Squirrel AI Learning by Yixue Group ranked among MIT TR 50, a list of 50 Smartest Companies

HANGZHOU, China, July 17, 2019 /PRNewswire/ -- On June 29, the 2019 Summit of Smartest Companies of MIT Technology Review was held in Hangzhou. At the summit, the MIT TR 50, a list of 50 smartest companies of the world selected by MIT Technology Review was unveiled. Squirrel AI Learning, a unicorn enterprise of AI+ education of China was included in the list. The list was honored as a "guide of emerging technology to the future world". Starting from the perspective of combining technology and commerce, MIT Technology Review predicted the leading companies in the future. It is one of the most influential ranking lists in the global circle of science and technology.

MIT Technology Review is a business media of emerging technology with the longest history and greatest influence in the world. Since 2010, it selects "50 smartest companies (TR50) from around the world based on multiple dimensions such as breakthrough and innovation created by companies, industries and even the world as a whole according to their technical competence and commercial sensitivity and the core competence of emerging technologies.

As a leading enterprise in the AI+ education industry of China, Squirrel AI Learning is the first brand of AI Adaptive Education devoted to the education of primary and secondary schools in the Asia-Pacific region as well as the first high-tech company of China that has developed an AI Adaptive Learning system for primary and secondary schools with independent intellectual properties. It has been included in the MIT TR50. Other excellent enterprises that were also selected include Huawei, Tesla, TSMC, Ali Cloud, Iflytek, Tencent and Jinri Toutiao.

In this year's TR50, we have seen the ceaseless pursuit of giants such as Huawei, TSMC and Xilinx in emerging technology, the prominent performance of start-ups such as Qihan Bio-tech, Coherent AI and Xizhi Technology, and more companies that have accumulated a solid strength and foundation in their respective fields such as Ping An Insurance, Envision Group and Tesla which have bravely crossed the borders to integrate new technologies and business models. We have also witnessed the continuous breakthroughs made by AI companies such as Squirrel AI Learning and Sense Time in the past two years.

The reason offered by the MIT Technology Review for selecting Squirrel AI Learning is: It has perfected teaching quality through the independently developed PKS model, MIBA system and AI system. The PKS model can work out the central probability values of test questions, and evaluate the dynamic mastery degree of the students in each knowledge point. The MIBA system, by collecting eye-ball movement data, facial expression data, behavior and action data, brain wave data, helps the self-adaption engine to accurately judge and form persona of students and change the teachers' teaching mode and students' learning mode.

The comment fully recognized the achievements of Squirrel AI Learning in scientific researches. Particularly, the core algorithm engine of Squirrel AI Learning has significantly enhanced the learning efficiency of students by handling 2 million knowledge points, 3 million videos, 10 million test questions and 225 billion learning behaviors in the peak period of summer vacation. Ranked top among the peers, it is leading the development of AI self-adaptive education in China.

The summit has invited industrial leaders, experts and scholars such as representatives from those companies that have ever been listed in MIT TR50, the smartest companies to be listed in the newest TR50, scientists from international academic research institutions and investors from cutting-edge technology investment institutions of China. The event was participated in by industrial leaders, renowned scholars and journalists such as Mi Lei, a founding partner of Casstar, Daniel Starta, a partner of AT Kearney, Zhou Wei, the co-founder and CEO of Squirrel AI Learning, Patrick Chang, the investment director of Samsung NEXT, Chen Huiling, the Chief Content Officer of MIT Technology Review China Region, Professor Donald Sadoway for John F. Elliott Lecture of MIT. Zhou Wei was invited to deliver a thematic speech "Guessing the seventh wave of education" and shared his innovative ideas with the guests and audiences on reforming the education sector with AI technology, which had received a strong response from the audiences.

A new entry of MIT TR50, Squirrel AI Learning: Steady accumulation of AI technology

As a unicorn enterprise in the AI+ education industry of China, Squirrel AI Learning has made dazzling achievements in the past year. It has stood out in the evaluation of several indicators and finally been listed in the MIT TR50.

