Zeroth-Order Optimization Finds Flat Minima

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Liang, Li, Bingcong, Thekumparampil, Kiran Koshy, Oh, Sewoong, Muehlebach, Michael, He, Niao
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908642692825088
author Zhang, Liang
Li, Bingcong
Thekumparampil, Kiran Koshy
Oh, Sewoong
Muehlebach, Michael
He, Niao
author_facet Zhang, Liang
Li, Bingcong
Thekumparampil, Kiran Koshy
Oh, Sewoong
Muehlebach, Michael
He, Niao
contents Zeroth-order methods are extensively used in machine learning applications where gradients are infeasible or expensive to compute, such as black-box attacks, reinforcement learning, and language model fine-tuning. Existing optimization theory focuses on convergence to an arbitrary stationary point, but less is known on the implicit regularization that provides a fine-grained characterization on which particular solutions are finally reached. We show that zeroth-order optimization with the standard two-point estimator favors solutions with small trace of Hessian, which is widely used in previous work to distinguish between sharp and flat minima. We further provide convergence rates of zeroth-order optimization to approximate flat minima for convex and sufficiently smooth functions, where flat minima are defined as the minimizers that achieve the smallest trace of Hessian among all optimal solutions. Experiments on binary classification tasks with convex losses and language model fine-tuning support our theoretical findings.
format Preprint
id arxiv_https___arxiv_org_abs_2506_05454
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Zeroth-Order Optimization Finds Flat Minima
Zhang, Liang
Li, Bingcong
Thekumparampil, Kiran Koshy
Oh, Sewoong
Muehlebach, Michael
He, Niao
Machine Learning
Artificial Intelligence
Optimization and Control
Zeroth-order methods are extensively used in machine learning applications where gradients are infeasible or expensive to compute, such as black-box attacks, reinforcement learning, and language model fine-tuning. Existing optimization theory focuses on convergence to an arbitrary stationary point, but less is known on the implicit regularization that provides a fine-grained characterization on which particular solutions are finally reached. We show that zeroth-order optimization with the standard two-point estimator favors solutions with small trace of Hessian, which is widely used in previous work to distinguish between sharp and flat minima. We further provide convergence rates of zeroth-order optimization to approximate flat minima for convex and sufficiently smooth functions, where flat minima are defined as the minimizers that achieve the smallest trace of Hessian among all optimal solutions. Experiments on binary classification tasks with convex losses and language model fine-tuning support our theoretical findings.
title Zeroth-Order Optimization Finds Flat Minima
topic Machine Learning
Artificial Intelligence
Optimization and Control
url https://arxiv.org/abs/2506.05454