1 paper
Florentin Guth, Brice Ménard, Gaspar Rochette +1
A central question in deep learning is to understand the functions learned by deep networks. What is their approximation class? Do the learned weights and representations depend on…