Eric Bruillard

ICT, educational technology and educational instruments. Will what has worked work again elsewhere in the future?

Computers in education have been around for more than 30xa0years. They have been the vector not only of diverse forms of educational technology, but also of many new kinds of learning activities using software instruments. With the gradual trivialization of computer systems, the initially rather positive expectations about computers in education have progressively turned into more skeptical attitudes about the use of ICT. What is really working with computers in education? Was the investment worth the outcomes? The number of scientific studies focusing on these issues is huge. Most of their results have been obtained either by very focused experiments on some educational technology, generally involving a limited number of students, or by surveys, taking into account large numbers of learners but with little control about the kind of activities that were actually organized. Overall, hardly any conclusive results about possible causal implications of using ICT in learning have been obtained and there is no evidence about the stability of existing results over time. This paper, particularly taking into account the French experience, draws attention on the specific case of software instruments. These instruments do not exactly aim at performing more efficiently existing activities, but lead to invent new learning activities better in line with constructivist views. It suggests that, besides studies about what works, more research is in order about how it works and contends that what will work in the future is linked to the capability of the teaching profession and their allies to collectively build and discuss new interrogations and solutions to pedagogical problems, which will sustain and regulate the many discoveries of innovators.

Automated Test Assembly for Handling Learner Cold-Start in Large-Scale Assessments

In large-scale assessments such as the ones encountered in MOOCs, a lot of usage data is available because of the number of learners involved. Newcomers, that just arrive on a MOOC, have various backgrounds in terms of knowledge, but the platform hardly knows anything about them. Therefore, it is crucial to elicit their knowledge fast, in order to personalize their learning experience. Such a problem has been called learner cold-start. We present in this article an algorithm for sampling a group of initial, diverse questions for a newcomer, based on a method recently used in machine learning: determinantal point processes. We show, using real data, that our method outperforms existing techniques such as uncertainty sampling, and can provide useful feedback to the learner over their strong and weak points.