Learning to Learn without Gradient Descent by Gradient Descent
Abstract
We learn recurrent neural network optimizers trained on simple synthetic functions by gradient descent. We show that these learned optimizers exhibit a remarkable degree of transfer in that they can be used to efficiently optimize a broad range of derivative-free black-box functions, including Gaussian process bandits, simple control objectives, global optimization benchmarks and hyper-parameter tuning tasks. Up to the training horizon, the learned optimizers learn to trade-off exploration and exploitation, and compare favourably with heavily engineered Bayesian optimization packages for hyper-parameter tuning.
Cite
@article{arxiv.1611.03824,
title = {Learning to Learn without Gradient Descent by Gradient Descent},
author = {Yutian Chen and Matthew W. Hoffman and Sergio Gomez Colmenarejo and Misha Denil and Timothy P. Lillicrap and Matt Botvinick and Nando de Freitas},
journal= {arXiv preprint arXiv:1611.03824},
year = {2017}
}
Comments
Accepted by ICML 2017. Previous version "Learning to Learn for Global Optimization of Black Box Functions" was published in the Deep Reinforcement Learning Workshop, NIPS 2016