Parallelizing Over Artificial Neural Network Training Runs with Multigrid

08/07/2017
by   Jacob B. Schroder, et al.
0

Artificial neural networks are a popular and effective machine learning technique. Great progress has been made parallelizing the expensive training phase of an individual network, leading to highly specialized pieces of hardware, many based on GPU-type architectures, and more concurrent algorithms such as synthetic gradients. However, the training phase continues to be a bottleneck, where the training data must be processed serially over thousands of individual training runs. This work considers a multigrid reduction in time (MGRIT) algorithm that is able to parallelize over the thousands of training runs and converge to the exact same solution as traditional training would provide. MGRIT was originally developed to provide parallelism for time evolution problems that serially step through a finite number of time-steps. This work recasts the training of a neural network similarly, treating neural network training as an evolution equation that evolves the network weights from one step to the next. Thus, this work concerns distributed computing approaches for neural networks, but is distinct from other approaches which seek to parallelize only over individual training runs. The work concludes with supporting numerical results for two model problems.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
05/25/2018

Training of photonic neural networks through in situ backpropagation

Recently, integrated optics has gained interest as a hardware platform f...
research
11/12/2019

92c/MFlops/s, Ultra-Large-Scale Neural-Network Training on a PIII Cluster

Artificial neural networks with millions of adjustable parameters and a ...
research
10/30/2016

A Theoretical Study of The Relationship Between Whole An ELM Network and Its Subnetworks

A biological neural network is constituted by numerous subnetworks and m...
research
11/08/2018

Measuring the Effects of Data Parallelism on Neural Network Training

Recent hardware developments have made unprecedented amounts of data par...
research
06/14/2018

Multilevel Artificial Neural Network Training for Spatially Correlated Learning

Multigrid modeling algorithms are a technique used to accelerate relaxat...
research
01/18/2017

On the Performance of Network Parallel Training in Artificial Neural Networks

Artificial Neural Networks (ANNs) have received increasing attention in ...
research
02/14/2019

Superposition of many models into one

We present a method for storing multiple models within a single set of p...

Please sign up or login with your details

Forgot password? Click here to reset