Sample-Efficient Bayesian Optimization with Transfer Learning for Heterogeneous Search Spaces

TL;DR

This paper introduces two novel transfer learning methods for Bayesian optimization that effectively handle heterogeneous search spaces, enabling more efficient optimization with limited function evaluations.

Contribution

It proposes Gaussian process-based approaches with conditional kernels and hyperparameter inference to transfer knowledge across different search spaces in BO.

Findings

Both methods outperform baseline approaches on benchmark problems.

The approaches effectively transfer information between heterogeneous search spaces.

Methods improve sample efficiency in black-box optimization tasks.

Abstract

Bayesian optimization (BO) is a powerful approach to sample-efficient optimization of black-box functions. However, in settings with very few function evaluations, a successful application of BO may require transferring information from historical experiments. These related experiments may not have exactly the same tunable parameters (search spaces), motivating the need for BO with transfer learning for heterogeneous search spaces. In this paper, we propose two methods for this setting. The first approach leverages a Gaussian process (GP) model with a conditional kernel to transfer information between different search spaces. Our second approach treats the missing parameters as hyperparameters of the GP model that can be inferred jointly with the other GP hyperparameters or set to fixed values. We show that these two methods perform well on several benchmark problems.