BOHB: Robust and Efficient Hyperparameter Optimization at Scale

TL;DR

BOHB combines Bayesian optimization and bandit-based methods to deliver robust, efficient, and scalable hyperparameter tuning, outperforming existing approaches across diverse machine learning tasks.

Contribution

The paper introduces BOHB, a novel hyperparameter optimization method that merges Bayesian and bandit strategies for improved performance and scalability.

Findings

BOHB outperforms Bayesian optimization and Hyperband on various tasks.

BOHB is robust and versatile across different models and problem types.

BOHB is simple to implement and computationally efficient.

Abstract

Modern deep learning methods are very sensitive to many hyperparameters, and, due to the long training times of state-of-the-art models, vanilla Bayesian hyperparameter optimization is typically computationally infeasible. On the other hand, bandit-based configuration evaluation approaches based on random search lack guidance and do not converge to the best configurations as quickly. Here, we propose to combine the benefits of both Bayesian optimization and bandit-based methods, in order to achieve the best of both worlds: strong anytime performance and fast convergence to optimal configurations. We propose a new practical state-of-the-art hyperparameter optimization method, which consistently outperforms both Bayesian optimization and Hyperband on a wide range of problem types, including high-dimensional toy functions, support vector machines, feed-forward neural networks, Bayesian neural networks, deep reinforcement learning, and convolutional neural networks. Our method is robust and versatile, while at the same time being conceptually simple and easy to implement.