Hey there, I’m Fynn!
I am a PhD researcher at TU Graz, working on information-theoretic analyses of neural networks. My research asks how deep learning models use, compress, and preserve information: what they retain, what they discard, and how redundancy, synergy, and structure shape generalization. I am especially interested in mutual information, partial information decomposition, and their role in understanding different neural network architectures. More broadly, my work sits between machine learning and information theory, with the goal of making deep learning more interpretable, robust, and resource-efficient.
PhD researcher at the Signal Processing and Speech Communication Lab, TU Graz 
ELLIS PhD student, advised byBernhard Geiger
Giuseppe Durisi
Working on the FWF project Information Planes and Decompositions