Inductive Methods for Acquiring Task Knowledge in Adaptive Systems

摘要

In order to achieve the adaptation of interactive systems to situation-specific task requirements, models of tasks as well as methods of deducing possible tasks from the user's input actions are required. The paper gives a detailed description of two inductive, similarity-based methods for acquiring task knowledge from the dialogue history. The first approach is semi-automatic and relies on interactions with a human referee, whereas the second is completely automated based on certain heuristics. The methods are analyzed according to the underlying machine learning principles. Finally, an analysis-based approach for acquiring operational task schemata based on declarative descriptions of generic task concepts is briefly explained. The different methods have been implemented and successfully tested in the FINIX environment. They are general in that they may be used to acquire procedural task knowledge in different domains.