MIT TR50 emphasizes the R&D capacity of an enterprise in emerging technology. Having emphasized R&D of science and technology and invested in the team of scientists continually, Squirrel AI Learning has won the favor of the judges. In the investment of laboratories, in 2016, Squirrel AI Learning has established a Yixue AI Lab with SRI as the primary research partner. In June 2018, it co-established a joint laboratory of AI Adaptive education with the Automation Institute of Chinese Academy of Sciences. In May 2019, it co-established a joint laboratory with Carnegie Mellon University to develop artificial intelligence, machine learning, cognitive science and man-machine interface technology and new methods to improve the adaptive learning experience of K-12 students worldwide.

Through constantly cooperating with cutting-edge scientific research teams and institutes worldwide in development and research of laboratories, Squirrel AI Learning has made milestone achievements at the technical level. By applying PKS (Probabilistic Knowledge State) model, MIBA (Multi-modal Integrated Behavioral Analysis) system and strengthening the learning RL, Squirrel AI Learning has further enhanced the intelligence of its self-adaptive learning system. PKS model can work out the central probabilistic value of each test question and evaluate the mastery degree of each student in each knowledge at each moment. Just like X-light scanning, it can clearly and accurately analyze the learning status of each student. MIBA system applies machine learning algorithms such as deep learning to analyze the data of students collected from multiple channels, including eyeball movement data, facial expression data, behavior data and brain wave data to evaluate the comprehensive learning state of the students, accurately judge whether the students have been attentive and concentrated, or felt puzzled in the process of learning, helped the self-adaption engine to form a complete persona of students, thereby recommending more suitable learning contents and enhancing the learning efficiency and effects. The personalized recommendation model based on deep learning for strengthening learning algorithms and real-time training can provide more accurate and personalized content recommendation service to further strengthen the intelligence of its self-adaptive learning system.

This has not only further optimized the efficiency of AI Adaptive Learning system, but also enabled AI teachers to be more familiar with the conditions of students. The application of PKS system, MIBA system and MCM system in the robot teacher system of Squirrel AI Learning is more consistent with the substance of education. In the past, AI education was imparting knowledge points. Through efforts at the technical level, Squirrel AI Learning has been able to teach the students some learning capacity with lifelong benefits.

PKS system has enhanced the accuracy of testing and significantly reduced the demands for algorithms and data, which has thus solved a great tough issue in the field of education. The quantity of data is relatively limited in the field of education and medicine. The formats of collected data vary. It's very difficult to obtain data. Squirrel AI Learning system has solved the limitation of data quantity. Through testing, it can depict the user's persona more clearly.

MIBA system has solved the difficulty for AI virtual teachers to communicate and dialogue with students, which has increased a multi-dimension description of the users and thus made the system develop a clearer and more comprehensive knowledge of the learning state of students. If MIBA system could be promoted on a large scale in three or five years to come, it will enable AI teachers to surpass human teachers in such aspects as the range of knowledge, material library, teaching path and strategy and in judging and perceiving the overall state of students throughout their learning process.

The team of scientists of Squirrel AI Learning has made great contribution to the achievement of so many technical breakthroughs. The substance of science and technology lies in people. Squirrel Learning has always tried its best to maintain its own talent pool of scientists, because only an excellent scientific research team can reach farther and farther in technical breakthroughs.

The scientists of Squirrel AI Learning come from unicorn AI education companies worldwide such as Knewton, RealizeIt and ALEKS. Its chief scientist Dr. Cuiwei has a post-doctoral degree of AI from National University of Ireland and used to be a core scientist of RealizeIt. The chief architect Richard is the technical director in charge of the Asia-Pacific Region of Knewton, the earliest self-adaptive education company of the US. The chief data scientist Dr. Dan Bindman is a co-founder of ALEKS, a giant company of self-adaption in the US, who is a chief architect in charge of the algorithms of key products.

