Optimistic Rates: A Unifying Theory for Interpolation Learning and Regularization in Linear Regression
ACM / IMS Journal of Data Science · 2024
DOI: 10.1145/3594234 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACM / IMS Journal of Data Science, 2024. Verfasst von Lijia Zhou und 3 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 54 Literaturangaben.
Abstract (Originalsprache)
We study a localized notion of uniform convergence known as an “optimistic rate” [ 34 , 39 ] for linear regression with Gaussian data. Our refined analysis avoids the hidden constant and logarithmic factor in existing results, which are known to be crucial in high-dimensional settings, especially for understanding interpolation learning. As a special case, our analysis recovers the guarantee from Koehler et al. [ 21 ], which tightly characterizes the population risk of low-norm interpolators under the benign overfitting conditions. Our optimistic rate bound, though, also analyzes predictors with arbitrary training error. This allows us to recover some classical statistical guarantees for ridge and LASSO regression under random designs and helps us obtain a precise understanding of the excess risk of near-interpolators in the over-parameterized regime.
Zitieren
APA
Zhou, L., Koehler, F., Sutherland, D. J., & Srebro, N. (2024). Optimistic Rates: A Unifying Theory for Interpolation Learning and Regularization in Linear Regression. ACM / IMS Journal of Data Science. https://doi.org/10.1145/3594234
Vancouver
Zhou L, Koehler F, Sutherland DJ, Srebro N. Optimistic Rates: A Unifying Theory for Interpolation Learning and Regularization in Linear Regression. ACM / IMS Journal of Data Science. 2024. doi:10.1145/3594234
BibTeX
@article{Zhou2024,
author = {Zhou, Lijia and Koehler, Frederic and Sutherland, Danica J. and Srebro, Nathan},
title = {Optimistic Rates: A Unifying Theory for Interpolation Learning and Regularization in Linear Regression},
journal = {ACM / IMS Journal of Data Science},
year = {2024},
doi = {10.1145/3594234}
}
EUVIMED