, Author: eggs, butter, Zenan, original title : ” Is China’s AI technology leading because of good math? American magazine points directly at the gap between Chinese and American mathematics education
The picture from: Visual China
Last week, a Chinese student’s video titled “The Mathematical Abilities of Asians Is Actually Poor?” appeared on Hot Search.
“Knowing it but not knowing why, they are only good at using it instead of pursuing the truth.” This girl thinks that most Asians pay attention to the use of mathematics when they study mathematics but don’t care about the derivation of the principle. The argument has some truth, but the conclusion has led to Few people vomit.
Soon, the blogger “They all call me Aunt Xue” released a new video to apologize to everyone, saying that personal experience cannot refer to the group, and the remarks about the poor mathematics of Asians are very incorrect. In a country where examination-oriented education is the mainstay and the pressure for further education is high, if students spend so much energy but cannot understand a science, it is indeed terrible. Mathematics is the most important foundation of contemporary higher education.
In July, the 2020 Alibaba Global Mathematics Competition ended, the event founder Ma Yun commented”https://img.huxiucdn.com/article/content/202007/14/130043623412.jpg?imageView2/2/w/1000/format/jpg/interlace/1/q/85″ data-w=”795″ data-h=”530″>
But this is not a problem of artificial intelligence itself, but a problem of mathematics.
Artificial intelligence cannot simply be seen as a “black box”-as long as unlimited investment is made, it can grow and develop. If American talents cannot master the basic knowledge of mathematics required in the field of artificial intelligence, so many think tank projects and government reports will be meaningless. Without the ability to abstract math, billions of dollars of investment requirements are also unrealistic.
What we call “artificial intelligence” today is actually a combination of various algorithms and their characteristic development, which draws heavily on advanced mathematical and statistical knowledge. Taking deep neural networks as an example, it is not an artificial brain, but a large number of repeated calculations of a series of gradients (high school calculus is hardly taught)” The “learning” information conversion module, the gradient is also the backbone of the back propagation algorithm family.
Backpropagation is a common method used in conjunction with optimization methods (such as gradient descent) to train artificial neural networks .
All machine learning can do similar analysis. This research is about how to write a computer program to learn a task instead of performing a pre-coded task. The ability to quickly classify data, identify patterns, predict results and self-learning is all due to increasingly complex algorithms, as well as increasingly powerful computing power and massive amounts of data.
From iPhone to Summit, the most powerful supercomputer in the world, from GoogleTo Facebook, these computing platforms and computing programs use extremely complex mathematical calculations, ranging from simulated nuclear explosions to online search results.
Contrary to Li Kaifu, author of “Artificial Intelligence Superpowers: China, Silicon Valley, and the New World Order,” artificial intelligence is more than data. Kai-Fu Lee once mentioned that today’s data is like the oil resources of the early 20th century, and China with the most data is the new Saudi Arabia. However, if there is no matching mathematical ability, and some people who promote the creative development of the field of mathematics, all the data in the world together can only take you to this step-not enough to reach the frontier researchers in the field of AI We boldly imagine the future.
No matter how you divide it, the world runs on the basis of 1 and 0, and the algorithms that run them are repeatedly tempered. The talents who created these algorithms have been trained for many years to master these complex mathematical knowledge.
Unfortunately, middle school and college students in the United States do not have basic mathematical knowledge, such as statistical theory and differential geometry, and this knowledge is precisely the basis for them to enter a higher stage and the basis of artificial intelligence.
In the Organization for Economic Cooperation and Development (OECD) in the 2018 International Student Assessment Program test, the math scores of 15-year-old American students are 35 points, far below the average level of the member states of the organization. Even at the university level, American students usually have to memorize algorithms and then insert algorithms when needed because they do not have the basic knowledge required to solve abstract problems.
The failure to develop advanced mathematics skills for students means that fewer and fewer American citizens are pursuing advanced degrees in mathematics and science. According to data from the National Science Foundation, in 2017, more than 64% of doctoral candidates and nearly 70% of masters students in computer science in the United States were international students. Half of the doctoral degrees in mathematics were awarded to non-U.S. citizens, including China. Students and Indian students make up the majority. Excellent Asian students hope to go to American universities for further study. This is largely because the training system of American universities is more advanced than that of their own countries, although the educational gap between China and the United States continues to narrow.
