Generation of Demand Feedback in Intelligent Tutors for Programming

Abstract. In this paper, we will examine how a model-based tutor can automatically generate demand feedback. We will propose a two-stage feedback generation algorithm that maintains the principle of modu-larity characteristic of model-based representation, while automatically generating detailed demand feedback. We will evaluate model-based pro-gramming tutors to demonstrate that the feedback generated using this algorithm is coherent and effective at improving learning.