Teachers are using facial recognition to see if students are paying attention


Remember those days of being in school struggling to concentrate in class? Well, a school in France is hoping to tackle those wandering minds by using technology.

ESG Business School in Paris is starting an artificial intelligence program which will identify whether students are paying attention in class. 

From September, software called Nestor (like the butler in The Adventures of Tintin comic book) will use students’ webcams to examine eye movements and facial expressions to establish whether they’re actually attentive to a video lecture. 

“In the past decades, everybody thought taxi cabs would live forever. Uber destroyed the system to rebuild it. I wanted to do something similar in education, specifically for poor students and e-learning,” Marcel Saucet, founder of LCA Learning, the company which created Nestor, told Mashable.

Using 20 key landmarks on the face of the user — eyes, brows, lips, jaw — Nestor can understand how the user behaves in front of a 5-minute video lesson.

“We are using machine learning algorithms to improve performances of the tool. Besides, we are working with psychoanalysts, like Dr. Saverio Tomasella, to keep a human understanding in the solution and make it more accurate,” LCA Learning’s Nicolas Delhaume said. 

At the end of the clip, students have to take a quiz focused on the moments when they were not learning.

Study times 

Nestor — which was presented at the UN’s World Council of Peoples event in New York — will collect data from several sessions to build a pattern which predicts when students are more likely to lose focus, according to Saucet.

“If a certain student’s attention wanes after 2 minutes and 30 seconds, for instance, a notification will pop up on the screen 10 seconds before that timestamp to warn them,” Saucet told Mashable.

The software can also integrate with the students’ social networks and calendar to suggest study times.

“If a student’s routine is to watch YouTube at 10pm on weekdays, Nestor could suggest that they study instead,” Saucet said.

Initially, Nestor will work only for volunteering students watching lectures remotely for two online classes at the ESG Business School. 

However, Saucet plans to extend the technology for all students, who will receive notifications whenever they’re not listening in class. 

It could also help teachers improve the effectiveness of their teaching as they’d be able to get feedback on when students were paying attention or not. 

The tech isn’t intended to replace teachers but the makers believe it will help develop education among poor children. 

“You can have students who are going to pay £100,000 at Oxford University or 100,000 poor kids in Bangladesh or Uganda who can afford only one AI-driven video,” Saucet said.

Privacy concerns

It’s an interesting application of tech in the field of education at a time when facial recognition is being put to use for commercial and security purposes — from Amazon scanning faces in its grocery store to police using the software on 117 million Americans — albeit with privacy concerns over how the data is collected and used. 

But Saucet is keen to alleviate those concerns, adding that data would be encrypted and anonymised, and will not be collected by LCA Learning, but destroyed immediately.  

“We currently keep only the analysis data, and no information about the user,” said Saucet. 

 LCA also says it has no plans to sell data to third parties, but that could change in the future. 

“In 10 years data will be collected and used for financial matters. In a couple of years, if the collection of data is worth 10bn dollars, I’ll give 5bn to Bangladesh and 5bn I’ll keep for me,” Saucet jokes. 

“The data is obviously the source for our big data work, and a need to improve the understanding of what is happening in front of the camera,” adds Delaume. 

It’s clear that LCA Learning is keen to stress the positive side. But even if they don’t initially sell the data, the school will also need to make a decision about how it the information is used. 

Https%3a%2f%2fblueprint api production.s3.amazonaws.com%2fuploads%2fvideo uploaders%2fdistribution thumb%2fimage%2f570%2fd36f0ff2 6662 4a07 b4de a367e1a74b6c



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *