Batch Bayesian Optimization via Local Penalization

TL;DR

This paper introduces a simple, efficient batch Bayesian optimization method using local penalization based on Lipschitz continuity, enabling parallel exploration with minimal computational overhead.

Contribution

It proposes a novel heuristic for batch selection in Bayesian optimization that captures local repulsion efficiently, improving speed over existing methods.

Findings

Compared favorably in runtime with more complex methods

Achieved significant speed-up in computational experiments

Effective in optimizing expensive functions with parallel evaluations

Abstract

The popularity of Bayesian optimization methods for efficient exploration of parameter spaces has lead to a series of papers applying Gaussian processes as surrogates in the optimization of functions. However, most proposed approaches only allow the exploration of the parameter space to occur sequentially. Often, it is desirable to simultaneously propose batches of parameter values to explore. This is particularly the case when large parallel processing facilities are available. These facilities could be computational or physical facets of the process being optimized. E.g. in biological experiments many experimental set ups allow several samples to be simultaneously processed. Batch methods, however, require modeling of the interaction between the evaluations in the batch, which can be expensive in complex scenarios. We investigate a simple heuristic based on an estimate of the Lipschitz constant that captures the most important aspect of this interaction (i.e. local repulsion) at negligible computational overhead. The resulting algorithm compares well, in running time, with much more elaborate alternatives. The approach assumes that the function of interest, $f$, is a Lipschitz continuous function. A wrap-loop around the acquisition function is used to collect batches of points of certain size minimizing the non-parallelizable computational effort. The speed-up of our method with respect to previous approaches is significant in a set of computationally expensive experiments.