When to restart? Exploring escalating restarts on convergence

TL;DR

This paper introduces SGD-ER, an adaptive learning rate restart strategy that triggers escalations based on training stagnation, improving convergence and test accuracy across multiple datasets and architectures.

Contribution

The paper proposes a novel convergence-aware restart method, SGD-ER, which adaptively escalates learning rates upon stagnation to escape local minima.

Findings

SGD-ER improves test accuracy by 0.5-4.5% over standard schedulers.

SGD-ER effectively escapes sharp local minima.

The method is validated on CIFAR-10, CIFAR-100, and TinyImageNet with various architectures.

Abstract

Learning rate scheduling plays a critical role in the optimization of deep neural networks, directly influencing convergence speed, stability, and generalization. While existing schedulers such as cosine annealing, cyclical learning rates, and warm restarts have shown promise, they often rely on fixed or periodic triggers that are agnostic to the training dynamics, such as stagnation or convergence behavior. In this work, we propose a simple yet effective strategy, which we call Stochastic Gradient Descent with Escalating Restarts (SGD-ER). It adaptively increases the learning rate upon convergence. Our method monitors training progress and triggers restarts when stagnation is detected, linearly escalating the learning rate to escape sharp local minima and explore flatter regions of the loss landscape. We evaluate SGD-ER across CIFAR-10, CIFAR-100, and TinyImageNet on a range of architectures including ResNet-18/34/50, VGG-16, and DenseNet-101. Compared to standard schedulers, SGD-ER improves test accuracy by 0.5-4.5%, demonstrating the benefit of convergence-aware escalating restarts for better local optima.