Speeding up Hyper-parameter Optimization by Extrapolation of Learning Curves Using Previous Builds.

Akshay Chandrashekaran Ian R. Lane

Published in: ECML/PKDD (1) (2017)

Keyphrases

learning curves
hyperparameters
learning curve
average case
gaussian processes
random sampling
power law
closed form
prior information
incomplete data
parameter settings
data sets
cross validation
support vector
learning algorithm
model selection
em algorithm
maximum likelihood
decision trees