However, this also means that those who are trained by American universities to open up new directions in the field of computer science and mathematics are not Americans. Some of these people will stay in the United States, but many people will also return to China to devote themselves to the development of their country’s technology industry.
There are good reasons to believe that the United States should relaxThe entry visa restrictions encourage more people to stay in the United States after completing their studies. But the most important issue is that there are simply not enough U.S. citizens who take advanced mathematics as a major subject. These details have caused corresponding effects, from coping with global competition to the entrepreneurial culture in Silicon Valley, from concerns about national security to whether American companies I think I need to think about the interests of the United States.
The stagnation of mathematics education in the United States is in sharp contrast to China’s active promotion of the AI plan in 2030. In the field of artificial intelligence technology today, the United States and China occupy the most important positions. China has invested a lot of money in artificial intelligence-related technologies, making it the core focus of domestic industrial modernization. In the “Made in China 2025” strategy, the term “intelligent manufacturing in China” was proposed. The system combining manufacturing technology and information technology became the development direction, and technologies such as artificial intelligence dominated.
At present, China’s AI market is about 24.5 billion yuan. In the “New Generation Artificial Intelligence Development Plan” issued by the State Council, the basic theory of artificial intelligence will achieve a major breakthrough by 2025, and some technologies and applications will reach the world’s leading level. Artificial intelligence has become the main driving force for my country’s industrial upgrading and economic transformation. At that time, the scale of the core industry of artificial intelligence is expected to exceed 400 billion yuan, driving the scale of related industries to exceed 5 trillion yuan.
To achieve this goal, China has invested 14 billion yuan to build an AI industrial park in Beijing. And technology companies such as Huawei that lead the development of the global industry are also launching their own AI chips to create a “full-scenario AI” strategy.
In China, artificial intelligence has stepped out of the laboratory and entered the stage of technology exploration. Based on highly accurate facial recognition technology, we can implement many fast and convenient applications, such as authentication when a bank transfers large amounts, remote account opening by a securities firm, or even a virtual ID card. AI labs of some technology companies have also launched deep learning models that can recognize Deepfake face-changing technology. These technologies will play a very important role in combating black production and evidence fraud.
Although China’s AI industry is constantly getting closer to the United States, China still takes a longer-term view-they are trying to improve the mathematics of young people. According to the latest statistics from the Organization for Economic Cooperation and Development (OECD) for 15-year-old students worldwide, Chinese students have the world’s first mathematics ability (Science and reading ability are also the first). There is no doubt that China is focusing on STEM (that is, science, technology, engineering, and mathematics) education, which surpasses the United States and Europe country.
In January 2020, the Ministry of Education issued a notice to carry out the pilot reform of basic subjects enrollment in some colleges and universities (also known as the strong foundation plan), It pointed out: “To serve the national strategy, recruit a group of young students who are ambitious, interested, and talented to carry out special training, and to provide backup talents for major national strategic areas… The main selection and training are those who are interested in serving major national strategic needs and have excellent comprehensive qualities. Or top students in basic subjects.”
On May 6, universities participating in the Qiangji Program have published their first year enrollment guidelines, including Tsinghua University, Peking University, and other 36 “first-class university” construction colleges and universities. This plan focuses on key areas such as high-end chips and software, intelligent technology, new materials, advanced manufacturing and national security, as well as the humanities and social sciences where national talent is scarce. It mainly highlights the supporting and leading role of basic disciplines, all limited to basic disciplines, and It is not allowed to change majors, which is also the biggest difference from previous independent admissions. (University’s independent admissions have also been cancelled).
It seems that both Chinese and Americans have the same idea to improve their basic education and improve their strength in cutting-edge fields such as artificial intelligence. But beyond mathematics, there may be many determinants.
Reference content:
Almost Human (ID: almosthuman2014) , author: egg sauce, Zenan < /p>