Batch Bayesian Optimization via Local Penalization
[edit]
Related Material
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 could either be computational or physical facets of the process being optimized. Batch methods, however, require the modeling of the interaction between the different evaluations in the batch, which can be expensive in complex scenarios. We investigate this issue and propose a highly effective heuristic based on an estimate of the function’s Lipschitz constant that captures the most important aspect of this interaction–local repulsion–at negligible computational overhead. A penalized acquisition function is used to collect batches of points minimizing the nonparallelizable computational effort. The resulting algorithm compares very well, in runtime, with much more elaborate alternatives.


Gonzalez, J., Dai, Z., Hennig, P. & Lawrence, N.D.. (). Batch Bayesian Optimization via Local Penalization. , in PMLR :648657
