SGD for Structured Nonconvex Functions: Learning Rates, Minibatching and Interpolation.

Robert M. Gower Othmane Sebbouh Nicolas Loizou

Published in: CoRR (2020)

Keyphrases

learning rate
learning algorithm
convergence rate
gaussian kernels
covering numbers
uniform convergence
convergence speed
data mining
global optimization
neural network
objective function
convergence theorem
stochastic gradient descent
matrix factorization
convex optimization
lower bound
feature selection