ABITS: An Agent Based Intelligent Tutoring System for Distance Learning

The purpose of this paper is to describe an Intelligent Tutoring Framework highly re-usable and suitable to several knowledge domains. In particular the system, named ABITS, has been realized in the context of the InTraSys ESPRIT project. It is able to support a Web-based Course Delivery Platform with a set of “intelligent” functions providing both student modeling and automatic curriculum generation. Such functions found their effectiveness on a set of rules for knowledge indexing based on Metadata and Conceptual Graphs following the IEEE Learning Object Metadata (LOM) standard. Moreover, in order to ensure the maximal flexibility, ABITS is organized as a Multi Agent System (MAS) composed by pools of three different kind of agents (evaluation, pedagogical and affective agents). Each agent is able to solve in autonomous way a specific task and they work together in order to improve the WBT learning effectiveness adapting the didactic materials to user skills and preferences.