Using AI to personalize learning

AU researcher Dr. Oscar Lin uses AI to explore adaptive learning and delivering a more personalized education experience for online learners.

Educators have long understood that different students have different learning styles and diverse needs. It’s one of the reasons so many are drawn to a flexible online education at Athabasca University. 

Dr. Oscar Lin, a researcher and professor in the School of Computing and Information Systems in the Faculty of Science and Technology, hopes to use AI to further customize the learning experience, so that the course content itself can be adapted to fit each individual learner’s needs. In other words, he’s researching opportunities for adaptive learning. 

“Our educational paradigm prioritizes industrial-era efficiency over personalized intellectual development, transforming classrooms into mechanical processing environments,” he said. “Recognizing that every learner is unique in their skills, interests, and requirements, we see adaptive learning as the key to fulfilling their potential.” 

Adaptive learning does not mean changing the learning outcomes of a course, which are typically included in the syllabus for each online course. These would remain the same for all students, as consistent learning outcomes are critical to the integrity and validity of any given course. 

Instead, Lin said this would mean customizing the pathway through which each student achieves those learning objectives for any given course. This could include changing the sequence of a course to best conform to an individual student’s needs. 

“What works for one student at one time, doesn’t always work for another student at another time,” he said. “It’s always changing.” 

The goal, he explained, is to have an AI application in an online learning environment that could support an educator in adapting to an individual student’s preferences—but not replace the educator altogether. Similar to how AI works with an autonomous vehicle, a human would still maintain the ability to intervene in decision-making.

What works for one student at one time, doesn't always work for another studnet at another time. It's always changing.

Dr. Oscar Lin, Professor, Faculty of Science and Technology

Adapting and personalizing learning with AI 

Lin explained there is some element of customization already at play in most educational settings.  

“Different teachers have different pedagogical strategies, but which one will work better? We don’t know; it depends on the students,” he said. “Teaching isn’t just a science, it’s an art.” 

In designing a course, Lin identified three main components: what to teach, how to teach it, and how to assess the student’s learning. 

An AI tool could be used to help teachers by making high-quality predictions about what sequence of learning might help each individual student. Learners would still be expected to reach the stated learning objectives in a course, but an AI could help assess their current knowledge level and make suggestions for how to bridge the gap between current knowledge level and learning outcomes. 

“The sequence is so important. That’s where we feel we need our algorithm,” he said. “We still need to assess students to understand their level, and then the algorithm can help identify the next step, to tell us how we can close the gap for this student.” 

This project, Eliciting adaptive sequences for online learning, is funded through a Natural Sciences and Engineering Research Council of Canada Discovery Grant. Lin collaborates on this work with other researchers at AU, including students, and researchers outside the institution. 

He is part of an international effort through the Institute of Electrical and Electronics Engineers (IEEE) to standardize the data that different learning management systems use to describe student activity, so that data from multiple different systems can be used to make better adaptive learning predictions to benefit students. 

“This also paves the way for commercialization, so that our research results in not just published papers, but it can be commercialized,” he said.

AI research opportunities for students 

Lin said the goal of his work is not just the research itself, but also to provide learning opportunities for AU students. Many students at both the graduate and undergraduate levels have been able to develop projects that contribute to his ongoing research. 

One student in the Master of Science in Information Systems program, for example, has worked on a project related to creating adaptive quizzes and feedback tailored to individual learners. This would mean an AI could support instructors in more effectively and efficiently assessing a student’s knowledge level and recommend a personalized learning sequence. 

Lin added there are also many opportunities for undergraduate students in the Bachelor of Science in Computing and Information Systems program who are doing a project or independent study in their final year. 

“I have several students who have been very successful,” he said. 

Raymond Morland (Bachelor of Science in Computing and Information Systems ’24) worked with Lin during his undergraduate degree, and now as part of the Master of Science in Information Systems program, on many of these adaptive learning projects. 

“These projects have enabled me to apply the knowledge I have gained through my coursework at Athabasca University to meaningful, real-world research,” he said. “This experience has not only deepened my understanding of cutting-edge AI but also fuelled my passion for pursuing further research in this dynamic field.” 

These projects have enabled me to apply the knowledge I have gained through my coursework at Athabasca University to meaningful, real-world research.

Raymond Morland (Bachelor of Science in Computing and Information Systems '24)

AI to support Space-Defence Technologies Alberta 

While the adaptive learning projects have been keeping Lin busy for the past several years, he has recently become involved with the Space-Defence Technologies Alberta project, a multi-institutional collaboration funded through Alberta’s Major Innovation Fund. 

The overall scope of this project is to support Alberta’s competitiveness in space commerce and defence by advancing technologies and training the future workforce. 

Lin is the institutional lead on this project for AU, and he explained his work specifically is focused on proposing AI algorithms to train groups of unmanned aerial vehicles, also called UAVs or drones, to perform specific tasks intelligently and collaboratively with each other and with human commanders, while adapting to new information. These specific tasks are related to situational awareness and uncertainty assessment. 

“The higher level will be how can we control a fleet of drones to complete a task, like putting out wildfires,” he said. 

One example he cited is if a fleet of drones could be used to perform firefighting activities, such as pinpointing hotspots in a wildfire and delivering water precisely to the right spot. Artificial intelligence could be beneficial in helping the drones to adapt to rapidly changing situations on the ground via human-AI collaboration.