Is there still a gap between China and the United States in the field of artificial intelligence?

Reference news network reported on February 8, 65438? According to Hong Kong media, China has made great efforts to develop artificial intelligence (AI) in recent years, and has become a big AI country, ranking second in the world in terms of both the number of talents and the number of enterprises. Only the United States is clearly ahead of China in all fields. It is worth noting that the United States accounts for 45.7% of the institutions of higher learning that study AI. If China wants to catch up with the United States, it must strengthen the training of talents in institutions of higher learning.

According to the website of Hong Kong Economic Daily (65438+February 7), the Fortune Global Forum opened in Guangzhou yesterday. The theme of this year's forum is "opening up and innovating to build a new economic structure". In recent years, the central government has promoted scientific and technological innovation to drive economic transformation, and artificial intelligence is a key link. In recent years, China's leading technology has also invested a lot of resources in the AI field, and has achieved results.

The report pointed out that the most important talent development in the AI industry, the "White Paper on Global Artificial Intelligence Talents" shows that the total number of talents in the global AI field is about 300,000, but the current market demand is in the order of one million, and there is a big gap in the supply of AI talents.

As of June 20 17, there were 26 17 artificial intelligence start-ups in the world. The United States ranks first with 1078, that is, American enterprises account for more than 40% of the global total, and China ranks second with 592 enterprises, followed by Britain, Israel, Canada and other countries. Among them, there are about 78,700 employees in US 1078 artificial intelligence start-ups and about 39,200 employees in 592 companies in China.

Specifically, among the 300,000 AI talents in the world, there are about 65,438+million universities and about 200,000 industries. At present, there are 367 universities in the world with AI research direction, but there are only about 20,000 graduates in AI-related fields every year. Among these 367 universities, there are 168 in the United States, accounting for 45.7% of the world, while there are only 20 in China.

The report believes that the industries where AI talents gather in China and the United States are not the same. Investors in the United States pay more attention to the basic layer, while investors in China pay more attention to the application layer.

In fact, in recent years, China's development in the field of AI is indeed at the forefront of the world, and its achievements are better than those of many developed countries. But compared with the United States, its foundation is still relatively weak, and there is still a long way to go to catch up with the United States. In particular, the research on AI in American universities accounts for nearly half of the world, and a large number of talents are added every year, which is beyond the reach of other countries, including China.

According to the report, it is more noteworthy that American investors and enterprises pay more attention to basic technology research. These achievements may not be commercialized immediately, but they will have a far-reaching impact on future technologies; China investors pay more attention to application, which is understandable at first, but don't be eager for quick success.

Profile picture: On February 2nd, 65438, the 4th world internet conference Internet Light Expo kicked off in Wuzhen, Tongxiang City, Zhejiang Province, where visitors learned about the intelligent floating platform.

Further reading American media claims that the United States uses artificial intelligence to analyze satellite photos to search for missile launch sites in other countries.

Reference news network165438+1reported on October 26th? According to American media, only a few well-trained analysts in intelligence agencies are engaged in searching for undeclared nuclear facilities or secret military locations from massive satellite images. However, deep learning artificial intelligence that allows Google and Facebook to automatically filter pictures of faces and cats may also prove invaluable in the tit-for-tat spy world. An early example is that American researchers trained a deep learning algorithm to find the China ground-to-air missile launch site, which was hundreds of times faster than manual analysis.

According to Wired magazine website165438+1October 2 1, the deep learning algorithm has been proved to be able to help people who have no previous experience in image analysis find ground-to-air missile launch sites scattered in the southeast of China, covering an area of nearly 90,000 square kilometers. This artificial intelligence based on neural network, that is, layered artificial neurons that can filter and learn from massive data, can rival the overall accuracy of 90% achieved by human image analysis experts in locating missile launch sites. Perhaps even more amazing, deep learning software has helped humans shorten the time to find potential missile launch sites from 60 hours to only 42 minutes.

