Communication Quantization for Data-Parallel Training of Deep Neural Networks.

Nikoli Dryden Tim Moon Sam Ade Jacobs Brian Van Essen

Published in: MLHPC@SC (2016)

Keyphrases

neural network
data sets
data analysis
synthetic data
data points
data distribution
data collection
input data

sensor data
missing data
database
knowledge discovery
data structure
feed forward neural networks
training algorithm
training process

parallel processing
network structure
databases
learning algorithm
high quality
end users
training samples

statistical analysis
feature space
data processing
image quality
artificial neural networks
supervised learning
data sources