SHANGHAI, China, Dec. 20, 2018 (GLOBE NEWSWIRE) -- During Sept. 22-23, the Machine Learning-Summit (ML-Summit) 2018 co-sponsored by Boolan and IT Boda was grandly held at Crowne Plaza Shanghai. Squirrel AI Learning’s Cofounder and Chief Scientist Dr. Wei Cui discussed the development and application of AI, shared the innovative technologies and application prospects of AI in various fields including social interactions, retail, transportation, industry, logistics, Internet of Things, finance, medical care, social security, real estate, and tourism, and looked into the future developments and challenges faced by AI with more than 30 global senior technical experts in the frontiers of AI, including “the Father of Machine Learning” Michael I. Jordan, Facebook Engineering Manager Shengbo Guo, Uber Machine Learning Platform’s Technical Director Eric Chen, Ctrip Vice President and Chief Data Officer Wei Feng, etc.
As the forerunner of machine learning and one of the developers of the root directory in the field of AI, Michael I. Jordan gave a keynote speech titled “The Principle of Human-Centered AI: The Learning System Meets the Economic System” at the summit. Based on his prospective judgment on the current AI industry and academic research, he put forward the important idea of “AI = Data + Algorithm + Market”.
He pointed out that the researches on AI have long ignored the fact that human society is essentially an economic society and the invisible hand of market is also advanced intelligence developed by humans. Prof. Jordan reviewed the several breakthroughs in machine learning since the 1990s. He pointed out that designing AI in full imitation of human from the God's-eye view is a wrong direction, and our goal should be developing small-scale intelligence that is well coordinated with each other and with human.
At the summit, Squirrel AI Learning’s Chief Scientist Dr. Wei Cui shared the implementation and practice of Squirrel AI intelligent adaptive technology and introduced Squirrel AI intelligent adaptive technology. When entering the Squirrel AI intelligent adaptive system, students need to take a knowledge test. This test can find out which knowledge points students have mastered, which students have not mastered, and which only know a little.
A photo accompanying this announcement is available at http://www.globenewswire.com/NewsRoom/AttachmentNg/ec889b8d-8194-47c9-8f4d-3c81c695a287
In his speech, Dr. Wei Cui deeply interpreted the Squirrel AI intelligent adaptative learning system and how to use AI to bring forth disruptive innovations. In the knowledge test, the intelligent adaptative system adopts knowledge space theory and information theory technology, which can accurately, quickly and effectively check students' mastery of knowledge. For example, for a subject covering more than 100 knowledge points, it can finish testing using over 30 exercises. The principle behind it is the intelligent adaptative system has built a knowledge state space. For example, the knowledge state space formed by 100 knowledge points is 2 to the 100th power. The purpose of the test is to find out the corresponding knowledge state of each student in the space. The method of exhaustion will take too much time. AI technologies such as knowledge space theory and information theory can narrow the search scope down and quickly identify each student's knowledge state. Knowledge state space technology can filter out invalid knowledge state. Information theory technology can select the most critical knowledge points and give them priority in the test and get more information about how well a student knows. The accuracy rate is now up to 90%.
Dr. Wei Cui also said that with the advent of AI, the infinite computing power of AI has helped to solve the problem of knowledge splitting. For example, Squirrel AI Learning’s knowledge splitting has reached the world-leading nanoscale. One knowledge point in the syllabus may become all of 13 knowledge points in the knowledge system..Moreover, Squirrel AI has a knowledge correlation theory system. It can extend from the knowledge points that a student has not mastered on the surface to the prerequisite knowledge points the student has not mastered. As a result, in just one hour, it can clearly know a student’s mastery of 5,000 knowledge points and truly diagnose the student’s knowledge loopholes.
A photo accompanying this announcement is available at http://www.globenewswire.com/NewsRoom/AttachmentNg/08d4e399-7125-402e-8325-ab95c7c91831
After that, when it’s used in learning, the intelligent adaptative system can collect the learning data generated in the learning process in real time, update student portraits constantly based on their past learning records, plan the optimal learning path timely and recommend the most suitable learning content dynamically for each student, thereby maximizing learning efficiency.
That is teaching students in accordance of their aptitudes, which is always stressed in education. At present, there is still no way to teach students according to their aptitudes in the classroom. All students listen to the teacher at a fixed pace, whether you are a 50-point student or a 95-point student. Such uniform education is like Tencent news and Sina news in the past. If we want to distribute different contents to different users like Toutiao does, we need machine learning and genetic algorithms, to let the system constantly and more and more accurately push exercises and contents that match each student's weak points, to achieve the most efficient learning effects.
Squirrel AI Learning’s AI intelligent adaptive learning model and technology represent the most advanced level in the Chinese market, mainly because the company has gathered three of the world's leading experts in adaptive learning technology, including Wei Cui, Richard Tong and Dan Bindman, who are respectively the company’s Chief Scientist, Chief Architect and Chief Data Scientist. They respectively came from RealizeIT, Knewton, and ALEKS, three world-renowned AI adaptive education companies. Using nearly 10 years of first-hand experience in intelligent adaptive education technology application and research, they make Squirrel AI stand on the shoulders of giants, helping Squirrel AI build its evolving technological barriers with proprietary intellectual property rights.
Squirrel AI Learning has established a Joint Lab for AI Intelligent Adaptive Learning with Stanford Research Institute (SRI) and a Joint Lab for Parallel AI Intelligent Adaptation with China Academy of Sciences (CAS) Institute of Automation. Moreover, AI technology papers on Squirrel AI intelligent adaptive learning system have been accepted and presented at the world-top AI academic conference AIED for two years in succession. Squirrel AI Learning has been invited to give speeches and got papers accepted at many global top-class academic conferences on AI, such as CSEDU, UMAP, and ACM TURC.
Then, Dr. Wei Cui introduced the four human vs. machine teaching competitions organized by Squirrel AI Learning, which repeated the success of Alpha Go defeating world champions Sedol Lee and Jie Ke. In Oct. 2017. In the first human vs. machine teaching competition in the Asia-Pacific region in Zhengzhou, Henan, Squirrel AI teaching robot defeated a senior teacher with 17 years of teaching experience. In April 2018, in the second human vs. machine teaching competition in Chengdu, Sichuan, Squirrel AI teaching robot defeated a senior high school entrance examination board member. In June 2018, in the third human vs. machine teaching competition in Dongying, Shandong, Squirrel AI teaching robot defeated a senior teacher who conducted 1V3 teaching.
Globally, there are already more than 100 companies developing AI intelligent adaptive education, with nearly 100 million users. Predictably, there has been a global wave of AI-enabled education, but it is just the beginning in China. Dr. Wei Cui hopes Squirrel AI can solve the problem of unequal distribution of teachers in traditional education and revolutionize the uniform traditional teaching mode, thereby giving every child a super AI teacher.
About Squirrel AI Learning
At present, Squirrel AI Learning already has more than 1 million users. Next, it will develop an OpenAI platform, so that all institutions and individuals no longer need to develop their own intelligent adaptative systems. In addition, Squirrel AI Learning will offer 1 million learning accounts to children from poor families for free.
To learn more about Squirrel AI Learning, please visit http://www.squirrelai.us/.
Guangming Yang