Kurt Davis, director of the Geospatial Information Center of the University of Missouri and professor of electrical engineering and computer science, said: "The algorithm is used to find the location that is said to have a highly suspected missile launch site, and then manually evaluate the search results to ensure accuracy and find out how much time the algorithm saves."

According to the report, the study of the University of Missouri was published in the Journal of Applied Remote Sensing on June 6, 2008/KLOC-0. This research was conducted against the background that satellite imaging analysts were overwhelmed by the flood of big data. DigitalGlobe, a famous commercial satellite imaging company, generates about 70 megabytes of original satellite imaging data every day, not to mention all the imaging data of other commercial satellites and government spy satellites.

Davis and his colleagues showed how the existing deep learning model can train and improve satellite image analysis, and can find potential targets of interest to intelligence agencies and national security experts. These deep learning models-including GoogleNet and Microsoft Research's RESNET-were originally established to find targets and classify them from traditional photos and video information. Davis and his colleagues adjusted these models to adapt to the difficulties and limitations of interpreting satellite imaging data, such as training some deep learning models that can explain color and black-and-white images, so as to be used when only black-and-white images of ground-to-air missile launch bases can be obtained.

In fact, analysts rely extensively on satellite imaging technology to track the development of North Korea's weapons program. In this relatively small country, human analysts may have discovered most or even all existing ground-to-air bomb launch sites. But similar deep learning tools can help automatically mark new ground-to-air launch sites in North Korea or other countries. Knowledge of existing and new missile launch sites can sometimes lead analysts to find other suspicious locations, because countries usually set up ground-to-air missile launch sites in specific areas to prevent valuable equipment nearby from being attacked by air.

According to the report, in the end, researchers only used about 90 samples from China missile launch base to train artificial intelligence. Such a small training data set may not achieve accurate deep learning results under normal circumstances. In order to get around this problem, Davis and his colleagues slightly changed the original image in different directions, and converted these more than 90 training samples into 893,000 training samples.

According to the report, the impressive deep learning results of this research are likely to benefit from the large scale of the ground-to-air missile launch site and the unique shape from the satellite overlooking photos. Davis reminded that when trying to analyze smaller targets such as vehicle-mounted missile launchers, radar antennas, vehicle-mounted radar systems and military vehicles, deep learning algorithms will face greater challenges because there are not so many pixels to use when extracting identification features from existing satellite images.

Profile picture: This satellite picture taken on February 2, 2006, 5438+02, 65438 shows the traffic route map of the Fengxili nuclear test site in North Korea. (The satellite image was taken by DigitalGlobe Company and marked on the website "38 North" of the Institute of American and Korean Studies, School of Advanced International Studies, Johns Hopkins University, and published on February 28th, 2065438+02/kloc-0. )

According to foreign media, China's artificial intelligence will rival the United States in five years: the PLA will gain an advantage.

Reference news network165438+1reported on October 29th? According to foreign media reports, a research institute in the American intelligence community recently held a competition to see who developed the best face recognition technology. Among all the contestants, a China startup named Yi Tu Technology won the top prize of $25,000.

Reuters165438+1October 28th reported that a report by a think tank in the United States cited many similar examples to prove that China's military may use the country's rapid progress in the field of artificial intelligence to promote the modernization of its armed forces and may seek military superiority over the United States.

The report, written by elsa Cania of the Center for New American Security, said: "Compared with the United States, China is no longer at a technological disadvantage, but has become a real rival, perhaps surpassing the United States in artificial intelligence."

Cania wrote that the future competition between the United States and China in the field of artificial intelligence "may change the future balance of economic and military power."

Eric Eric schmidt, the executive chairman of Alphabet, recently issued a similar warning about China's potential at a gathering in Washington. He pointed out that China's "New Generation Artificial Intelligence Development Plan" published in July required to catch up with the United States in the next few years and eventually become the world's leading artificial intelligence innovation center.

