In This Turbulent Autumn, Who Can Use AI to Balance the World?
More and more countries have realized that the development of artificial intelligence technology will affect national competitiveness and productivity. Therefore, countries are also increasing their investment in AI development and joining this AI competition. At present, what is the strength of the three leading regions of the United States, China, and the European Union? Who will win this competition? The latest report released by the US CDI provides a comprehensive analysis of this.
Artificial intelligence is driving the fourth industrial revolution.
exist Gartner A forecast earlier this year indicated that by 2020,Artificial intelligence will eliminate 1.8 million jobs.But it will also create 2.3 million jobs;
Chairman of SoftBankMasayoshi SonAt the SoftBank Group's annual supplier and customer meeting on July 18, it was stated that artificial intelligence will change the way we live within 30 years.And urged Japanese universities to accelerate the training of AI talents;
Huawei predicted in its "Top Ten Trends for 2025" released on August 8 that by 2025, smart technology will penetrate every person, family and organization.97% All large companies will adopt AI technology.
As we can feel, AI is everywhere. For individuals, artificial intelligence provides convenience in all aspects of our lives;As for the country,AI can improve competitiveness, productivity, protect national security, and help solve problems in society, which will have a significant impact on national competitiveness.
Therefore, in recent years, more and more countries have begun to deploy AI strategies, and the competition in artificial intelligence has become increasingly fierce. Some media even call it an AI arms race.
How is the race going so far and who will emerge victorious?
AI race report: US leads, China second, EU lags behind
August 19, United StatesData Innovation CenterThe Center for Data Innovation released an analysis report on the development of artificial intelligence in China, the United States and Europe."Who Is Winning the AI Race: China, the EU or the United States?"
Report fromTalent, research, development, applications, data and hardwareSix categories compare the relative positions of China, the European Union, and the United States in the AI economy.
The report concluded: Overall,The United States leads in AI, China is second but catching up quickly, and the European Union lags behind.
According to the scoring method in the report,With a total score of 100, the United States scored 44.2 points, China scored 32.3 points, and the European Union scored 23.5 points.
Furthermore, the report notes that this order may change in the coming years.
To understand the relationship between each region’s AI prowess and its population size, the authors also calculated the score for each indicator by adjusting for the size of its workforce.
When controlling for size, the United States leads (58.2 percentage points), the European Union is second (24.3 percentage points), and China is third (17.5 percentage points).
Let's take a look at the rankings of the six indicators:
Talent:
Research:
develop:
application:
data:
hardware:
By comparing the strength of these six indicators, we can have a macro understanding of the AI competitiveness of the three regions.
For each category, the authors used multiple indicators to measure. For example, for research, AI Paper Quantity, Quality, and R&D IndicatorsThe calculation method is to calculate the score for each indicator based on the ratio. For example, in terms of hardware, China has 219 supercomputers, the EU has 92, and the US has 116. Therefore, China's score is 0.5 [(219/(219+92+116)], the EU is 0.2, and the US is 0.3.
Finally, sum the scores for each category to get the total score.
In the AI competition, what are China's shortcomings?
1. Talent: China is short of AI talent, while the United States is attractive
As the key to the development of artificial intelligence, AI talent has received much attention in recent years. As David Wipf, chief researcher at Microsoft Research in Beijing, said: “The future of AI will be a battle for data and talent."
The governments of China, the European Union, and the United States have announced or begun taking measures to improve and expand AI talent, such as adding new undergraduate majors in AI and investing more money in AI education.
Talent comparison indicator 1:quantity
People who have published journal articles or owned patents on AI-related topics between 2007 and 2017 were selected as statistical criteria.
In total, the number of AI researchers in the EU far exceeds that of the US and China. If calculated by population ratio, the US has 173 AI researchers per 1 million workers, the EU also has 173, and China has 23.
Talent comparison indicator 2: Quality
In addition to quantity, quality is also important. According to the h-index ranking, among the top 10% AI researchers in the world, there are about 5,787 in the EU (data for some EU countries are not released), 5,158 in the US, and 977 in China.
Talent comparison indicator 3: Number of authors of papers published in top AI conferences
According to statistics on the authors of papers published in 21 AI conferences in 2018, the United States is far ahead of the European Union and China with 10,295 authors. In terms of population ratio, the United States has 62 authors, the European Union has 19 authors, and China has 3 authors.
