Saved in:
Bibliographic Details
Main Authors: Dagréou, Mathieu, Moreau, Thomas, Vaiter, Samuel, Ablin, Pierre
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2302.08766
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909409474510848
author Dagréou, Mathieu
Moreau, Thomas
Vaiter, Samuel
Ablin, Pierre
author_facet Dagréou, Mathieu
Moreau, Thomas
Vaiter, Samuel
Ablin, Pierre
contents Bilevel optimization problems, which are problems where two optimization problems are nested, have more and more applications in machine learning. In many practical cases, the upper and the lower objectives correspond to empirical risk minimization problems and therefore have a sum structure. In this context, we propose a bilevel extension of the celebrated SARAH algorithm. We demonstrate that the algorithm requires $\mathcal{O}((n+m)^{\frac12}\varepsilon^{-1})$ oracle calls to achieve $\varepsilon$-stationarity with $n+m$ the total number of samples, which improves over all previous bilevel algorithms. Moreover, we provide a lower bound on the number of oracle calls required to get an approximate stationary point of the objective function of the bilevel problem. This lower bound is attained by our algorithm, making it optimal in terms of sample complexity.
format Preprint
id arxiv_https___arxiv_org_abs_2302_08766
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle A Lower Bound and a Near-Optimal Algorithm for Bilevel Empirical Risk Minimization
Dagréou, Mathieu
Moreau, Thomas
Vaiter, Samuel
Ablin, Pierre
Machine Learning
Optimization and Control
Bilevel optimization problems, which are problems where two optimization problems are nested, have more and more applications in machine learning. In many practical cases, the upper and the lower objectives correspond to empirical risk minimization problems and therefore have a sum structure. In this context, we propose a bilevel extension of the celebrated SARAH algorithm. We demonstrate that the algorithm requires $\mathcal{O}((n+m)^{\frac12}\varepsilon^{-1})$ oracle calls to achieve $\varepsilon$-stationarity with $n+m$ the total number of samples, which improves over all previous bilevel algorithms. Moreover, we provide a lower bound on the number of oracle calls required to get an approximate stationary point of the objective function of the bilevel problem. This lower bound is attained by our algorithm, making it optimal in terms of sample complexity.
title A Lower Bound and a Near-Optimal Algorithm for Bilevel Empirical Risk Minimization
topic Machine Learning
Optimization and Control
url https://arxiv.org/abs/2302.08766