Schmidt said at this meeting: "I think we will still be in the leading position in the next five years, but China will catch up soon. So, after five years, we may be almost there. "

Earlier this year, an unpublished document of the Pentagon rendered that China enterprises evaded American supervision by buying shares in American companies, and acquired American artificial intelligence sensitive technology with potential military applications.

In response to this situation, some US congressmen proposed a bill this month to strengthen the management of foreign investment in the United States. The report of the Center for New American Security pointed out China's acquisition activities, and said that there were some obstacles in training domestic artificial intelligence industry to compete with the United States, including recruiting top talents.

However, Schmidt said, "If you think that their system and education system can't train the people I mentioned, you are wrong."

Artificial intelligence technology has promoted the emergence of self-driving cars, which may completely change the face of transportation. In addition, this technology has brought great progress to medicine. It is estimated that there are also military uses that can change the battlefield situation.

The Pentagon has a project that focuses on letting computers help screen images taken by drones, thus reducing the workload of human analysts. Some machine learning techniques have been used for this.

According to the report, the report quoted public documents as saying that the China People's Liberation Army is investing in various projects related to artificial intelligence, and its research institutions are cooperating with China's defense industry. The report said: "The PLA expects that the emergence of artificial intelligence will fundamentally change the nature of war."

Kania admits that since the development of artificial intelligence is still in the early stage, the policies of China and other countries on artificial intelligence are not perfect, and a considerable part of her research results are speculative. Nevertheless, she said in the report that some PLA thinkers predicted that there would be a "strange scene" on the battlefield, when human beings would not keep up with the speed and rhythm of machine-led decision-making in the course of combat.

Data Map: The Youth AI Artificial Intelligence Design Competition started in Suzhou.

Further reading American media: Is the fourth industrial revolution coming? Artificial intelligence will eliminate outdated jobs.

Reference news network165438+1reported on October 24th? The US media said that driverless cars and trucks occupy roads and robots "operate" factories. This super smart phone, called Uber Helicopter, sent its owner to the rapidly expanding urban area. Machines use algorithms for self-learning to complete cognitive tasks that once required human wisdom, eliminating countless management and industrial jobs.

According to the article on Washington post website165438+1October 22nd, this is the vision that the fourth industrial revolution triggered by technological progress will reshape the world-it is likely to be realized in the next five to ten years. Today, this scene can be seen not only in Silicon Valley, but also in think tanks in Paris, electric car factories in China and even around the Sahara desert.

2 1 century, the impact of technology is different from the past. When steam engines, electricity and computers were born, society had several years to adapt to the changes brought about by technology. Today, change is anytime and anywhere, and it spreads all over the world in digital form in an instant.

Governments at all levels around the world suddenly realize that social media, other forms of algorithms and artificial intelligence have quickly gone beyond their control and even made them unaware.

According to the article, the life of Americans today is far more widely influenced by technology than in the past few decades. Self-driving cars, cloud (technology) and swarms of drones are well-known concepts. However, these technologies are staring at you and me step by step, and this reality is really unprepared.

"China is not only a global factory, but also a global laboratory," said a speaker at the annual meeting of the French Institute of International Relations in Morocco this month. He refers to the rapidly developing military and civilian artificial intelligence technology in Beijing.

The article said that the US government stayed out of it and let market forces develop technology giants with global influence. China chose head-on competition. Europe allows American technology companies to enter the market and manage them instead of competing with them. Russia weaponized information technology and added social media to the army, missiles and tanks.

Diplomats and strategists have begun to explore the intersection of technology and international affairs, hoping to find a way to apply the principles of cold war deterrence and arms control agreements to the field of cyber threats.

Some people also call on the government to start to find ways to solve the impact of technology on the domestic labor market and the urgent problems brought about by the increasingly fragile political system.

According to the article, artificial intelligence and automation are creating new jobs and eliminating outdated jobs, and new jobs often need constant retraining and changes in various occupations and places. Employers in the United States report that there are currently 6,654,380,000 job vacancies, mainly because job seekers lack the required skills or job mobility.

Artificial intelligence robot

Further reading Foreign media said that China is leading in the field of artificial intelligence: high-tech non-western exclusive games.

