Abstract:

Nonmonotonicities of students' behavior and the student model inference process itself are discussed including the usage of the model in tutoring. They are classified into two classes, i.e., single world contradictions and multi-world contradictions (student knowledge contradictions). Student knowledge contradictions are the essentials of the learning processes of a student. This paper presents a new perspective to capture them formulating a student discrimination structure. A student model description language SMDL and modeling algorithm HSMIS, which is a nonmonotonic inductive student modeling system, are formulated to cope with single world contradictions. SMDL is based on a logic programming language taking 4 truth values. THEMIS is a new nonmonotonic and inductive model inference system which incorporates deKleer's ATMS as a vehicle for formulating both nonmonotonicities (contradictions). The formulated THEMIS embodies advanced representation power, sufficiently high adaptability and generality. Not only can it follow a student's change of understanding, but it can model a student who has inconsistent knowledge.