AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

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

Xin Ning, Qiankun Li, Xiaolong Huang, Qiupu Chen, He Feng, Weijun Li, Prayag Tiwari, X Y Liu

TPAMI 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Low-rank compression and merging

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add DualOpt_Scratch real-time layer-wise weight decayAdam (ResNet-50) · CIFAR-100 (from scratch)accuracy: 63.47 -> 64.67Improved
add DualOpt_ScratchSGD (ResNet-50) · CUB-200-2011 (from scratch)accuracy: 15.59 -> 20.19Improved
add DualOpt_ScratchAdam (Inception) · OfficeHome (from scratch)accuracy: 31.15 -> 42.98Improved
add DualOpt_ScratchSGD (ResNet-50) · OfficeHome (from scratch)accuracy: 43.81 -> 42.01Hurt
add DualOpt fine-tuning (weight rollback, layer-wise penalty, diversified decay)Full fine-tuning (ViT-B) · PACSaccuracy: 87.79 -> 94.38Improved
add DualOpt fine-tuningFull fine-tuning (ConvNeXt-B) · CIFAR-100accuracy: 92.72 -> 92.86No clear effect
add DualOptViT-B from scratch · ImageNettop-1 accuracy: 81.34 -> 82.58Improved
add DualOptSwin-B from scratch · ImageNettop-1 accuracy: 83.05 -> 83.89Improved
add DualOptMask R-CNN full fine-tuning · COCO2017Bbox mAP: 40.20 -> 41.10Improved
add DualOptUperNet full fine-tuning · ADE20KmIoU: 43.65 -> 44.62Improved
add layer-wise decay rate lambda*i/n instead of uniform decay (ablation)DualOpt_Scratch (Adam) · CIFAR-100accuracy: 63.47 -> 64.67Improved
use inverse layer-wise decay instead of proposed (ablation)DualOpt_Scratch (Adam) · CIFAR-100accuracy: 64.67 -> 62.61Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy