On June 11, 2025, M.Sc. Dirk Leffrang successfully completed his doctoral examination as part of his dissertation entitled:
“Imperfections of Machine Learning: Experimental Investigation of Human Advice-taking Behavior.”
In his dissertation, Mr. Leffrang explores the imperfections of ML-based decision support systems and how these imperfections affect users’ trust in the outputs of such systems. His work contributes to the ongoing discourse on algorithm aversion/appreciation, which examines how people respond to algorithmic recommendations: whether they distrust and avoid them (aversion) or trust and (perhaps even overly) appreciate them (appreciation).
While most existing research focuses on how different levels of algorithm explainability impact user trust, Mr. Leffrang's research emphasizes less-explored factors — such as the communication of uncertainty in time series forecasts, hallucinations by large language models (LLMs), or the energy consumption of complex neural networks.
Dirk Leffrang's dissertation was supervised and reviewed by Prof. Dr. Oliver Müller and Prof. Dr. Simon Thanh-Nam Trang. With the successful completion of his disputation, he has now taken the final step toward earning his doctorate. The chair extends its warmest congratulations on this significant achievement and wishes him all the best for his future, both professionally and personally.
