Gunnar König

I am a postdoctoral researcher at the Tübingen AI Center, working with Ulrike von Luxburg on explainable AI, causal inference, and foundation models. Before Tübingen I did my PhD with Bernd Bischl and Moritz Grosse-Wentrup at LMU Munich, the Munich Center for Machine Learning (MCML), and the University of Vienna.

My research in explainable AI (XAI) revolves around three questions: Which conclusions do existing XAI methods allow, and under what assumptions? What should explanations look like to fulfill their purposes, with a particular focus on recourse, contestability, and scientific inference? And, importantly, how can we estimate them efficiently and accurately? These questions are closely tied to causality, so I often use tools from causal inference to answer them.

Lately, my focus has shifted towards the explainability of tabular foundation models. If you would like to exchange ideas or collaborate, feel free to reach out.