Fabian Gröger

I am a final-year PhD student at the University of Basel, supervised by Alexander Navarini and Marc Pouly. I also spent time as a visiting researcher at EPFL with Maria Brbić.

I study how representation spaces form, how to measure them reliably, and how to make them useful, especially when data is messy, scarce, or biased. My recent work on the Aristotelian Representation Hypothesis shows that much of the reported convergence across neural networks is a measurement artifact, and that what truly aligns are local neighborhood structures, not global geometry.

This perspective on representations threads through my work: I develop methods for multimodal alignment under limited supervision (NeurIPS ‘25), build tools for representation-based data quality auditing (NeurIPS ‘24), and apply these ideas to real-world problems in medical imaging and audio (ICASSP ‘25 & ‘26).

Outside of research, I spend my time with my family and on the bike.