Training Two-Layer ReLU Networks with Gradient Descent is Inconsistent.

David Holzmüller Ingo Steinwart

Published in: J. Mach. Learn. Res. (2022)

Keyphrases

recurrent networks
training set
social networks
multi layer
objective function
cost function
stochastic gradient descent
training phase
network structure
error function
complex networks
online learning
end to end
network analysis
data sets
loss function
training samples
network size
machine learning