Meta reasoning and student modelling

The student modelling process in Intelligent Educational Systems is the process of inferring a model of the student's knowledge by analysing his or her behaviour. When the student's behaviour is the result of a reasoning activity, the task of building a student model involves the system's reasoning about another agent's knowledge and reasoning (meta-reasoning). In this work we propose a method to describe the student modelling process and give a clear an explicit account of the assumptions which can be made on the student's reasoning activities. The approach is based on a meta-level architecture, where the domain specific knowledge of the student and the expert are represented at the object level, and the meta-reasoning principles are described in a first order logical theory at the meta-level. The abstraction work needed to formulate the basic principles informing the student model activity of some existing educational systems, in a neat and uniform way, has naturally brought some new insights and proposals.