From the Director – Pymble Institute
The Pymble Institute newsletter has been published with updates on a range of activities for Term 4. Click here to visit the Pymble Institute site and read the newsletter in full.
Research with UTS Education and Data Science Institute into AI and online learning tasks
One of the research projects under the Digital Intelligence Strategic Pillar is a collaborative pilot project with teams from the University of Technology Sydney. The project is investigating how we can use Artificial Intelligence technology to better understand what it is like for students working on online tasks. Involving participating students from Year 9 and 10 Data Science classes, the goal of the research is to help teachers design even more effective online tasks in future. This joint project brings together academics from UTS Education and International Studies (Professor Nick Hopwood, Dr Tracey-Ann Palmer, Dr Mun Yee Lai) and UTS Data Science Institute (Dr Kun Yu, Dr Yifei Dong) with Data Science teachers (Cedric Le Bescont, Anthony England, Kim Maksimovic and Dr Glen McCarthy).
Technology developed by Dr Yu, Dr Dong and colleagues from the Data Science Institute can track where students look on the screen, for how long, where they move the cursor, how intensely they focus, and facial expressions including eye and head movement which suggest concentration and distraction. The data collected is not in video nor photo-form, but has been transferred into code. The code helps the researchers understand student engagement – what students pay attention to when they are working on a task, how long they remain focused on the screen, what might be more or less challenging in the task and what happens to eyes and the body in different parts of the task.
A group of helpful and inquisitive students from Data Science classes completed an online task while the software monitored their faces and computer clicks, and they then participated in a focus groups to explore the experience. The researchers are currently aligning both types of data to understand the features of tasks that work most effectively to engage students.
