Pierre Dillenbourg

Some noise for my protégé

Pierre Dillenbourg
EPFL, Switzerland

A lot of research on robot-learner interactions seems to follow the postulate that a nice learner-robot interaction is a condition for learning to occur. It is indeed a necessary condition, but not a sufficient one. Learning first depends on the activity that the learner has to perform with the help or guidance of the robot. In this talk, I explore one of these activities, learning by teaching or the so-called ‘protégé effect’: a child is expected to improve his writing by teaching Nao how to write. In a learning context, the relationship between the quality of interactions and learning is non-linear. If one could design a robot that perfectly understands the child and vice-versa, learning would actually be diminished: what produces learning is not the level of mutual understanding among the robot and the child,, but the effort that both agents have to perform (e.g. number of explanations, rephrasing, deictics,…) to strive towards a mutual understanding of the task at hand. This is why some noise, some misunderstandings are useful for learning, as long as the environment provides the resources for detecting and repairing misunderstanding in a smooth way.

Short bio: A former teacher in elementary school, Pierre Dillenbourg graduated in educational science (University of Mons, Belgium). He started his research on learning technologies in 1984. He obtained a PhD in computer science from the University of Lancaster (UK), in the domain of artificial intelligence applications for educational software. He has been assistant professor at the University of Geneva. He joined EPFL in  2002. He is currently full professor in learning technologies in the  School of Computer & Communication Sciences, where he is the head of the CHILI Lab: “Computer-Human Interaction for Learning & Instruction ». He is also the academic director of  Center for Digital Education, which implements the MOOC strategy of EPFL. His work on robotics only started a few years ago, when Frédéric Kaplan and then Severin Lemaignan joined his lab. He hence became member of the Swiss NCCR on robotics where he worked on several projects with Aude Billard, Francesco Mondada, Auke Ijspeert as well as Ana Paiva.