DeepAI AI Chat
Log In Sign Up

Accumulated Gradient Normalization

by   Joeri Hermans, et al.
Université de Liège
Maastricht University

This work addresses the instability in asynchronous data parallel optimization. It does so by introducing a novel distributed optimizer which is able to efficiently optimize a centralized model under communication constraints. The optimizer achieves this by pushing a normalized sequence of first-order gradients to a parameter server. This implies that the magnitude of a worker delta is smaller compared to an accumulated gradient, and provides a better direction towards a minimum compared to first-order gradients, which in turn also forces possible implicit momentum fluctuations to be more aligned since we make the assumption that all workers contribute towards a single minima. As a result, our approach mitigates the parameter staleness problem more effectively since staleness in asynchrony induces (implicit) momentum, and achieves a better convergence rate compared to other optimizers such as asynchronous EASGD and DynSGD, which we show empirically.


page 1

page 2

page 3

page 4


Gradient Sparification for Asynchronous Distributed Training

Modern large scale machine learning applications require stochastic opti...

Sharper Convergence Guarantees for Asynchronous SGD for Distributed and Federated Learning

We study the asynchronous stochastic gradient descent algorithm for dist...

Momentum Centering and Asynchronous Update for Adaptive Gradient Methods

We propose ACProp (Asynchronous-centering-Prop), an adaptive optimizer w...

SPI-Optimizer: an integral-Separated PI Controller for Stochastic Optimization

To overcome the oscillation problem in the classical momentum-based opti...

On the Convergence of Adam and Adagrad

We provide a simple proof of the convergence of the optimization algorit...

Anytime Minibatch with Delayed Gradients

Distributed optimization is widely deployed in practice to solve a broad...

The effect of Target Normalization and Momentum on Dying ReLU

Optimizing parameters with momentum, normalizing data values, and using ...