BOSH: Bayesian Optimization by Sampling Hierarchically
This addresses a key limitation in BO for stochastic settings, offering improved performance for applications like simulation optimization and reinforcement learning, though it appears incremental as it builds on existing BO frameworks.
The paper tackled the problem of Bayesian Optimization (BO) for functions with stochastic evaluations, such as in hyper-parameter tuning, by proposing BOSH, a method that uses a hierarchical Gaussian process and information theory to dynamically generate realizations, resulting in more efficient and higher-precision optimization across various tasks.
Deployments of Bayesian Optimization (BO) for functions with stochastic evaluations, such as parameter tuning via cross validation and simulation optimization, typically optimize an average of a fixed set of noisy realizations of the objective function. However, disregarding the true objective function in this manner finds a high-precision optimum of the wrong function. To solve this problem, we propose Bayesian Optimization by Sampling Hierarchically (BOSH), a novel BO routine pairing a hierarchical Gaussian process with an information-theoretic framework to generate a growing pool of realizations as the optimization progresses. We demonstrate that BOSH provides more efficient and higher-precision optimization than standard BO across synthetic benchmarks, simulation optimization, reinforcement learning and hyper-parameter tuning tasks.