Particularly in 2018, Professor Tom Mitchell, the Godfather of Machine Learning, Dean of Computing School of Carnegie Mellon University, an academician of National Academy of Engineering and National Academy of Arts and Sciences of the US, a fellow of American Association for the Advancement of Science (AAAS) and the Association for the Advancement of Artificial Intelligence (AAAI), officially accepted the invitation of Squirrel AI Learn to serve as its Chief AI Scientist. As the primary person in charge of the artificial intelligence field of Squirrel AI Learning, Mitchell will lead dozens of AI scientists and hundreds of AI application engineers and technicians of his team to carry out basic research on artificial intelligence in the field of self-adaptive education and develop related products.

In 2019, Koedinger, a professor with Computer and Psychology Department of Carnegie Mellon University (CMU), director of LearnLab Research Institute, and one of the three giants of self-adaptive education science of America, officially served as Chief Learning Scientist of Squirrel AI Learning.

The man-machine interaction system proposed by Professor Tom Mitchell is currently a globally leading algorithm scheme. The algorithm scheme is no longer self-training of AI model, but enables every human user to train the AI model.

We can expect that the above technologies will bring an overwhelming change to the education sector of China in three or five years to come.

In the traditional education model, too much emphasis has been placed on the teaching of teachers. Due to the lack of quality education resources, the workload of teachers becomes very heavy. If AI system can perfect teaching quality, the role of teachers can shift from teaching contents to monitoring and supervision. Teachers can shape the character of students, have emotional communication with the students and shape their values. The Ministry of Education of China has ever claimed that AI will not replace the position of teachers, but those teachers who cannot apply AI will be discarded.

Zhou Wei: Guessing of the seventh wave of education

As a high-tech company listed in the MIT TR 50 this year, the CEO of Squirrel AI Learning, Zhou Wei was invited to deliver a thematic speech "Guessing of the seventh wave of education" in the summit, who has reshaped the education sector from the perspective of industrial development. Zhou Wei believed that currently the network education development has undergone six waves, which are the Internet + wave, the tool wave, the O2O wave, the human one-to-one education wave, the content wave and the AI wave. The guessing of Squirrel AI Learning for the seventh wave of education is unmanned classroom.

Zhou Wei believed that each wave has bred new technologies and upgrade of hardware equipment. "We have a beautiful imagination of future education, unmanned classroom. This is an exploration of future education. We will build a learning system centered on the students completely." Zhou said, "We are going to make a precedent for the new retail of Squirrel AI Learning. Through artificial intelligence technology, we will control the costs of teachers to a low level. There will be no teachers, but only assistants. The operating efficiency will be improved significantly, and the education will be equal. The unmanned classrooms online and offline will be allocated with 30% of human assistants. The education team behind it will be full-time senior and excellent teachers. In this way, quality education resources can be spread in each classroom in the forms of artificial intelligence and Internet."

Education needs to form a close loop. Squirrel AI Learning will provide an excellent education scenario for the students so that their learning can be more effective. This education mode is called by Squirrel AI Learning as free of disturbance under supervision. Zhou also said, new retail is allocated according to needs. Education will be allocated according to the learning progress of students. We will build a new learning relationship based on artificial intelligence. Previous learning relationships established a teaching system centered on teachers. After the introduction of artificial intelligence, we can build a learning system centered on students completely.

Now Squirrel AI Learning has nearly 2,000 learning centers nationwide, which covers more than 400 cities.

Finally Zhou said, "Hopefully through AI technology, we can provide every kid with a super AI teacher combining Confucius, da Vinci and Einstein. It will thoroughly solve the ratio between students and teachers. In the traditional teaching, one teacher has to teach 100 students. The learning effect depends on the matching degree between the 100 students and the capability value of the teacher. In the future, we will reshape this teaching relationship based on AI so that 100 teachers will serve one student only in our learning system."

SOURCE Squirrel AI Learning