Foundations for Complete Exercise Generation

Abstract

This paper introduces a new approach to exercise generation for university courses, that aims for high domain coverage, variability as well as adaptability. For this, we utilize a domain-specific language (DSL) which allows the creation of models describing subject matter. These models are further aggregated with a topic-oriented representation of a course and serve as the basis for an exercise generator, which then uses various templates to process different types of knowledge into different types of exercises. This novel approach is initially developed and tested by applying it to two engineering-related university courses.