HOW ROTATIONAL INVARIANCE OF COMMON KERNELS PREVENTS GENERALIZATION IN HIGH DIMENSIONS

von DONHAUSER, KONSTANTIN
Veröffentlicht in International Conference on Machine Learning (38. : 2021 : Online) International Conference on Machine Learning (ICML 2021 ; Part 4 of 16 (2022)

Fast Rates for Noisy Interpolation Require Rethinking the Effect of Inductive Bias

von Donhauser, Konstantin
Veröffentlicht in International Conference on Machine Learning (39. : 2022 : Baltimore, Md.; Online) International Conference on Machine Learning (ICML 2022) ; Part 7 of 33 (2023)

Interpolation can hurt robust generalization even when there is no noise

von Donhauser, Konstantin
Veröffentlicht in NeurIPS (35. : 2021 : Online) 35th Conference on Neural Information Processing Systems (NeurIPS 2021) ; Volume 28 of 36 (2022)