ActiveMath - generation and reuse of interactive exercises using domain reasoners and automated tutorial strategies

In this thesis we present an approach to interactive exercises for mathematical domains. We have defined a generic knowledge representation for multi-step interactive exercises and developed the exercise sub-system within the ActiveMath learning environment. In order to achieve deep domain diagnosis and automated feedback generation in multiple mathematical domains, a distributed semantic evaluation service architecture was developed. This architecture allows to remotely query multiple domain reasoning services using a domain-independent query format. In our approach, we allow for hand-crafted exercises for which we provide authoring tools, as well as fully or partially automatically generated exercises. We have also built an extensible framework for tutorial and presentation strategies, which allows for reusing the same exercise with different strategies.