A framework for ICAI systems based on inductive inference and logic programming

The main components of an Intelligent Computer-Assisted Instruction (ICAI) system are the expertise, the student model and tutoring strategies. The student model manages what the student dose and dose not understand, and the performance of an ICAI system depends largely on how well the student model approximates the human student. We propose a new framework for ICAI systems which uses the inductive inference for constructing the student model from the student’s behavior. In the framework, both the expertise and the student model are represented as Prolog programs, which enables to express the meta-knowledge that is the knowledge of how to use the knowledge. Since the construction of the student models is performed independently of the expertise, the framework is domain-independent. Therefore, an ICAI system for any subject area can be built with the framework. As an example, the ICAI system teaching chemical reaction is presented together with a sample performance. The authors believe that the new framework for ICAI systems based on logic programming and inductive inference could be a breakthrough of the future ICAI systems.