NANJING, China, July 17, 2019 /PRNewswire/ -- On June 27, Slush 2019 Nanjing Technology Innovation Conference was held at Nanjing International Expo Center of China and was attended by more than 2,000 attendees and industrial leaders from around the world. The event was also participated in by 243 start-up businesses, 124 institutional investors, 40-odd news media, 200-odd volunteers and 1,500-odd high-tech talents and university students from all quarters of the society. Liao Yifang, a partner of Squirrel AI Learning and directing manager of its campus operation department, was invited to deliver a speech of "How AI self-adaptive technology can upgrade education". Her wonderful speech aroused keen interest among experts, scholars, investors and other audiences.
Dozens of guests from enterprises and investors discussed and shared their views on the topics and latest trend of industries such as artificial intelligence, big data, and gaming, and included Peter Vesterbacka, the founder of Angry Birds and Slush, Zhou Jian, the CEO of Microsoft Accelerator, Liao Yifang, a partner of Squirrel AI Learning, Chen Yijing, the person in charge of product operation of LinkedIn China, Dong Bo, a senior supervisor of SAP, Gu Dawei, the secretary general of the technical committee of Xiaomi Corp, and Zeng Haoshen, a partner of Chenhui Venture Partners.
AI technology revolutionizes traditional education mode, multi-dimension innovation realizes intelligent tutoring
Liao mentioned, education is an important part in our life. The artificial intelligence technology can upgrade the traditional education.
In traditional education, what difficulties were we faced with? First, as we know, education resources are seriously scarce. It is difficult to develop an excellent teacher. Last July when I traveled to Tibet, I found that the education resources there were poor. After a discussion, we found that if artificial intelligence was applied, we could spread quality education to every corner without any limitation of time and location. Through artificial intelligence technologies and algorithms, we can develop excellent AI teachers and duplicate and spread the good teaching experience beyond the limitation of time and location.
The second difficulty we faced in traditional education is that the learning contents are almost the same for every student. Fixed plans and contents of learning are used for all students in the class, yet every student varies in individuality, so their learning methods should be different too. How can we solve this problem?
The AI self-adaptive education can solve this issue. Truly personalized education will enable every kid to learn in a suitable way. This is also what can be brought by AI self-adaptive education.
A third difficulty we faced in traditional education is that teaching or learning goes on at the same pace for all students. It is clear that everyone varies in learning speed and efficiency. In fact, the AI education will enable every kid to learn at a suitable speed and achieve a better learning effect, which is impossible in traditional education.
In the past, our learning focused on imparting and absorption of knowledge points. We have spent a lot of time in reciting, memorizing and doing assignments. Under the heavy learning pressure, it is usually difficult to cover all learning contents, and sometimes we may even neglect the kids' learning of thoughts, abilities and methods to a certain extent. That is to say, we learn so much knowledge not for taking a test, but to train the abilities of our kids to solve problems. What on earth can AI self-adaptive education bring to us? If we can include more learning thoughts, abilities and methods, apart from the teaching and learning of knowledge points, I am sure that it will make a difference to our kids.
Liao said, taking Squirrel AI Learning as an example, it can "teach students in accordance with their aptitude", target their knowledge gaps and weak areas and recommend learning contents accordingly so that the students can learn at a suitable time. Besides, it can trace and identify the students' mastery of basic knowledge and fill their earlier knowledge gaps.
In addition, Squirrel AI self-adaptive learning system can strategically abandon some knowledge points which are not high in score and probability of appearing in the exams, and let the students learn those that can be better absorbed by them and enable them to get a better score. In a way, Squirrel AI Learning has made education and learning more simple and efficient, liberated the students from the excessive assignments, strengthen their interest and confidence in learning and realize intelligent teaching at different dimensions.
Superiority of Squirrel AI Learning in terms of engine architecture
As a domestic leader of AI+ education, Squirrel AI Learning has developed China's first AI self-adaptive learning engine with complete independent intellectual properties and centered on advanced algorithm. Squirrel AI self-adaptive learning engine imitates an excellent teacher, with an aim to improve or accelerate the learning progress of students through personalized teaching methods. The system can identify what students need to learn, what they have mastered and what they have not and how well they have mastered the knowledge. Students use the system for learning, conducting evaluation and getting help until their learning objectives are realized.
