paper-with-me

Papers

OUIDecay: Adaptive Layer-wise Weight Decay for CNNs Using Online Activation Patterns

2026-05-11 · Alberto Fernández-Hernández, Jose I. Mestre, Cristian Pérez-Corral, Manuel F. Dolz, Jose Duato, Enrique S. Quintana-Ortí arxiv

Weight decay remains one of the most widely used regularization mechanisms for training convolutional neural networks, yet it is still commonly applied as a fixed coefficient shared by all layers throughout training. This uniform treatment ignores that different layers may follow different structural dynamics and therefore may require different regularization strengths. In this work, we propose OUIDecay, an adaptive layer-wise and time-dependent weight decay scheduler for CNNs driven by the Overfitting-Underfitting Indicator (OUI), an activation-based metric previously shown to provide early information about regularization quality. OUIDecay uses a lightweight batch-based formulation of OUI to monitor the structural behavior of each layer online and periodically rescales its weight decay relative to the other layers in the network. Unlike gradient-based adaptive decay methods, our approach relies on functional information extracted from activation patterns and does not require validation data. Experiments on EfficientNet-B0 with Stanford Cars, ResNet50 with Food101, DenseNet121 with CIFAR100, and MobileNetV2 with CIFAR10 show that OUIDecay achieves the best mean best-validation-loss in 7 out of 8 evaluated settings. These results indicate that activation-driven weight decay adaptation is a practical and effective alternative to fixed decay and gradient-based adaptive decay, while keeping the method lightweight and suitable for online use.

📄 PDF Abstract BibTeX arXiv:2605.10161

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

AlphaDecay: Module-wise Weight Decay for Heavy-Tailed Balancing in LLMs

2025-06-17 · Di He, Ajay Jaiswal, Songjun Tu, Li Shen 외

Weight decay is a standard regularization technique for training large language models (LLMs). While it is common to assign a uniform decay rate to every layer, this approach overlooks the structural diversity of LLMs an…

Diversity

Stochastic Gradient Methods with Layer-wise Adaptive Moments for Training of Deep Networks

2019-05-27 · Boris Ginsburg, Patrice Castonguay, Oleksii Hrinchuk, Oleksii Kuchaiev 외

We propose NovoGrad, an adaptive stochastic gradient descent method with layer-wise gradient normalization and decoupled weight decay. In our experiments on neural networks for image classification, speech recognition, m…

General Classificationimage-classificationLanguage Modelingspeech-recognition+2

Training Deep Networks with Stochastic Gradient Normalized by Layerwise Adaptive Second Moments

2020-01-01 · ICLR 2020 1 · Boris Ginsburg, Patrice Castonguay, Oleksii Hrinchuk, Oleksii Kuchaiev 외

We propose NovoGrad, an adaptive stochastic gradient descent method with layer-wise gradient normalization and decoupled weight decay. In our experiments on neural networks for image classification, speech recognition, m…

General Classificationimage-classificationImage ClassificationLanguage Modeling+5

Accelerating Learnt Video Codecs with Gradient Decay and Layer-wise Distillation

2023-12-05 · Tianhao Peng, Ge Gao, Heming Sun, Fan Zhang 외

In recent years, end-to-end learnt video codecs have demonstrated their potential to compete with conventional coding algorithms in term of compression efficiency. However, most learning-based video compression models ar…

DecoderVideo Compression

OrScale: Orthogonalised Optimization with Layer-Wise Trust-Ratio Scaling

2026-05-08 · Yuxuan Lou, Yang You arxiv

Muon improves neural-network training by orthogonalizing matrix-valued updates, but it leaves each layer's update magnitude controlled mostly by a global learning rate. We introduce OrScale, a trust-ratio extension of Mu…