Eyebrow raising, eyelid tightening, and mouth dimpling are facial expressions that indicate the highest level of a student’s engagement.
Students at a Dubai university have created a tool using artificial intelligence (AI) for increasing the effectiveness of online education, which will come handy amid the raging Covid-19 pandemic.
Learners at Murdoch University Dubai under the mentorship of their faculty have created a prototype AI tool that could hold the key to enhancing the delivery and effectiveness of virtual learning.
The Emotion-Affective Domain Mapping System (EADMS) is an innovative solution that aims to bridge the gap in the efficacy of learning that naturally exists between onsite and online learning.
It is a prediction tool that uses facial data collected from students’ webcams during online classes, extracts emotion attributes and maps it with Bloom’s Taxonomy, a global standard of behavioural and cognitive learning outcomes.
In a virtual classroom environment, educators lose the opportunity to interact with students and tailor their teaching style to what suits them best.
During face-to-face learning teachers normally use students’ facial expressions and emotional responses to predict their understanding levels subjectively, which is often not possible in a virtual environment.
Dr James Trotter, Murdoch University Dubai’s Dean and Academic President, said, “It’s very difficult to assess student comprehension levels subjectively during virtual classes. It’s also estimated that the problem of poor integration of technology within education systems is set to cost us roughly 0.3 to 0.9 years’ worth of education per year. The proposed EADMS solution, by two of our bright students and their faculty mentor, is a novel way to improve the comprehension assessment and subsequently improve the course content and delivery.”
Experts agreed that the EADMS system captures “students’ facial data during online classes in the form of a video and uses AI to determine emotions such as contempt, anger, fear, happiness, disgust, surprise and a neutral state of emotion”.
The mechanism translates facial data into emotional data, which is mapped with the ‘Affective Domain’ of Bloom’s Taxonomy.
This, in turn, helps generate a graphical chart that plots the understanding level of students.
The EADMS has demonstrated success in extracting information from videos on the internet to a reasonably reliable level.
A combination of geometric and appearance-based video analysis, as outlined by the EADMS, has provided accurate results for detecting engagement levels.
Mark Brown, General Manager (GM), Murdoch University Dubai, said that the institution is involved in taking the EADMS prototype to its next level of logical development.
“We anticipate that after the due refining, testing and development process, that is essential to bringing such a technical product to market, we will soon see it reach the product stage. It would be a great achievement, not only for our University but also for the UAE, when the product finds popularity with educators across the world,” he added.
© Khaleej Times
Twitter