Project 3: Design and Evaluation of User-centric XAI Methods

Many Artificial Intelligence (AI) systems constitute so-called "black boxes", which means that the reasons for their decisions and recommendations remain obscure to their users. Against this background, the research field of Explainable AI (XAI) provides approaches to automatically generate explanations along with AI systems’ outputs. These explanations are intended to make the functioning of the AI system as a whole or single decisions made by it transparent and comprehensible for lay users, domain experts, or developers. Within this project, we aim to contribute to raising the potential of XAI methods. To this end, new user-centric XAI methods will be designed and demonstrated for real-world applications. In addition, the users' perception of explanations generated by

First supervisor:

Prof. Dr. Matthias Klier, Institut für Business Analytics, Universität Ulm

 

Tandem partner:

Prof. Dr. Volker Herbort, Technische Hochschule Ulm

 

Consutling experts:

Prof. Dr. Henning Bruhn-Fujimoto, Institut für Optimierung und OR, Universität Ulm

Prof. Dr. Birte Glimm, Institut für Künstliche Intelligenz, Universität Ulm

Prof. Manfred Reichert, Insitut für Datenbanken und Informationssysteme, Universität Ulm

Prof. Dr. Stephan Schlüter, Technische Hoschschule Ulm