Improving cooperation in Virtual Learning Environments using multi-agent systems and AIML
Márcio Aurélio dos Santos Alencar, José Francisco de Magalhães Netto · 2011
The large number of messages posted on the forum, a key element in Distance Education Courses based on Virtual Learning Environment, which does not receive adequate feedback from the tutor in sufficient time is a typical problem faced by students in these environments. The tutors, in turn, feel a lack of tools to monitor activities carried out by the student. This article proposes an approach for solving these problems based on the concept of perception, using the multiagent system paradigm. The system is composed by intelligent agents, which act in a Moodle discussion forum using an AIML knowledge base. Agents solve questions about matters discussed at the forum, and they use perception to recommend the implementation of activities that the student has not done. The results from simulations based on real courses already completed show that there is a decrease in the workload of tutors. Students are reminded the deadline for the tasks automatically. It was necessary to create hundreds of AIML rules to get answers to good level. The partial results indicate that the approach of combining AIML and MAS is promising to improve the feedback from tutors and motivate students to conclude their work on time.