Finding approximate local minima faster than gradient descent.

Naman Agarwal Zeyuan Allen Zhu Brian Bullins Elad Hazan Tengyu Ma

Published in: STOC (2017)

Keyphrases

cost function
decision making
multiscale
real time
knowledge base
metadata
objective function
method finds
learning rules
piecewise linear
loss function
expert systems
support vector
learning environment
website
feature selection
social networks
artificial intelligence