paper-with-me

Papers

Rotational Equilibrium: How Weight Decay Balances Learning Across Neural Networks

2023-05-26 · Atli Kosson, Bettina Messmer, Martin Jaggi

This study investigates how weight decay affects the update behavior of individual neurons in deep neural networks through a combination of applied analysis and experimentation. Weight decay can cause the expected magnitude and angular updates of a neuron's weight vector to converge to a steady state we call rotational equilibrium. These states can be highly homogeneous, effectively balancing the average rotation -- a proxy for the effective learning rate -- across different layers and neurons. Our work analyzes these dynamics across optimizers like Adam, Lion, and SGD with momentum, offering a new simple perspective on training that elucidates the efficacy of widely used but poorly understood methods in deep learning. We demonstrate how balanced rotation plays a key role in the effectiveness of normalization like Weight Standardization, as well as that of AdamW over Adam with L2-regularization. Finally, we show that explicitly controlling the rotation provides the benefits of weight decay while substantially reducing the need for learning rate warmup.

📄 PDF Abstract BibTeX arXiv:2305.17212

Code (2)

epfml/req 공식 구현 pytorch
epfml/rotational-optimizers 공식 구현 pytorch

Tasks

L2 Regularization

Methods 이 논문이 사용한 방법론

Weight Standardization Weight Standardization is a normalization technique that smooths the loss landscape by standardizing the weights in convolutional layers. Different from the previous…
Weight Decay 설명 없음
Adam 설명 없음
Lion The Lion optimizer is discovered by symbolic program search. It is more memory-efficient than most adaptive optimizers as it only needs to momentum. The update of Lion is produced…
SGD Stochastic Gradient Descent is an iterative optimization technique that uses minibatches of data to form an expectation of the gradient, rather than the full gradient using…
AdamW AdamW is a stochastic optimization method that modifies the typical implementation of weight decay in Adam, by decoupling [weight…

Similar Papers 제목 키워드 기반

Demystifying Manifold Constraints in LLM Pre-training

2026-05-06 · Kang An, Jiaxiang Li, Donald Goldfarb, Shiqian Ma arxiv

The empirical success of large language model (LLM) pre-training relies heavily on heuristic stabilization techniques, such as explicit normalization layers and weight decay. While recent constrained optimization approac…

Fantastic Pretraining Optimizers and Where to Find Them II: Hyperball Optimization

2026-06-15 · Kaiyue Wen, Xingyu Dang, Kaifeng Lyu, Tengyu Ma 외 arxiv

Matrix based optimizers such as Muon can substantially speed up language model pretraining, but their gains over AdamW are observed to shrink as model size and data scale grow when using standard constant decoupled weigh…

Why Do We Need Weight Decay in Modern Deep Learning?

2023-10-06 · Francesco D'Angelo, Maksym Andriushchenko, Aditya Varre, Nicolas Flammarion

Weight decay is a broadly used technique for training state-of-the-art deep networks from image classification to large language models. Despite its widespread usage and being extensively studied in the classical literat…

Deep Learningimage-classificationImage ClassificationLearning Theory+1

Spherical Motion Dynamics: Learning Dynamics of Normalized Neural Network using SGD and Weight Decay

2021-12-01 · NeurIPS 2021 12 · Ruosi Wan, Zhanxing Zhu, Xiangyu Zhang, Jian Sun

In this paper, we comprehensively reveal the learning dynamics of normalized neural network using Stochastic Gradient Descent (with momentum) and Weight Decay (WD), named as Spherical Motion Dynamics (SMD). Most related …

Spherical Motion Dynamics: Learning Dynamics of Neural Network with Normalization, Weight Decay, and SGD

2020-06-15 · Ruosi Wan, Zhanxing Zhu, Xiangyu Zhang, Jian Sun

In this work, we comprehensively reveal the learning dynamics of neural network with normalization, weight decay (WD), and SGD (with momentum), named as Spherical Motion Dynamics (SMD). Most related works study SMD by fo…