adaQN: An Adaptive Quasi-Newton Algorithm for Training RNNs.

Nitish Shirish Keskar Albert S. Berahas

Published in: ECML/PKDD (1) (2016)

Keyphrases

cost function
worst case
objective function
np hard
dynamic programming
neural network training
neural network
training algorithm
linear programming
optimization algorithm
quasi newton
global optimum
convergence rate
convex hull
segmentation algorithm
simulated annealing
k means
training set
optimal solution