Hyperband: Bandit-Based Configuration Evaluation for Hyperparameter OptimizationDownload PDF

Published: 21 Jul 2022, Last Modified: 05 May 2023ICLR 2017 PosterReaders: Everyone
Abstract: Performance of machine learning algorithms depends critically on identifying a good set of hyperparameters. While recent approaches use Bayesian Optimization to adaptively select configurations, we focus on speeding up random search through adaptive resource allocation. We present Hyperband, a novel algorithm for hyperparameter optimization that is simple, flexible, and theoretically sound. Hyperband is a principled early-stoppping method that adaptively allocates a predefined resource, e.g., iterations, data samples or number of features, to randomly sampled configurations. We compare Hyperband with state-of-the-art Bayesian Optimization methods on several hyperparameter optimization problems. We observe that Hyperband can provide over an order of magnitude speedups over competitors on a variety of neural network and kernel-based learning problems.
Conflicts: cs.ucla.edu, eecs.berkeley.edu, google.com
17 Replies