Generalization in Deep Learning
Kenji Kawaguchi, Yoshua Bengio, Leslie Pack Kaelbling · Cambridge University Press eBooks · 2022
This chapter provides theoretical insights into why and how deep learning can generalize well, despite its large capacity, complexity, possible algorithmic instability, non-robustness, and sharp minima, responding to an open question in the literature. We also discuss approaches to provide non-vacuous generalization guarantees for deep learning. On the basis of the theoretical observations, we propose new open problems.