Reference news network1reported on October 30th 10? British media said that high-tech was once regarded as the exclusive game of developed economies such as the United States, Europe and Japan, but now China is subverting this prejudice. In the field of artificial intelligence research and development, China unexpectedly leads the United States.

According to the Financial Times website1reported on October 20th, 18, "By 20 18, the China government's R&D expenditure on artificial intelligence is expected to reach1500 million US dollars." Dominic barton, chairman and global president of McKinsey, predicted this. Dominic barton is a China hand. Before becoming the head of McKinsey, he was stationed in Shanghai for a long time. At the end of September this year, he parachuted into Beijing and once again called on China to intensify its efforts to develop the artificial intelligence industry. He believes that China has great potential in this respect, and one of the most fundamental reasons is that it has a huge amount of life data. He said that there was a "false dawn" in the world 20 years ago, but there was no matching computing power at that time. Now, with the improvement of computing technology and the development of mobile Internet, more data has been collected, and artificial intelligence is changing from an idea to an applicable reality.

Previously, the McKinsey Global Institute made three judgments in a report: artificial intelligence investment has entered the stage of the world's leading technology companies competing for patents and intellectual property rights; Companies that enter the artificial intelligence industry in the early days are often closer to the digital frontier; High technology, communication and financial services will become the leading industries of artificial intelligence within three years. McKinsey predicts that the market size of artificial intelligence applications will reach $654.38+027 billion in eight years. Dominic barton said in his speech at the China Development Forum held in March 20 17 that some Internet companies in China are at the forefront in natural language processing, image and speech recognition, and can be integrated into new products such as personal assistants and autonomous driving.

As a veteran who has been in charge of this top global partner consulting company for nearly nine years, dominic barton has cultivated a very keen insight in his daily contact with various clients. An anecdote is that more and more institutions are beginning to take seriously the Belt and Road Initiative proposed by 20 13 China. In fact, as early as 20 10, Dominic Barton talked about the possibility of a "new trade axis", "This was the Silk Road".

Dominic Barton is very interested in China and knows how to exert his influence. Like many other industries, today's consulting industry is also facing the fate of being transformed by artificial intelligence. Dominic barton's daily job is to make decisions, and he needs to carefully weigh between hiring employees with high salaries and introducing artificial intelligence technology at high prices. He said that the company should undertake a "dual mission"-on the one hand, introduce robots and automated programs to meet customer needs, on the other hand, train employees to master the relevant skills necessary to use artificial intelligence.

The following are some excerpts from dominic barton's video interview with High-end point:

Q: Why are you calling on China to increase its efforts to develop artificial intelligence?

Dominic Barton: One of the reasons why I come to China regularly is that if you don't understand what is happening in China, you will be marginalized. Artificial intelligence is an area of interest to me, because I think it is an important technology that will change the world. I think China will have an important position in the field of artificial intelligence, mainly because of the talents and intellectual resources here. China's participation will also promote the development of the world, and we can solve many medical problems, such as cancer; There are many world-class problems such as the treatment of autism, which can be solved by artificial intelligence. In addition, it helps to improve the level of productivity. Therefore, China needs to play an important role in the field of artificial intelligence for various reasons.

Q: Both China and the United States focus on the development of artificial intelligence, while the population of China is nearly five times that of the United States, and China is still a developing country. What are the advantages of China?

Dominic Barton: The United States is in a leading position in this field, while Britain is very advanced in research. For example, there is a company like Deep Thinking, which has gathered many neuroscientists, biologists and computer scientists. Even Canada is trying to promote the development of artificial intelligence, and Toronto and Montreal are also trying their best to promote related industries. For me, I think it is more important to grasp talents and leadership, and artificial intelligence is an important opportunity. China already has data, which once again shows that it is not enough to develop artificial intelligence only by technology, but also by data. You mentioned that the population of China is five times that of the United States, and these data are important information for the development of artificial intelligence.

