Non-human Modelers: Challenges and Roadmap for Reusable Self-explanation

Abstract

Increasingly, software acts as a “non-human modeler” (NHM), managing a model according to high-level goals rather than a predefined script. To foster adoption, we argue that we should treat these NHMs as members of the development team. In our GrandMDE talk, we discussed the importance of three areas: effective communication (self-explanation and problem-oriented configuration), selection, and process integration. In this extended version of the talk, we will expand on the self-explanation area, describing its background in more depth and outlining a research roadmap based on a basic case study.