Dynamically Weighted Momentum with Adaptive Step Sizes for Efficient Deep Network Training

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wang, Zhifeng, Li, Longlong, Zeng, Chunyan
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914121602039808
author Wang, Zhifeng
Li, Longlong
Zeng, Chunyan
author_facet Wang, Zhifeng
Li, Longlong
Zeng, Chunyan
contents Within the current sphere of deep learning research, despite the extensive application of optimization algorithms such as Stochastic Gradient Descent (SGD) and Adaptive Moment Estimation (Adam), there remains a pronounced inadequacy in their capability to address fluctuations in learning efficiency, meet the demands of complex models, and tackle non-convex optimization issues. These challenges primarily arise from the algorithms' limitations in handling complex data structures and models, for instance, difficulties in selecting an appropriate learning rate, avoiding local optima, and navigating through high-dimensional spaces. To address these issues, this paper introduces a novel optimization algorithm named DWMGrad. This algorithm, building on the foundations of traditional methods, incorporates a dynamic guidance mechanism reliant on historical data to dynamically update momentum and learning rates. This allows the optimizer to flexibly adjust its reliance on historical information, adapting to various training scenarios. This strategy not only enables the optimizer to better adapt to changing environments and task complexities but also, as validated through extensive experimentation, demonstrates DWMGrad's ability to achieve faster convergence rates and higher accuracies under a multitude of scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2510_25042
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Dynamically Weighted Momentum with Adaptive Step Sizes for Efficient Deep Network Training
Wang, Zhifeng
Li, Longlong
Zeng, Chunyan
Machine Learning
Neural and Evolutionary Computing
Within the current sphere of deep learning research, despite the extensive application of optimization algorithms such as Stochastic Gradient Descent (SGD) and Adaptive Moment Estimation (Adam), there remains a pronounced inadequacy in their capability to address fluctuations in learning efficiency, meet the demands of complex models, and tackle non-convex optimization issues. These challenges primarily arise from the algorithms' limitations in handling complex data structures and models, for instance, difficulties in selecting an appropriate learning rate, avoiding local optima, and navigating through high-dimensional spaces. To address these issues, this paper introduces a novel optimization algorithm named DWMGrad. This algorithm, building on the foundations of traditional methods, incorporates a dynamic guidance mechanism reliant on historical data to dynamically update momentum and learning rates. This allows the optimizer to flexibly adjust its reliance on historical information, adapting to various training scenarios. This strategy not only enables the optimizer to better adapt to changing environments and task complexities but also, as validated through extensive experimentation, demonstrates DWMGrad's ability to achieve faster convergence rates and higher accuracies under a multitude of scenarios.
title Dynamically Weighted Momentum with Adaptive Step Sizes for Efficient Deep Network Training
topic Machine Learning
Neural and Evolutionary Computing
url https://arxiv.org/abs/2510.25042