Deep Double Descent: Where Bigger Models and More Data Hurt
Machine Learning
2019-12-06 v1 Computer Vision and Pattern Recognition
Neural and Evolutionary Computing
Machine Learning
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.
Cite
@article{arxiv.1912.02292,
title = {Deep Double Descent: Where Bigger Models and More Data Hurt},
author = {Preetum Nakkiran and Gal Kaplun and Yamini Bansal and Tristan Yang and Boaz Barak and Ilya Sutskever},
journal= {arXiv preprint arXiv:1912.02292},
year = {2019}
}
Comments
G.K. and Y.B. contributed equally