Bilevel Learning with Inexact Stochastic Gradients

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Salehi, Mohammad Sadegh, Mukherjee, Subhadip, Roberts, Lindon, Ehrhardt, Matthias J.
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910949210849280
author Salehi, Mohammad Sadegh
Mukherjee, Subhadip
Roberts, Lindon
Ehrhardt, Matthias J.
author_facet Salehi, Mohammad Sadegh
Mukherjee, Subhadip
Roberts, Lindon
Ehrhardt, Matthias J.
contents Bilevel learning has gained prominence in machine learning, inverse problems, and imaging applications, including hyperparameter optimization, learning data-adaptive regularizers, and optimizing forward operators. The large-scale nature of these problems has led to the development of inexact and computationally efficient methods. Existing adaptive methods predominantly rely on deterministic formulations, while stochastic approaches often adopt a doubly-stochastic framework with impractical variance assumptions, enforces a fixed number of lower-level iterations, and requires extensive tuning. In this work, we focus on bilevel learning with strongly convex lower-level problems and a nonconvex sum-of-functions in the upper-level. Stochasticity arises from data sampling in the upper-level which leads to inexact stochastic hypergradients. We establish their connection to state-of-the-art stochastic optimization theory for nonconvex objectives. Furthermore, we prove the convergence of inexact stochastic bilevel optimization under mild assumptions. Our empirical results highlight significant speed-ups and improved generalization in imaging tasks such as image denoising and deblurring in comparison with adaptive deterministic bilevel methods.
format Preprint
id arxiv_https___arxiv_org_abs_2412_12049
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Bilevel Learning with Inexact Stochastic Gradients
Salehi, Mohammad Sadegh
Mukherjee, Subhadip
Roberts, Lindon
Ehrhardt, Matthias J.
Optimization and Control
Machine Learning
Bilevel learning has gained prominence in machine learning, inverse problems, and imaging applications, including hyperparameter optimization, learning data-adaptive regularizers, and optimizing forward operators. The large-scale nature of these problems has led to the development of inexact and computationally efficient methods. Existing adaptive methods predominantly rely on deterministic formulations, while stochastic approaches often adopt a doubly-stochastic framework with impractical variance assumptions, enforces a fixed number of lower-level iterations, and requires extensive tuning. In this work, we focus on bilevel learning with strongly convex lower-level problems and a nonconvex sum-of-functions in the upper-level. Stochasticity arises from data sampling in the upper-level which leads to inexact stochastic hypergradients. We establish their connection to state-of-the-art stochastic optimization theory for nonconvex objectives. Furthermore, we prove the convergence of inexact stochastic bilevel optimization under mild assumptions. Our empirical results highlight significant speed-ups and improved generalization in imaging tasks such as image denoising and deblurring in comparison with adaptive deterministic bilevel methods.
title Bilevel Learning with Inexact Stochastic Gradients
topic Optimization and Control
Machine Learning
url https://arxiv.org/abs/2412.12049