The following is the 2021 global AI index report: more citations in China than in the United States From Stanford recommended by recordtrend.com. And this article belongs to the classification: artificial intelligence.
Stanford University publishes the AI index report every year. Recently, the 2021 annual report has been released, which is also the fourth issue of the report, focusing on the changes in the AI field from 2019 to 2020.
Stanford’s AI index report includes policy, country analysis, technical report and other key parts All relevant data are tracked, collected, refined and visualized for a long time, aiming to provide a detailed, unbiased and carefully reviewed data report for the public, so as to help global policy makers, researchers, company executives, media reporters and other people better understand the latest technological progress and changes in the complex field of AI.
The report was compiled by the Institute for human centered artificial intelligence of Stanford University. The Institute is jointly led by Professor Li Feifei, a well-known AI researcher, and John etchemendy, Dean of Stanford University.
It is worth noting that this year’s AI index report points out some highlights of global AI development in 2020, including:
1) Last year, the number of citations of AI papers from Chinese researchers in journals surpassed the United States (19.8%) for the first time with a weak advantage (20.7%) and took the lead in the world;
2) The proportion of foreign graduates of AI direction in the United States and Canada increased by 4.3% to 64.3%, of which 81.8% chose to study in the United States;
3) 65% of doctoral graduates in AI field work in the industry, which is 20% higher than that in 2010, indicating that the industry plays an increasingly important role in the development of AI technology;
4) The total amount of AI financing has increased by 9.3%, but the number of AI start-ups that have completed new financing has declined for three consecutive years;
5) influenced by COVID-19, the AI private equity investment in pharmaceuticals increased by 4.5 times in 2020 compared with the previous year.
The number of citations of papers exceeds the United States for the first time in China
The authors found that in 2020, for the first time, China will surpass the United States in the number of journal papers cited. Previously, in 2004 and 2017, China surpassed the United States in the total number of journal papers published twice.
The data of this discovery comes from the Microsoft academic graph. As for the specific country definition method of this thesis, the author’s unit is used.
In other words, the assertion that “the number of citations of Chinese papers exceeds the United States for the first time” actually means that the number of citations of papers published in global journals by authors from Chinese universities, scientific research institutions and companies exceeds that of authors working for American university scientific research institutions and companies for the first time. This definition method is not limited to the author’s own country.
In other words, this finding in the AI index report is enough to show that China, as a country, has begun to show a slightly better strength in AI research than the United States.
Let’s take a detailed look at the published and cited data of the major AI research powers in the world in the past year. According to the OECD definition, all papers covered in the category of artificial intelligence or machine learning are counted.
In 2020, a total of 228000 AI papers will be published in journals, an increase of 34.5% compared with that in 2019
The total number of citations of these AI journals reached 937000, of which China accounted for 20.7%, surpassing the United States for the first time (19.8%). At the same time, the share of journal citations in the EU continues to decrease to 11%
However, the AI index report also pointed out that although the number of citations of journal papers dropped to the second place this year, the number of citations of conference papers published by the United States still leads China by a large margin. This advantage has lasted for ten years:
Moreover, in 2019, China’s total number of papers published in AI academic conferences is basically equal to that of the United States, but in 2020, it will decline sharply, accounting for 15%, lagging behind the United States by 18.9%
Other noteworthy data:
1 / in all the countries and regions with strong AI research strength, academic institutions still account for the highest proportion of AI journal papers, but the second source is beginning to show differences: in the United States, 19.2% of the papers published by industry companies and their related research institutions; in China and the European Union, 15.6% and 17.2% of the papers published by state-owned research institutions, respectively;
2/ COVID-19 has influenced the AI Academic Summit, and most of the meetings have been moved online. As a result, the number of registered participants has increased. Data shows that the total number of participants of the nine most popular AI top conferences almost doubled in 2020;
3 / the status of AI research in the whole academic industry has also been improved. In 2011, AI journal papers accounted for 1.3% of the annual published papers, and this proportion has increased to 3.8% by 2019.
