Neural Network Optimization Reimagined: Decoupled Techniques for Scratch and Fine-Tuning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ning, Xin, Li, Qiankun, Huang, Xiaolong, Chen, Qiupu, He, Feng, Li, Weijun, Tiwari, Prayag, Liu, Xinwang
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911621773787136
author Ning, Xin
Li, Qiankun
Huang, Xiaolong
Chen, Qiupu
He, Feng
Li, Weijun
Tiwari, Prayag
Liu, Xinwang
author_facet Ning, Xin
Li, Qiankun
Huang, Xiaolong
Chen, Qiupu
He, Feng
Li, Weijun
Tiwari, Prayag
Liu, Xinwang
contents With the accumulation of resources in the era of big data and the rise of pre-trained models in deep learning, optimizing neural networks for various tasks often involves different strategies for fine-tuning pre-trained models versus training from scratch. However, existing optimizers primarily focus on reducing the loss function by updating model parameters, without fully addressing the unique demands of these two major paradigms. In this paper, we propose DualOpt, a novel approach that decouples optimization techniques specifically tailored for these distinct training scenarios. For training from scratch, we introduce real-time layer-wise weight decay, designed to enhance both convergence and generalization by aligning with the characteristics of weight updates and network architecture. For more importantly fine-tuning, we integrate weight rollback with the optimizer, incorporating a rollback term into each weight update step. This ensures consistency in the weight distribution between upstream and downstream models, effectively mitigating knowledge forgetting and improving fine-tuning performance. Additionally, we extend the layer-wise weight decay to dynamically adjust the rollback levels across layers, adapting to the varying demands of different downstream tasks. Extensive experiments across diverse tasks, including image classification, object detection, semantic segmentation, and instance segmentation, demonstrate the broad applicability and state-of-the-art performance of DualOpt. Code is available at https://github.com/qklee-lz/OLOR-AAAI-2024.
format Preprint
id arxiv_https___arxiv_org_abs_2604_22838
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Neural Network Optimization Reimagined: Decoupled Techniques for Scratch and Fine-Tuning
Ning, Xin
Li, Qiankun
Huang, Xiaolong
Chen, Qiupu
He, Feng
Li, Weijun
Tiwari, Prayag
Liu, Xinwang
Computer Vision and Pattern Recognition
Artificial Intelligence
With the accumulation of resources in the era of big data and the rise of pre-trained models in deep learning, optimizing neural networks for various tasks often involves different strategies for fine-tuning pre-trained models versus training from scratch. However, existing optimizers primarily focus on reducing the loss function by updating model parameters, without fully addressing the unique demands of these two major paradigms. In this paper, we propose DualOpt, a novel approach that decouples optimization techniques specifically tailored for these distinct training scenarios. For training from scratch, we introduce real-time layer-wise weight decay, designed to enhance both convergence and generalization by aligning with the characteristics of weight updates and network architecture. For more importantly fine-tuning, we integrate weight rollback with the optimizer, incorporating a rollback term into each weight update step. This ensures consistency in the weight distribution between upstream and downstream models, effectively mitigating knowledge forgetting and improving fine-tuning performance. Additionally, we extend the layer-wise weight decay to dynamically adjust the rollback levels across layers, adapting to the varying demands of different downstream tasks. Extensive experiments across diverse tasks, including image classification, object detection, semantic segmentation, and instance segmentation, demonstrate the broad applicability and state-of-the-art performance of DualOpt. Code is available at https://github.com/qklee-lz/OLOR-AAAI-2024.
title Neural Network Optimization Reimagined: Decoupled Techniques for Scratch and Fine-Tuning
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2604.22838