Squirrel AI Learning is the first artificial intelligence technology company in China to apply AI-adaptive technology to K–12 education.
Derek Haoyang Li elaborates on the four elements of the AI-adaptive education framework, including business model, advanced AI technology, data and computing capabilities, and technology implementation methods and the connection between business processes, pedagogy, architecture, operations, and theoretical foundations of adaptive learning. He shares how more than a dozen of artificial intelligence algorithm technologies, such as knowledge space theory, Bayesian theory, logistic regression, genetic algorithm, and deep learning are integrated in the Squirrel AI intelligent adaptive system.
You’ll have the opportunity to see how the students’ learning map, the content map, the learner model, and the learning goals are integrated in the Squirrel AI ontology. Through the knowledge map of tens of thousands of knowledge points created by the Squirrel AI, it can accurately capture the real-time learning data of students, analyze the learning situation of each individual student and student groups through data, and accurately diagnose the knowledge mastery and provide learning materials for students. It can optimize learning content in real time, tailor different learning contents to suit each student’s different learning paths, and help students achieve their learning goals.
Derek details the continuous data-based system optimization of Squirrel AI, including student learning content assessment, intelligent adaptive model assessment, teaching methods and product improvement under reinforcement learning. He brings the latest projects developed by Squirrel AI and SRI including the combination of different AI algorithms and the adoption of different AI models, adaptive learning strategies based on knowledge maps, multimodal behavioral analysis integration, research on biological neural input signals, learning tools such as natural language-based interactive intelligent robots, and cognitive simulators.
The natural language-based, interactive, intelligent robots project applies the speech-based two-party or even three-party dialogue mechanism. In the real interactive scenes where the Squirrel AI engine, the coaches, and students interact with each other, a lot of AI-driven methods such as deep learning, natural language processing, and logical thinking reasoning methods are used. In the communication between students and intelligent robots, the system can find the root cause of students’ mistakes and collect students’ data on mistakes in large quantities and use these data to optimize the intelligent adaptive learning system and create the comprehensive student database. Through dialogues, it will help students find the real reasons of mistakes and enhance the confidence and verbal skills of students who are too timid and shy to interact with human teachers. This system also enhances the learning ability of intelligent robots through interaction with students.
In the application of innovations used in different scenarios of the AI algorithm, Derek also shares the combination of the different research that Squirrel AI Learning has been conducting for testing and learning, such as the integration of item response theory, which has been widely used in adaptive learning systems to evaluate students’ current abilities and question-difficulty parameters, as well as Bayesian knowledge tracing and deep knowledge tracing methods commonly used in adaptive systems.
Derek Haoyang Li is the founder and chief education technology scientist at Squirrel AI Learning, the top 20 AI-Unicon in China. As a serial entrepreneur, Derek cofounded the first education company listed in China’s A-shares market. He was awarded the Top 30 AI-Entrepreneur in China. Derek also created several ingenious educational innovations in the world, “Concepts on Nano-Scaled Knowledge Components,” “AI-Model-Adapted Learning-Skills-Decomposition Methods,” “Reconstructing Knowledge Space Theory (KST) with Students’ Reasons for Mistakes,” and “Algorithms on Calculating the Relevance of Probability Between Non-Relevant Knowledge Components.”
Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?
Join the conversation here (requires login)
©2019, O'Reilly Media, Inc. • (800) 889-8969 or (707) 827-7019 • Monday-Friday 7:30am-5pm PT • All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners. • firstname.lastname@example.org