This is an interim version of our Electronic Legal Deposit Catalogue-eJournals and eBooks while we continue to recover from a cyber-attack.
Deep double descent: where bigger models and more data hurt*This article is an updated version of: Nakkiran P, Kaplun G, Bansal Y, Yang T, Barak B and Sutskever I 2020 Deep double descent: where bigger models and more data hurt Int. Conf. Learning Representations. (29th December 2021)
Record Type:
Journal Article
Title:
Deep double descent: where bigger models and more data hurt*This article is an updated version of: Nakkiran P, Kaplun G, Bansal Y, Yang T, Barak B and Sutskever I 2020 Deep double descent: where bigger models and more data hurt Int. Conf. Learning Representations. (29th December 2021)
Main Title:
Deep double descent: where bigger models and more data hurt*This article is an updated version of: Nakkiran P, Kaplun G, Bansal Y, Yang T, Barak B and Sutskever I 2020 Deep double descent: where bigger models and more data hurt Int. Conf. Learning Representations.
Abstract: We show that a variety of modern deep learning tasks exhibit a 'double-descent' phenomenon where, as we increase model size, performance first gets worse and then gets better. Moreover, we show that double descent occurs not just as a function of model size, but also as a function of the number of training epochs. We unify the above phenomena by defining a new complexity measure we call the effective model complexity and conjecture a generalized double descent with respect to this measure. Furthermore, our notion of model complexity allows us to identify certain regimes where increasing (even quadrupling) the number of train samples actually hurts test performance.