Type:

• Bachelor Thesis Business Information Systems

Status:

offered

Tutor:

• Prof. Dr. Mario Nadj

Abstract

The majority of explanations that have emerged from XAI research are static and have been based on the assumption that there is a single message to be conveyed by the explanation. Thus, empowering humans to explore explanations for algorithmic decisions interactively seems promising and there are already studies in this regard. For example, research is investigating how data scientists work or tools are being offered to help data scientists use and understand ML algorithms. However, we still need to adjust these interfaces to diverse target groups and capitalize on interactivity, especially for non-expert users.

Thus, as part of this Bachelor thesis, a literature review will be conducted to explore the state-of-the-art on empirical user-based studies in XAI research about interactive explanations and interfaces for diverse target groups.