Classification vs regression in overparameterized regimes: Does the loss function matter?

Vidya Muthukumar, Adhyyan Narang, Vignesh Subramanian, Mikhail Belkin, Daniel J. Hsu, Anant Sahai