Gradient Descent Finds Global Minima for Generalizable Deep Neural Networks of Practical Sizes.

Kenji Kawaguchi Jiaoyang Huang

Published in: CoRR (2019)

Keyphrases

neural network
global minima
global minimum
cost function
learning rules
neural network model
back propagation
artificial neural networks
energy minimization
multi layer
training process
multiresolution
loss function
feed forward
recurrent neural networks
pattern recognition
objective function
real world