Multiple Descent: Design Your Own Generalization Curve
Abstract
This paper explores the generalization loss of linear regression in variably parameterized families of models, both under-parameterized and over-parameterized. We show that the generalization curve can have an arbitrary number of peaks, and moreover, the locations of those peaks can be explicitly controlled. Our results highlight the fact that both the classical U-shaped generalization curve and the recently observed double descent curve are not intrinsic properties of the model family. Instead, their emergence is due to the interaction between the properties of the data and the inductive biases of learning algorithms.
Cite
Text
Chen et al. "Multiple Descent: Design Your Own Generalization Curve." Neural Information Processing Systems, 2021.Markdown
[Chen et al. "Multiple Descent: Design Your Own Generalization Curve." Neural Information Processing Systems, 2021.](https://mlanthology.org/neurips/2021/chen2021neurips-multiple/)BibTeX
@inproceedings{chen2021neurips-multiple,
title = {{Multiple Descent: Design Your Own Generalization Curve}},
author = {Chen, Lin and Min, Yifei and Belkin, Mikhail and Karbasi, Amin},
booktitle = {Neural Information Processing Systems},
year = {2021},
url = {https://mlanthology.org/neurips/2021/chen2021neurips-multiple/}
}