Tuning Hyperparameters without Grad Students: Scalable and Robust Bayesian Optimisation with Dragonfly

TL;DR

Dragonfly is an open source Python library that advances Bayesian Optimisation by integrating methods for high-dimensional, multi-fidelity, structured, and parallel optimization, significantly improving performance in complex real-world tasks.

Contribution

The paper introduces Dragonfly, a scalable and robust Bayesian Optimization library that incorporates novel methods for challenging optimization scenarios and demonstrates superior performance over existing tools.

Findings

Dragonfly outperforms other optimization packages in complex tasks.

New methods improve handling of high-dimensional and multi-fidelity problems.

Enhanced optimization over structured and constrained domains.

Abstract

Bayesian Optimisation (BO) refers to a suite of techniques for global optimisation of expensive black box functions, which use introspective Bayesian models of the function to efficiently search for the optimum. While BO has been applied successfully in many applications, modern optimisation tasks usher in new challenges where conventional methods fail spectacularly. In this work, we present Dragonfly, an open source Python library for scalable and robust BO. Dragonfly incorporates multiple recently developed methods that allow BO to be applied in challenging real world settings; these include better methods for handling higher dimensional domains, methods for handling multi-fidelity evaluations when cheap approximations of an expensive function are available, methods for optimising over structured combinatorial spaces, such as the space of neural network architectures, and methods for handling parallel evaluations. Additionally, we develop new methodological improvements in BO for selecting the Bayesian model, selecting the acquisition function, and optimising over complex domains with different variable types and additional constraints. We compare Dragonfly to a suite of other packages and algorithms for global optimisation and demonstrate that when the above methods are integrated, they enable significant improvements in the performance of BO. The Dragonfly library is available at dragonfly.github.io.