Project in the spotlights: PhD student Mirjam de Haas on project L2TOR

13 May 2022

A small white robot points to a wooden block. "One," calls out the girl sitting across from him on the floor, as she too points with her little finger. "Two," continues the robot. His hand moves to the adjacent cubes. "Two," echoes the child's voice again. It is a scene from a promo video of L2TOR ("el tutor"). Back in 2015, a robot was developed through this MindLabs project to teach foreign languages to preschoolers. Now, more than 7 years later, the project has 8 robots and 6 different PhD studies have already been done around this topic. And many more will follow. Because researcher Mirjam de Haas is 'far from finished'.

[Vertaal naar het "English":]
[Vertaal naar het "English":]

De Haas has been involved with L2TOR from the beginning. Last April she defended her thesis 'Staying engaged in child-robot interaction: A quantitative approach to studying preschoolers' engagement with robots and tasks during second-language tutoring'. With success; she received her PhD.

Gestures as an advantage for learning performance
De Haas on her research: "In education, we are increasingly dependent on digital tools. COVID-19 has demonstrated this once again. During the corona period, digital tools would have been very useful for teachers. Robots, for example, that, unlike tablets, can use their bodies to behave similarly to teachers."

That recognizability then lies in things like gestures while talking, the PhD student reveals. Or by pointing to objects. "Children can concentrate better this way, research shows. An advantage for their learning performance. Moreover, robots enable children to interact socially. Much more so than when using tablets. And that is especially important when learning a second language."

European collaboration
L2TOR was launched as part of Horizon 2020; an EU funding program for research and innovation. It is a collaboration between 6 European universities (including Tilburg University) and 2 companies that developed the robots. The main experiment was a large-scale field study in which about 200 Dutch children learned English words during 7 lessons. This study showed that children successfully learn from robot Robin.

By the way, there was also a "but. The robot did not necessarily perform better than a tablet. Children learned as many words with the robot as with the tablet. De Haas: "It did turn out that children are more involved when a robot uses gestures than when it does not. This prompted me to write my own dissertation."

The future of Robin
Because development in the field of AI has been very rapid in recent years, new technology can be added to social robots all the time. "So there is still plenty to experiment with." De Haas enthusiastically lists a number of new research questions. "Because of these new techniques, for example, a robot can increasingly work independently, without a human instructor. How does this affect teaching? Or what can Robin teach older children, who already have a broader basic knowledge?" Her enthusiasm is partly fueled by the environment in which she operates, the PhD student indicates. "The dynamic MindLabs ecosystem inspires. When I'm working here, it immediately motivates me."