A large number of international students are studying in the United States after graduation, and the industry is more sought after than the academia
Last year, we reported that the global AI talent tracker, a report by macropolo, a think tank under the Paulson foundation, found that the United States has attracted the world’s top AI researchers.
The vast majority of foreign talents who eventually engage in research in American universities or companies complete their master’s degree in the United States, and the most important source country is China. The report also points out that if we lose talents from overseas, especially from China, the advantage of AI talents in the United States will no longer exist.
Stanford University’s 2021 AI index report further strengthens this concept. The researchers found that the proportion of international students among the fresh doctoral graduates in AI direction in North America also reached a new high, reaching 64.3% in 2019, which is 4.3% higher than that in 2018, and the proportion is also higher than that in previous years, where international students are more keen on oriented majors, such as computer engineering, computer science and informatics.
The AI index report also found that 81.8% of the new year’s doctoral graduates in AI direction chose to work in the United States in 2019, and only 8.6% confirmed to go abroad for employment (another 9.6% were unknown)
Other computer related majors, such as network, software engineering, programming language, have been squeezed obviously, and the number of doctoral degrees awarded in these directions has decreased significantly. At the same time, as can be seen from the figure below, many directions that can be classified into the field of AI, such as machine learning (ML), robotics, computer vision, algorithms, etc., have increased the number of doctoral degrees awarded.
Among them, the number of doctoral degrees awarded in AI / ml direction and robot / computer vision direction increased the most, accounting for 22.8% and 7.3% of all AI fields. In 2019, nearly 300 AI / ml and nearly 100 robot / computer vision doctorates will be awarded in the United States.
There is another representative finding in this chapter: in the past decade, the proportion of doctoral graduates in AI field choosing to stay in academia has decreased year by year. In 2019, only 23.7% of doctoral graduates remain in the academic field, which is almost half of the level in 2010;
The proportion of doctoral graduates who choose to go to the industry to put what they have learned into practice or continue to work in the scientific research departments of top companies has increased significantly, from 44.4% in 2010 to 65.7% in 2019, with an increase of nearly 50%.
According to the report, this is mainly because the number of young scholars who choose to enter the academic field has remained stable, and most of the students in AI field who have increased sharply in the past decade have been absorbed by the industry.
After all, in the past decade, the application of AI in the industry has made a huge leap in both quality and quantity. These increased people may also be attracted by those top companies, or have already had a clear intention to apply for a job.
Not only do Ph.D. graduates prefer to go to the industry, but the AI index report also points out that the faculty level of American universities is also experiencing a very obvious brain drain.
A paper “artificial intelligence, education and entrepreneurship” published in 2019 found that from 2004 to 2018, the number of AI professors in American universities leaving their teaching posts to go to the industry increased significantly, which can be called an “unprecedented brain drain”.
This paper points out a very noteworthy phenomenon: the job hopping of AI professors in the industry, especially when they worship the recognized lifelong professors, and their teaching positions are filled by their own temporary professors or professors from lower ranking schools, will have a greater psychological impact on the doctoral students in AI, and will make them underestimate the significance of continuing to stay in the academic field.
In 2019, the overall number of AI professors’ job hopping is lower than that in 2018, but the situation is still worrying, because the temporary professors (green in the figure below) are more willing to stay in school, and the number of lifelong professors’ job hopping (purple in the figure below) is still a new high.
The following are the schools with the most serious loss of AI teaching staff from 2004 to 2018, with CMU, Georgia Institute of technology, University of Washington and UC Berkeley ranking the highest:
Other noteworthy data:
The world’s top universities are increasing investment in AI education. In the past four academic years, the number of courses that teach students the skills needed to build and deploy practical AI models has increased by 102.9% and 41.7% at undergraduate and graduate levels, respectively;
AI venture financing is difficult, it’s easier to connect with medicine
Based on data sources such as CrunchBase, the AI index report found that the amount of private investment in AI field increased by 9.3% in the past year (5.7% in 2019), with a total amount of US $40 billion.
