BDA-PCH: Block-Diagonal Approximation of Positive-Curvature Hessian for Training Neural Networks.

Sheng-Wei Chen Chun-Nan Chou Edward Y. Chang

Published in: CoRR (2018)

Keyphrases

block diagonal
neural network
training process
linear programming problems
semidefinite programming
pattern recognition
artificial neural networks
supervised learning
back propagation
training examples
linear programming
approximation algorithms
linear program
np hard
pairwise
optimal solution
training data