A robust adaptive stochastic gradient method for deep learning.

Çaglar Gülçehre Jose Sotelo Marcin Moczulski Yoshua Bengio

Published in: IJCNN (2017)

Keyphrases

deep learning
gradient method
unsupervised learning
convergence rate
unsupervised feature learning
machine learning
step size
negative matrix factorization
optimization methods
mental models
data sets
weakly supervised
learning algorithm
convergence speed