Talent comparison indicator 4: Number of AI education researchers (2018)
According to the 2017 h-index ranking, among the 20 companies with the most AI researchers, the EU has 85 researchers, the US has 232, and China has only 7.
A big threat facing the EU and China is thatAs a country with abundant AI talents, the United States continues to attract talents from other countries.Between 1998 and 2017, 1,283 foreign AI academic researchers came to the United States from abroad to take up U.S. industry positions.
Europe and China attracted only 834 and 58 such researchers, respectively.
In contrast,China lacks top AI talent.And only 25% of AI researchers have more than ten years of experience, while the proportion in the United States is close to 50%. This may also be because China started late in artificial intelligence.
But China is working hard to narrow the gap. Many universities have already strengthened the construction of AI disciplines and expanded the number of students enrolled. Tsinghua University has also establishedYao Class, Zhi ClassSuch an experimental class for top computer talents.
2. Research: China has more papers than the US, but the quality is the lowest
Although artificial intelligence has become a major trend, it is far from a mature technology. China, the European Union and the United States are all actively investing in AI research.
China's new generation of artificial intelligence development plan requires 2025Previously, China's Ministry of Industry and Information Technology planned to make a major breakthrough in artificial intelligence theory.Annual funding of US$950 million (approximately RMB 6.7 billion)Fund strategic AI projects.
The European Commission has pledged to 1.5 billion euros (12.8 billion yuan) to be invested between 2018 and 2020For artificial intelligence research, and plans to start from 2021 to 2027Invest at least 7 billion euros (about 59.9 billion yuan) in AI.
In September 2018, the Defense Advanced Research Projects Agency, part of the U.S. Department of Defense (DOD), announcedCostUS$2 billion (approximately RMB 14.1 billion) to develop next-generation artificial intelligence technology.
With great investment, various regions have also produced a large number of research papers to showcase their achievements. By analyzing the number and quality of papers and data on computer companies, the conclusion is:The United States leads, followed by the European Union and finally China.
Research comparison indicator 1:Number of papers (2017)
In 2017, the European Union published the most AI papers, followed by China, and the United States the least. However, in terms of population ratio, the United States still has the highest output.
Research comparison indicator 2:Paper quality
The competition of papers cannot be calculated only by quantity, but also by quality. The FWCI (Frequency Weighted Criteria) of the United States is 1.83, higher than that of the European Union (1.20) and China (0.94). The FWCI of China is 0.94, indicating that the frequency of citation of Chinese authors is lower than the average level of global AI authors.
Research comparison indicator 3: R&D expenditure
Number of software and computer service companies ranked in the top 100 in the world by R&D spending in China, the U.S. and Europe in 2018:
In 2018, there were 268 software and computer service companies in the top 2,500 in terms of total R&D spending. The United States, China, and the European Union spent $77 billion, $12 billion, and $11 billion on R&D, respectively.
Therefore, although China is not as good as the United States in research, it is catching up and already has higher quality research outputs.
A recent study of AI papers by the Allen Institute for Artificial Intelligence shows that by 2020 and 2025, the quality of AI research papers produced by China will surpass that of the United States.
3. Development - Strong US AI patents and AI acquisitions
existPatents and major AI company acquisitionsIn terms of AI, US companies performed strongly. For example, Microsoft and IBM each filed more patents than any other company in 8 of the 15 subcategories of machine learning, including supervised learning and reinforcement learning. Google and Microsoft also ranked in the top five in terms of AI patent applications.
Comparative Development Indicator 1: AI Venture Capital and Private Equity Funds (2017-2018)
Comparative Indicator 2: Number of Venture Capital and Private Equity Financing Transactions, 2017-2018
Comparative Indicator 3: Number of acquisitions of AI companies (January 2000 to May 2019)
Development Comparison Indicator 4: Number of AI Startups in 2017
Development comparison indicator 5: Number of AI companies (2019)
Development Comparison Indicator 6: Number of Highly Cited AI Patents (1960-2018)
4. Application - Chinese companies actively apply artificial intelligence
The governments of China, the European Union, and the United States have all publicly acknowledged the importance of AI applications.
In 2017, China’s Ministry of Industry and Information Technology released the “Three-Year Action Plan for Promoting the Development of the New Generation of Artificial Intelligence Industry (2018-2020)”, which calls for incorporating AI into the manufacturing industry.
