ICML 2015poster1174 citations
Gradient-based Hyperparameter Optimization through Reversible Learning
Dougal Maclaurin, David Duvenaud, Ryan Adams
Abstract
Tuning hyperparameters of learning algorithms is hard because gradients are usually unavailable. We compute exact gradients of cross-validation performance with respect to all hyperparameters by chaining derivatives backwards through the entire training procedure. These gradients allow us to optimize thousands of hyperparameters, including step-size and momentum schedules, weight initialization distributions, richly parameterized regularization schemes, and neural network architectures. We compute hyperparameter gradients by exactly reversing the dynamics of stochastic gradient descent with momentum.
BibTeX
@InProceedings{pmlr-v37-maclaurin15,
title = {Gradient-based Hyperparameter Optimization through Reversible Learning},
author = {Maclaurin, Dougal and Duvenaud, David and Adams, Ryan},
booktitle = {Proceedings of the 32nd International Conference on Machine Learning},
pages = {2113--2122},
year = {2015},
editor = {Bach, Francis and Blei, David},
volume = {37},
series = {Proceedings of Machine Learning Research},
address = {Lille, France},
month = {07--09 Jul},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v37/maclaurin15.pdf},
url = {https://proceedings.mlr.press/v37/maclaurin15.html},
abstract = {Tuning hyperparameters of learning algorithms is hard because gradients are usually unavailable. We compute exact gradients of cross-validation performance with respect to all hyperparameters by chaining derivatives backwards through the entire training procedure. These gradients allow us to optimize thousands of hyperparameters, including step-size and momentum schedules, weight initialization distributions, richly parameterized regularization schemes, and neural network architectures. We compute hyperparameter gradients by exactly reversing the dynamics of stochastic gradient descent with momentum.}
}