English

A General Recipe for Likelihood-free Bayesian Optimization

Machine Learning 2022-10-10 v2 Artificial Intelligence Machine Learning

Abstract

The acquisition function, a critical component in Bayesian optimization (BO), can often be written as the expectation of a utility function under a surrogate model. However, to ensure that acquisition functions are tractable to optimize, restrictions must be placed on the surrogate model and utility function. To extend BO to a broader class of models and utilities, we propose likelihood-free BO (LFBO), an approach based on likelihood-free inference. LFBO directly models the acquisition function without having to separately perform inference with a probabilistic surrogate model. We show that computing the acquisition function in LFBO can be reduced to optimizing a weighted classification problem, where the weights correspond to the utility being chosen. By choosing the utility function for expected improvement (EI), LFBO outperforms various state-of-the-art black-box optimization methods on several real-world optimization problems. LFBO can also effectively leverage composite structures of the objective function, which further improves its regret by several orders of magnitude.

Keywords

Cite

@article{arxiv.2206.13035,
  title  = {A General Recipe for Likelihood-free Bayesian Optimization},
  author = {Jiaming Song and Lantao Yu and Willie Neiswanger and Stefano Ermon},
  journal= {arXiv preprint arXiv:2206.13035},
  year   = {2022}
}

Comments

ICML 2022. This version fixes a typo in eq 33

R2 v1 2026-06-24T12:04:42.701Z