How Will AI Help China's Education
COVID-19 and China
Ma Gang speaking at the Global New Economy Conference. PHOTO: EqualOcean

From June 12 to 14 The 2019 Global New Economy Conference jointly organized by the Changning Chapter of Shanghai Youth Federation and EqualOcean under the guidance of Shanghai Municipal Commission of Economy and Informatization, Shanghai Municipal Commission of Commerce, and the Shanghai Changning District Government was successfully held at the World Trade Center in Shanghai. The event focused on the theme of "Technology innovation leads to the era of intelligence,” and emerging industry innovation hotspots such as 5G, New Energy, Intelligent Manufacturing, etc.

On June 12, 2019, at the Leaders’ Summit under Global New Economy Annual Conference - Squirrel AI partner Ma Gang (马刚) gave a speech.

Ma Gang: Good afternoon, my friends! Today, I'd like to talk to you about how to use AI to improve teaching.

Our understanding of education is “to read a book a hundred times and see for oneself", "there is no easy way of learning.” But these verses, in fact, are just another way to tell us how hard learning is.

"Diligence" requires a large amount of time, and "bitterness," tells that learning is impossible to be happy. Chinese students struggle the most in learning. The worst part is that they are required to work on a huge amount of exercises and quizzes to pass exams like “gaokao”. This is happening to all Chinese students across the country, from north to south, from tier-one cities to bottom-tier cities.

The reason behind this is the shortage of qualified teachers. On the one hand, good teachers are very rare in terms of time and resources. On the other hand, because of the unequal development of China's economic resources, the difference in teacher resources from first-tier cities to lower-tier cities is very large.

It is very difficult for the children of Xinjiang (新疆) and Tibet to enjoy the educational resources of Shanghai and Beijing, and for the children of the sixth-tier towns and villages to enjoy the resources of the first-tier cities. Because good teachers tend to be concentrated in cities with relatively developed economies.

Is there a way through which we can increase the number of qualified teachers? Yes, Artificial intelligence can help us achieve that.

In China, children spend a lot of time studying and the learning process is very hard. For my child, who is a third grader, goes home every night to do homework. It takes about two to three hours to do homework, if he plays with his friends for a little while, he will do his homework at about 11 o'clock in the night.

Staying up till 11 o'clock has an impact on the child's biological clock, and it also has an impact on his/her physical health, but there is no other way because the school requires you to do so much homework.

Right now, 90% of a child's learning time is wasted. My child, for example, comes home every night with about 80% of the work cannot do, but there is no way it must be done because it is almost impossible for a single teacher to assign different homework to 50 students. 

At present, there are a large number of companies in China and abroad who want to use artificial intelligence to solve the above problems. We have the same aim. Squirrel AI began to explore this track in 2014, and we are honored to start laying out this track before so many peers.

Below, I want to talk about how to use artificial intelligence to solve the problem mentioned above.

Firstly we divide all the learning materials into different types and build up a knowledge graph. The density of this graph is high. By doing this, the system can figure out the most important or urgent knowledge that students should learn.

We also hope to maintain focus throughout the child's learning process. Children like to play games on the Internet but are not willing to learn two hours of language, mathematics, and English at home. This is because their children lack a lot of sense of accomplishment in the learning process, and the sense of accomplishment is to keep the children's attention the most which is core point. How can I maintain my child's sense of accomplishment? That’s where the evaluation process of the AI helps the most.

“刷题” is a process of constantly looking for mistakes, reviewing and verifying. If students got 60% in the earlier years, it means there’s a 40% loophole. Therefore, they wouldn’t learn new knowledge when they didn’t make up the shortage. It’s like building construction. If the foundation is poor, the building is dangerous. So it’s important for kids to look back to the early point where they didn’t understand the basic knowledge.

The other is personalized matching. Different children are completely different in a lot of aspects. We will found out that each child spends a completely different time in learning the knowledge points. We want to know all the children's learning when they learn all the knowledge points, which is almost impossible for human teachers.

The demand for personalized education is very high, so we hope to achieve it through artificial intelligence.

From 2016 to the present, we have built nearly 2000 learning centers. At the earliest, we had a lot of doubts, but we aimed and believed that during the learning process, the children can go to the learning center in an atmosphere where they can learn together. We believe that such learning has twice the effect with a minimum effort. From first-tier to six-tier cities, we have learning centers in relatively remote areas, including Xinjiang and Tibet.

We hope that through artificial intelligence, the difficulties of current Chinese education and children's learning can be changed. I hope that the learning process that is bleak and dull will be slightly brightened. I hope that children can get very high-quality teachers in the learning process. Our expectation is that all children can have their own super teacher of "Socrates + Da Vinci + Einstein", I believe that this goal is not too far. Thank you all!

Latest Updates:

Communicate Directly with the Author!

Ask the author questions about the copied text

Research Reports
Editor's Picks