See Upcoming Public Defences

FULLY DIGITAL (UNTIL FURTHER NOTICE) : PhD Defence Daniel Davison | 'Hey robot, what do you think?' - How children learn with a social robot

'Hey robot, what do you think?' - How children learn with a social robot

Due to the COVID-19 crisis the PhD defence of Daniel Davison will take place online (until further notice).

The PhD defence can be followed by a live stream.

Daniel Davison is a PhD student in the research group Human Media Interaction (HMI). His supervisor is prof.dr. V. Evers from the Faculty of Electrical Engineering, Mathematics and Computer Science (EEMCS).

As robots become more available and accessible, we expect that they will gradually move into the classrooms of the future. Before such a transition can take place we need to better understand in what ways robots may excel and add value to the learning process. We also need to gain a better understanding of what it takes to develop and use them in real classrooms for longer periods of time.

This thesis explores how we can make a meaningful impact on the learning process of primary school children by delivering certain forms of social support through a robot. We argue that a robot---by virtue of being a robot---is in a unique position to offer these kinds of support. We consider the robot as a holistic entity consisting of many social features and modalities, together establishing it as a complete and convincing social agent. To show the effects of the robot as a whole, we compare it to situations where children learn without a robot; support is then delivered through a less-social tablet device.

We developed and evaluated four variants of a Robot-Extended Computer Assisted Learning (RECAL) system. When working with the robot, children gave better explanations and showed an improved mindset towards learning---this shows that robots can indeed make a meaningful contribution!

Furthermore, we share insights to gain a better understanding how technologies like the RECAL system may be used for conducting research in the wild over longer durations. We show how the autonomous robot and learning tasks offered sufficiently rich and challenging content to elicit unsupervised recurring spontaneous interactions with children throughout a four month period. We also discuss how in-task experience sampling and embedded sensors in the learning materials were used to follow their progression through tasks and difficulty levels. With these first steps we have shown that conducting long term studies in the wild is a feasible endeavour yielding valuable insights.