology in different real industrial environments, and constantly improve the technology maturity and practical performance. By further promoting the construction of a new generation of national open innovation platform for artificial intelligence, we should give full play to the leading role of leading enterprises and research institutions in the artificial intelligence industry, encourage the open source of all kinds of general software and technologies, and create a more perfect technological innovation ecology. Give full play to the local initiative to promote the development of artificial intelligence, strengthen the application demonstration of artificial intelligence, and comprehensively enhance economic innovation and international competitiveness.

Third, we should continue to take the construction of talent team as the top priority in the development of artificial intelligence. Adhere to the combination of training and introduction, improve the artificial intelligence education system, expand the enrollment scale of graduate students, strengthen the talent reserve and echelon construction, open up special channels, implement special policies, and realize the accurate introduction of high-end artificial intelligence talents.

Fourth, strengthen the ethical governance of artificial intelligence. Artificial intelligence has the characteristics of high integration of technical attributes and social attributes. We should focus on the possible risks and challenges of artificial intelligence, strengthen the research of artificial intelligence in law, safety, employment, ethics and government governance, and guide the safe and controllable development of artificial intelligence.

Fifth, deepen the opening and cooperation of artificial intelligence. We should adhere to international open cooperation, carry out R & D cooperation around the global technical problems of artificial intelligence, jointly promote the development and governance of artificial intelligence, jointly formulate relevant international standards and ethical norms in the field of artificial intelligence, and actively respond to the global challenges that artificial intelligence may cause.

Editor in charge: CC

Leave a Reply

Your email address will not be published. Required fields are marked *