LGNEOCNov 14, 2017

Decoupled Weight Decay Regularization

arXiv:1711.05101v35213 citationsHas Code
Originality Incremental advance
AI Analysis

This addresses a subtle but impactful issue for machine learning practitioners using adaptive optimizers, leading to better model performance and adoption in major frameworks.

The paper tackles the problem that L2 regularization and weight decay are not equivalent for adaptive gradient algorithms like Adam, proposing a simple modification to decouple weight decay from the optimization steps. This modification improves Adam's generalization performance, allowing it to compete with SGD with momentum on image classification datasets.

L$_2$ regularization and weight decay regularization are equivalent for standard stochastic gradient descent (when rescaled by the learning rate), but as we demonstrate this is \emph{not} the case for adaptive gradient algorithms, such as Adam. While common implementations of these algorithms employ L$_2$ regularization (often calling it "weight decay" in what may be misleading due to the inequivalence we expose), we propose a simple modification to recover the original formulation of weight decay regularization by \emph{decoupling} the weight decay from the optimization steps taken w.r.t. the loss function. We provide empirical evidence that our proposed modification (i) decouples the optimal choice of weight decay factor from the setting of the learning rate for both standard SGD and Adam and (ii) substantially improves Adam's generalization performance, allowing it to compete with SGD with momentum on image classification datasets (on which it was previously typically outperformed by the latter). Our proposed decoupled weight decay has already been adopted by many researchers, and the community has implemented it in TensorFlow and PyTorch; the complete source code for our experiments is available at https://github.com/loshchil/AdamW-and-SGDW

Code Implementations23 repos

Data from Papers with Code (CC-BY-SA-4.0)

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes