A meta-engineering approach for customized document-centered knowledge acquisition

Abstract

Enabling domain specialists to formalize domain knowledge using knowledge acquisition tools, also known as direct knowledge acquisition, has been a challenge in artificial intelligence research for decades. Building custom-tailored knowledge acquisition tools proved to be a convenient method to empower domain specialists to contribute to knowledge bases directly. Engineering on the meta level, i.e. building tools that allow to build customized knowledge acquisition tools, is the strategy to reduce implementation costs for the customized tools. In this paper, we present an approach to build customized knowledge acquisition environments for documentcentered knowledge base authoring. This editing paradigm, being an alternative to the use of graphical user interfaces, is well suited for the customization approach allowing for knowledge acquisition at low effort for both, the domain specialists and the knowledge engineers. We present a meta-engineering process coordinating the customization efforts and a metatool for the system KnowWE enabling efficient metaengineering. To illustrate the overall approach we describe a real-world case study in e-learning in the domain of Ancient History.