Using Pedagogical Agents In a Multi-strategic Intelligent Tutoring System

acquiring and modeling knowledge in Intelligent Tutoring Systems (ITS) design is more complex than in most other knowledge based systems since ITS incorporate different kinds of expertise (knowledge on the subject matter, knowledge on the learner’s knowledge, pedagogical expertise, etc.) that evolve continuously. The pedagogical expertise, in particular, tends to be distributed among several components such as virtual learners and a tutor, which play different roles to enhance learning in collaborative mode. In addition, all these knowledge need to be improved in order to be adapted to the learner and the different learning situations. To foster such an improvement we introduced the principle of actor. An actor is a form of intelligent agent which is reactive, instructable, adaptive and cognitive. The property of instructability is particularly important since by allowing the instructional designer to feed the agent with new knowledge during its execution, it helps to solve the knowledge acquisition problem. This paper describes the use of actors for implementing pedagogical strategies and more generally for detecting which strategy is more suited to a given learner. Our approach leads to the definition of a multi-strategic ITS, based on pedagogical actors, that is able to switch among various strategies. In this framework we use pedagogical actors to model the expertise of the various pedagogical strategies. We illustrate the paper with an example drawn from a small ITS in the domain of radiology which has been implemented using an actor platform. keywords : pedagogical agents, actors, intelligent agents, multi-strategic intelligent tutoring systems, pedagogical strategies, knowledge acquisition.