Learner modelling by expert teachers: learner information space and the minimal learner model

Abstract

Giving individualized guidance to a learner must rely on a learner model. Learner modelling has been little studied within education, although the specialised research field of Artificial Intelligence and Education (AI&ED) has emphasized its importance. We present results from the Mayday project's observations of learning interactions suggesting that the learner models used by expert human teachers contain diverse kinds of information much of which is sketchy and conjectural. We propose a theoretical framework: the Leamer Information Space and the Minimal Leamer Model.