US President Trump issued an executive order in 2019 calling for the development of technical standards for the application of artificial intelligence.
Application Comparison Indicator 1: Percentage of companies that applied AI in 2018 (successfully incorporated AI into their business processes)
Application Comparison Indicator 2: Percentage of Companies Piloting AI in 2018
Chinese companies and employees seem to realize the importance of leveraging AI, soLeading in AI applications.Artificial intelligence has gradually penetrated into Chinese culture.
On the contrary, the EU is still skeptical about artificial intelligence, and many employees have negative emotions towards artificial intelligence in the workplace.
5. Data: China has a demographic advantage
US and EU seek data sets that transcend borders
Artificial intelligence systems typically rely on large amounts of data for training. Large datasets help AI systems develop more highly accurate models.
In terms of AI data comparison,Individuals generate large amounts of data to study when they participate in various online and offline activities.For example, using search engines, posting information on social media, and purchasing items online. These activities generate data that is of great value to machine learning models.
In some cases, this gives China a natural advantage due to its larger population; however, it also suggests that the EU and US should seek to build population datasets that transcend national borders.
Data comparison indicator 1: Number of fixed broadband subscribers
Data comparison indicator 2: Number of people using mobile payment in 2018
6. Hardware: The United States is still far ahead, and China is catching up
AI systems rely on computers that can perform a large number of operations per second.Semiconductor equipment,For example, integrated circuits. In recent years, GPUs have catalyzed the development of AI. In addition,SupercomputerOther technologies also contributed greatly.
Hardware is very important for a country’s AI competitiveness. For example, the ZTE incident in 2018. ZTE almost went bankrupt at the time because it relied on semiconductor equipment from American companies.
Hardware comparison indicator 1: Number of top 15 semiconductor sales companies in 2019
Hardware comparison indicator 2: Number of top 10 semiconductor R&D companies in 2017
Hardware comparison indicator 3: Number of companies designing AI chips (2019)
Hardware comparison indicator 4: Number of top 500 supercomputers in 2019
Hardware Comparison Index 5: Overall performance of the top 500 supercomputers from 2009 to 2019
Of the 10 fastest supercomputers in the world, 6 are located in the United States.But figures from recent years suggest that in some respects, China has caught up with the United States and become the world leader in supercomputers.
It is the world's first supercomputer with a computing speed exceeding one exascale
In June 2010, 282 of the world’s 500 best-performing supercomputers were located in the U.S. However, in 2018, the number of U.S. supercomputers hit a record low, with only 109 making the top 500.
at the same time,Chinese AI chips are on the rise.Several technology companies, including Baidu, Alibaba, Tencent and Huawei, are developing AI-optimized integrated circuits.
In the past two years, Chinese AI chip startups have received at least $100 million in funding. Compared with the overall semiconductor market,China is in a more advantageous position in the competition in the artificial intelligence chip market.
The AI race is unpredictable. How can you dominate?
Through the analysis of these six aspects, we can see that in this AI competition, China, the United States, and Europe all have their own strengths and weaknesses. The author provides targeted "secrets to winning" on how to win the competition.
China:Leveraging the advantages of population base and data, we will focus on improving our ability to develop and retain local talent. We will encourage Chinese companies to set up R&D institutions abroad.
To cultivate students' creativity, the Ministry of Education has approved 35 universities to offer undergraduate majors in artificial intelligence. This number should be expanded, and like the UK, more funds should be spent to subsidize students and help more AI students obtain degrees.
In addition, the quality of papers and patents should be improved on the basis of quantity.
USA:The United States has benefited greatly from attracting talent. Therefore, the United States should focus on policies that encourage foreign talent to continue to immigrate to the United States and increase its domestic talent base.
EU:Compared with China and the United States, the EU has a smaller AI venture capital ecosystem. However, the EU is experiencing a serious brain drain in AI. The EU should also provide financial incentives to researchers to encourage them to stay in Europe.
In addition, the EU should use public data to amend the GDPR (General Data Protection Regulation), which will strictly restrict the development of their artificial intelligence.
Therefore, to win this competition, each region needs to learn from each other's strengths and overcome their weaknesses in order to change their position in the race. No one can make a final conclusion on the final result of the competition, so let's wait and see!
-- over--