Saved in:
Bibliographic Details
Main Authors: Graca, Manuel, Silveira, L. Miguel, Oliveira, Arlindo, Liu, Frank
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.16017
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911688490483712
author Graca, Manuel
Silveira, L. Miguel
Oliveira, Arlindo
Liu, Frank
author_facet Graca, Manuel
Silveira, L. Miguel
Oliveira, Arlindo
Liu, Frank
contents In this paper, we present CT-AGD (Curvature-Tuned Accelerated Gradient Descent), an optimization method for non-convex optimization problems in deep learning training tasks. CT-AGD is a general boosting procedure that accelerates first-order methods by explicitly capturing the local curvature using finite-difference quotients, and the development of heuristics aimed at mitigating noise and bias introduced by stochastic mini-batch training. CT-AGD has a comparable storage and computational overhead as adaptive gradient methods such as Adam. Our extensive experiments demonstrate that CT-AGD achieves the same level of accuracy as the baseline first-order methods, yet reduces the required training epochs by 33% on average.
format Preprint
id arxiv_https___arxiv_org_abs_2605_16017
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Accelerated Gradient Descent for Faster Convergence with Minimal Overhead
Graca, Manuel
Silveira, L. Miguel
Oliveira, Arlindo
Liu, Frank
Machine Learning
In this paper, we present CT-AGD (Curvature-Tuned Accelerated Gradient Descent), an optimization method for non-convex optimization problems in deep learning training tasks. CT-AGD is a general boosting procedure that accelerates first-order methods by explicitly capturing the local curvature using finite-difference quotients, and the development of heuristics aimed at mitigating noise and bias introduced by stochastic mini-batch training. CT-AGD has a comparable storage and computational overhead as adaptive gradient methods such as Adam. Our extensive experiments demonstrate that CT-AGD achieves the same level of accuracy as the baseline first-order methods, yet reduces the required training epochs by 33% on average.
title Accelerated Gradient Descent for Faster Convergence with Minimal Overhead
topic Machine Learning
url https://arxiv.org/abs/2605.16017