Sam Babichenko

Reading notes

Benign overfitting

Estimators that fit noisy training data exactly and still predict well.