But the number of AI start-ups that can complete a new round of financing is less than in previous years, that is to say, more money is put into fewer companies. This is also in line with the changes in AI investment in recent years. On the whole, the total amount of AI investment is increasing, but the growth rate has slowed down.
The report points out that one of the reasons is that in the year when the new coronavirus is rampant, investors are more willing to pursue hot spots and invest the money originally left to AI start-ups in various directions to a few companies specializing in drug discovery and design with the help of AI power.
As can be seen from the figure below, the private investment of AI drug design and development companies in 2020 is as high as US $13.8 billion, which is 4.5 times higher than that in 2019!
In addition, in order to avoid the spread of the virus, global school principals stopped face-to-face teaching for a long time. We can see that the investment in AI education and technology has also increased rapidly. (and, there’s also a huge increase in investment in games…)
Specific to the country, the data show that the active degree of private equity investment of American AI companies is still far higher than that of China, which ranks second. The total amount of financing is close to 24 billion US dollars, more than double that of China.
In 2020, all investment activities for AI companies, including private equity, listing, acquisition and merger, will increase by 40% compared with that in 2019, and the total amount will reach 68 billion US dollars. But influenced by COVID-19, small companies are disproportionately affected. The report points out that M & A activities constitute the vast majority of global AI investment and financing activities in 2020.
In this year, there were also some high-profile M & A interactions, such as NVIDIA’s acquisition of mellanox and Capgemini’s acquisition of ALTRAN.
Other noteworthy data:
1/ Brazil, India, Canada, Singapore and South Africa are the 5 countries with the highest growth rate of AI recruitment in the world (2016-2020 years), and COVID-19 has not significantly affected the recruitment of AI talents in key countries.
2 / however, for the first time in the past six years, the proportion of AI related job recruitment in the United States in the world will decrease in 2020. In 2019, the number of AI jobs in the United States will be 326000, which will decrease by 8.2% to 301 000 in 2020. This may mean that the competitive pressure of AI practitioners may increase slightly in the future.
3/ McKinsey’s survey of AI companies and investors showed that half of the respondents said COVID-19 would not continue to invest in the AI field. 27% said they had increased investment, and less than 1/4 had reduced investment;
You can download the 2021 AI index report from Stanford Hai’s website or read the condensed version.
The producers also released all the data behind the report on Google drive. Address: https://drive.google.com/drive/folders/1YY9rj8bGSJDLgIq09FwmF2y1k_ FazJUm
White House: the number of AI papers published by China in 2014-2015 exceeds that of the United States. It is estimated that China’s AI technology will surpass that of the United States to dominate the global technology market in 2022. Tencent Research Institute: a comprehensive interpretation of the development of AI industry in China and the United States Global AI index report in 2018: AI and human life in 2030 (with download) AI2: it is estimated that China’s AI research will lead the world in 2030, and the number of AI patents in China will be 5.5 times that of South Korea Japan economic news: China’s AI patent ranking surpasses that of the United States in 2019 Spain’s national news: China is becoming a leader in AI, and its strength can compete with the United States. In 2019, China’s AI enterprises rank second in the world. Stanford: AI index report in 2017
If you want to get the full report, you can contact us by leaving us the comment. If you think the information here might be helpful to others, please actively share it. If you want others to see your attitude towards this report, please actively comment and discuss it. Please stay tuned to us, we will keep updating as much as possible to record future development trends.
RecordTrend.com is a website that focuses on future technologies, markets and user trends. We are responsible for collecting the latest research data, authority data, industry research and analysis reports. We are committed to becoming a data and report sharing platform for professionals and decision makers. We look forward to working with you to record the development trends of today’s economy, technology, industrial chain and business model.Welcome to follow, comment and bookmark us, and hope to share the future with you, and look forward to your success with our help.