The Deep Bootstrap Framework: Good Online Learners are Good Offline Generalizers

arXiv2010.08127MaRDI QIDQ6351457

Author name not available (Why is that?)

Publication date: 15 October 2020

Abstract: We propose a new framework for reasoning about generalization in deep learning. The core idea is to couple the Real World, where optimizers take stochastic gradient steps on the empirical loss, to an Ideal World, where optimizers take steps on the population loss. This leads to an alternate decomposition of test error into: (1) the Ideal World test error plus (2) the gap between the two worlds. If the gap (2) is universally small, this reduces the problem of generalization in offline learning to the problem of optimization in online learning. We then give empirical evidence that this gap between worlds can be small in realistic deep learning settings, in particular supervised image classification. For example, CNNs generalize better than MLPs on image distributions in the Real World, but this is "because" they optimize faster on the population loss in the Ideal World. This suggests our framework is a useful tool for understanding generalization in deep learning, and lays a foundation for future research in the area.

Has companion code repository: https://github.com/preetum/cifar5m

This page was built for publication: The Deep Bootstrap Framework: Good Online Learners are Good Offline Generalizers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6351457)