Observation and adaptation of a learning session based on a multi-agent system: An experiment

The work reported here deals with the observation of pedagogical activities in educational platforms. Userspsila behavior is observed through the traces of their activity. In this paper we present a multi-agent system designed to bring more flexibility to the observation of learnerspsila behavior in educational platforms. Our system enables the teacher to dynamically adapt his/her observation scenario during the learning session by possibly reconfiguring dynamically both the behavior and the topology of dedicated agents. We describe our multi-agent system with a dedicated agent configuration language and we illustrate it through an example of use in a real learning context.