1 paper
Moritz Haas, David Holzmüller, Ulrike von Luxburg +1
The success of over-parameterized neural networks trained to near-zero training error has caused great interest in the phenomenon of benign overfitting, where estimators are statis…