Why is Squirrel AI Learning so confident? Liao said, the Squirrel AI self-adaptive learning engine architecture consists of three levels. The first level is fundamental level, showing a knowledge map of learning, including a map of contents and mistake analysis. In the course of learning, the kids try to master every knowledge point. Why did a mistake happen? Suppose ten thousand students answered the same question and two thousand of them gave a wrong answer. Did they make the mistakes for the same reason? Not at all. The causes vary among the students. We will sort out these causes, analyze the data and even design a way for the students to interact on the user's interface, thus finding out the actual causes of their mistakes.
The second level is algorithm level. In the entire structural design, say, we want to evaluate the knowledge points, recommend appropriate contents and the persona path of learning. In this process, algorithm accuracy is very important. That is to say, the level and degree of algorithm will determine the answer to be obtained in database analysis.
The third level is user interface level. It is a management system of learning, which can analyze different behaviors and even analyze the brain wave response of a kid when he is doing assignments, including the visual anchorage of his eyes so as to analyze his behaviors. Suppose I am to give a speech, but I am not born with a good linguistic competence. So how should I train myself? The breakdown of abilities can assist us to guide the students to learn about their advantages and disadvantages and pursue a personalized development.
Squirrel AI Learning has established different learning models for different grades and subjects, which can recommend suitable learning methods, assignments or tests to the students. In the course of learning, Squirrel AI Learning can record and analyze their learning status. Through the established knowledge map, it can identify what the students have mastered and what they have not and then recommend suitable assignments to fill their loopholes and enhance their learning efficiency.
Technology helps remove education barriers, but the substance and original aspiration of education remain unchanged
These years, the academic papers published by Squirrel AI Learning have been recognized by international academic conferences such as ACM, IJCAI, AIED, AERA, CSEDU, and KDD. It has established a Yixue AI Lab with SRI International as the primary research partner, Joint Labs with Chinese Academy of Sciences and Carnegie Mellon University and increased investment in AI education. Meanwhile, Squirrel AI Learning has launched the "Human vs Machine competition" four times in succession, which have been spread from Zhengzhou to 100 cities of different levels. Through a competition between "robot teaching team" and "human teaching team", it has effectively accelerated the fulfillment of Squirrel AI system.
Technically, the self-adaptive learning does not require presetting a learning process. Through algorithms, it can plan the knowledge points of several orders of magnitude and identify the best learning path for students according to the multi-dimension big data produced in the learning of students. Theoretically, adaptive learning focuses on the application of course theories, while self-adaptive learning focuses on how teachers teach and how students learn.
Liao believed that if kids can perceive the fun of learning from their learning achievements, they may become addicted to it. Today if we want to learn about the students, it is natural and understandable that we resort to exams. But it will be much more worthwhile if all learning states can be recorded. Education can be smarter. We should use more effective methods to gain the learning outcomes we desire. When your kids master all learning contents within a short time and efficient methods, what will happen to your kids? They will be confident and develop a great sense of achievement.
Squirrel AI Learning has globally initiated nano-level breakdown of knowledge points and breakdown of learning thoughts, abilities and methods, which can accurately detect the knowledge loopholes of every kid. As long as we master the real learning methods, we can improve our learning efficiency. While understanding the students' persona, we should upgrade their real-time persona. How can we identify the mastery degree of all knowledge points by every kid and their weak areas in different learning contents? Squirrel AI Learning can help the kids find a suitable learning strategy by analyzing all data. Therefore, ordinary teaching and extended learning will be provided according to the academic performance of students. Everyone will be matched with a different learning path.
Can we obtain a good learning outcome with 1/3 or 2/3 of time? What will happen if we do less assignments we have understood? In fact, whether AI empowers the application scenes, or AI self-adaptive optimizes the learning path, new technologies are set to revolutionize the learning environment. However, education will be changed by AI technology, the substance of education will remain unchanged. The tasks of education will always be to impart knowledge, clear confusion and inspire people.
Liao believed that every kid should be matched with an excellent AI teacher. The AI teacher will not only recommend knowledge points, but also infuse the kids with a scientific spirit and an artistic and philosophical spirit. The substance of education has remained unchanged. She hoped that all kids can receive quality and personalized education.This is a revolution to be brought by AI self-adaptive education.
SOURCE Squirrel AI Learning