Konstantin Donhauser

Welcome

My name is Konstantin Donhauser and I am a first year PhD student at the ETH Zurich AI center supervised by Fanny Yang and co-supervised by Afonso Bandeira. I have a Masters and Bachelors degree in Electrical Engineering and a Bachelors degree in Mathematics all from the ETH Zurich.

My research focus can broadly be summarized as anything related to the theoretical aspects of machine learning. In particular, I have recently worked on topics related to high dimensional statistics, interpolation theory and adversarial robustness.

Preprints

Tight bounds for minimum l1-norm interpolation of noisy data
Wang Guillaume*, Konstantin Donhauser*, Fanny Yang
[preprint]

Publications

Interpolation can hurt robust generalization even when there is no noise
Konstantin Donhauser*, Alexandru Tifrea*, Michael Aerni, Reinhard Heckel, Fanny Yang
Will be presented at Neurips 2021
Presented at the Adversarial machine learning AML workshop (classification part) and Overparameterization OPPO workshop (regression part) at ICML 2021
[preprint], AdvML: [paper, poster,video], OPPO: [paper, poster]

How rotational invariance of common kernels prevents generalization in high dimensions
Konstantin Donhauser, Mingqi Wu, Fanny Yang
Presented at International Conference on Machine Learning, ICML 2021
[arxiv, proceedings, poster, video]

Supervised Students