Research Focus

I am part of the research training group Dataninja (Trustworthy AI for Seamless Problem Solving: Next Generation Intelligence Joins Robust Data Analysis, https://dataninja.nrw), funded by the German federal state of North Rhine-Westphalia. My research goal is to develop better approximation techniques for the Shapley value. A game-theoretic concept to fairly distribute collective benefits achieved by a group of agents. It is commonly used in Explainable AI to attribute feature importance but also quantify the significance of data points, base learners in ensembles, and neurons in neural networks, among many other applications.

Publications