Q: You cherish talents and attach importance to education. Has your impression of China graduates changed in recent years?

Dominic Barton: Students in China have always been high-quality and talented. I remember when recruiting in several universities, I said that since these people have the ability to enter these universities, why bother to test their problem-solving ability? I may not even get into these schools myself. So, they are all very smart people. I think one of the changes I have noticed is that we have recruited many people who don't have that much social experience. Once I recruited two students in Shanghai, and they were both excellent. I let them work in KFC for two nights, but their parents were very depressed. A student's parents asked why they were recruited into McKinsey and worked in a fast food chain. I said, because you need to see people who don't flush the toilet, people who steal things and unreasonable customers. This is life. If you are always in an academic environment, you can't adapt to those who do strange things. I find it very important for more graduates to have such real social experience now, otherwise you will not be able to deal with people.

Dominic Barton.

Further reading American media articles: The next science and technology competition will focus on artificial intelligence.

Reference news network165438+1reported on October 9? On October 3rd, the bimonthly website of American Foreign Policy165438+/kloc-0 published the article "The goal of the next space race is artificial intelligence" by John Allen, the president of the Brookings Institution and former commander of the US military in Afghanistan, and amir hussein, an American artificial intelligence expert, saying that although the United States has most important innovation and research institutions in the field of artificial intelligence in history, international competitors are close behind.

Artificial intelligence has become a battlefield.

According to the article, about 60 years ago, lyndon johnson, then the majority leader of the US Senate, predicted that whoever won the space race would "control and completely control the earth". America finally won the race, not only landing on the moon, but also inspiring the next generation of scientists, technicians and optimists.

Recently, Russian President Vladimir Putin repeated Johnson's statement when he predicted the next great science and technology competition-artificial intelligence.

The article holds that there is a truth that transcends the times and geopolitics in Johnson's and Putin's understanding of technological power. However, now the United States is worried that it will face the risk of falling behind in this key competition.

According to the article, soon, in most social and economic fields, including work and entertainment, artificial intelligence will become the most potential changer, bringing us a competitive advantage, and its influence will far exceed the usual debate about automation replacing manufacturing work.

The article believes that although the United States is the birthplace of artificial intelligence and the location of most important innovation and research institutions in this field in history, international competitors are closely following.

According to the article, China recently announced a multi-billion dollar artificial intelligence development plan, which will lead the world in this technical field by 2030. Russia is developing the next generation MIG -4 1 fighter with artificial intelligence, which can control the fighter to fly at supersonic speed up to Mach 6.

The United States may be overtaken by China and Russia.

According to the article, the United States can see its weaknesses with the naked eye: the lack of artificial intelligence projects in the national high-level plan, the reduction of scientific and technological funds, and restrictions on immigration are all damaging its competitiveness. The question is whether the United States can correct its mistakes before it is too late.

The article has done a lot of thinking about the problems faced by the United States and the measures that should be taken:

First of all, encouraging American researchers to lead the world in innovation and opening up is also promoting their research results to enter the public eye quickly before they are protected. Although the United States cherishes an open academic culture, American enterprises need faster patent procedures and government support in order to have some advantages in intellectual property disputes with overseas infringers.

Secondly, the regulatory authorities make it very difficult to produce goods in the United States and sell them to other countries, thus creating a market where foreign competitors have no chance. For many years, the United States has banned the export of encryption technology and basic processors. What this practice brings is that international competitors meet the demand and create their own market.

Third, China has published more papers and supercomputers in deep learning than the United States. The United States needs more public investment in artificial intelligence.

Fourth, China is attracting better artificial intelligence talents and acquiring American technology companies. The solution is simple: provide more green cards for experts in this field; Provide more federal funds for research laboratories in public universities; Provide more investment in educational projects, such as Rhodes Scholarship, to attract future doctors.

The article said that the United States can never forget that artificial intelligence represents a breakthrough in human creativity. Artificial intelligence is the next great leap forward for mankind. The first country to land on the moon must now take the right step towards tomorrow.