Login / Signup
A Diffusion Theory for Deep Learning Dynamics: Stochastic Gradient Descent Escapes From Sharp Minima Exponentially Fast.
Zeke Xie
Issei Sato
Masashi Sugiyama
Published in:
CoRR (2020)
Keyphrases
</>
deep learning
stochastic gradient descent
loss function
least squares
unsupervised learning
step size
machine learning
random forests
matrix factorization
multi class
higher order
reinforcement learning
multiscale
image processing
multiple kernel learning
feature selection
data mining