On Penalty Methods for Nonconvex Bilevel Optimization and First-Order Stochastic Approximation

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Kwon, Jeongyeol, Kwon, Dohyun, Wright, Stephen, Nowak, Robert
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916120564334592
author Kwon, Jeongyeol
Kwon, Dohyun
Wright, Stephen
Nowak, Robert
author_facet Kwon, Jeongyeol
Kwon, Dohyun
Wright, Stephen
Nowak, Robert
contents In this work, we study first-order algorithms for solving Bilevel Optimization (BO) where the objective functions are smooth but possibly nonconvex in both levels and the variables are restricted to closed convex sets. As a first step, we study the landscape of BO through the lens of penalty methods, in which the upper- and lower-level objectives are combined in a weighted sum with penalty parameter $σ> 0$. In particular, we establish a strong connection between the penalty function and the hyper-objective by explicitly characterizing the conditions under which the values and derivatives of the two must be $O(σ)$-close. A by-product of our analysis is the explicit formula for the gradient of hyper-objective when the lower-level problem has multiple solutions under minimal conditions, which could be of independent interest. Next, viewing the penalty formulation as $O(σ)$-approximation of the original BO, we propose first-order algorithms that find an $ε$-stationary solution by optimizing the penalty formulation with $σ= O(ε)$. When the perturbed lower-level problem uniformly satisfies the small-error proximal error-bound (EB) condition, we propose a first-order algorithm that converges to an $ε$-stationary point of the penalty function, using in total $O(ε^{-3})$ and $O(ε^{-7})$ accesses to first-order (stochastic) gradient oracles when the oracle is deterministic and oracles are noisy, respectively. Under an additional assumption on stochastic oracles, we show that the algorithm can be implemented in a fully {\it single-loop} manner, i.e., with $O(1)$ samples per iteration, and achieves the improved oracle-complexity of $O(ε^{-3})$ and $O(ε^{-5})$, respectively.
format Preprint
id arxiv_https___arxiv_org_abs_2309_01753
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle On Penalty Methods for Nonconvex Bilevel Optimization and First-Order Stochastic Approximation
Kwon, Jeongyeol
Kwon, Dohyun
Wright, Stephen
Nowak, Robert
Optimization and Control
Machine Learning
In this work, we study first-order algorithms for solving Bilevel Optimization (BO) where the objective functions are smooth but possibly nonconvex in both levels and the variables are restricted to closed convex sets. As a first step, we study the landscape of BO through the lens of penalty methods, in which the upper- and lower-level objectives are combined in a weighted sum with penalty parameter $σ> 0$. In particular, we establish a strong connection between the penalty function and the hyper-objective by explicitly characterizing the conditions under which the values and derivatives of the two must be $O(σ)$-close. A by-product of our analysis is the explicit formula for the gradient of hyper-objective when the lower-level problem has multiple solutions under minimal conditions, which could be of independent interest. Next, viewing the penalty formulation as $O(σ)$-approximation of the original BO, we propose first-order algorithms that find an $ε$-stationary solution by optimizing the penalty formulation with $σ= O(ε)$. When the perturbed lower-level problem uniformly satisfies the small-error proximal error-bound (EB) condition, we propose a first-order algorithm that converges to an $ε$-stationary point of the penalty function, using in total $O(ε^{-3})$ and $O(ε^{-7})$ accesses to first-order (stochastic) gradient oracles when the oracle is deterministic and oracles are noisy, respectively. Under an additional assumption on stochastic oracles, we show that the algorithm can be implemented in a fully {\it single-loop} manner, i.e., with $O(1)$ samples per iteration, and achieves the improved oracle-complexity of $O(ε^{-3})$ and $O(ε^{-5})$, respectively.
title On Penalty Methods for Nonconvex Bilevel Optimization and First-Order Stochastic Approximation
topic Optimization and Control
Machine Learning
url https://arxiv.org/abs/2309.01753