Objective Bound Conditional Gaussian Process for Bayesian Optimization

Taewon Jeong, Heeyoung Kim
Proceedings of the 38th International Conference on Machine Learning, PMLR 139:4819-4828, 2021.

Abstract

A Gaussian process is a standard surrogate model for an unknown objective function in Bayesian optimization. In this paper, we propose a new surrogate model, called the objective bound conditional Gaussian process (OBCGP), to condition a Gaussian process on a bound on the optimal function value. The bound is obtained and updated as the best observed value during the sequential optimization procedure. Unlike the standard Gaussian process, the OBCGP explicitly incorporates the existence of a point that improves the best known bound. We treat the location of such a point as a model parameter and estimate it jointly with other parameters by maximizing the likelihood using variational inference. Within the standard Bayesian optimization framework, the OBCGP can be combined with various acquisition functions to select the next query point. In particular, we derive cumulative regret bounds for the OBCGP combined with the upper confidence bound acquisition algorithm. Furthermore, the OBCGP can inherently incorporate a new type of prior knowledge, i.e., the bounds on the optimum, if it is available. The incorporation of this type of prior knowledge into a surrogate model has not been studied previously. We demonstrate the effectiveness of the OBCGP through its application to Bayesian optimization tasks, such as the sequential design of experiments and hyperparameter optimization in neural networks.

Cite this Paper


BibTeX
@InProceedings{pmlr-v139-jeong21a, title = {Objective Bound Conditional Gaussian Process for Bayesian Optimization}, author = {Jeong, Taewon and Kim, Heeyoung}, booktitle = {Proceedings of the 38th International Conference on Machine Learning}, pages = {4819--4828}, year = {2021}, editor = {Meila, Marina and Zhang, Tong}, volume = {139}, series = {Proceedings of Machine Learning Research}, month = {18--24 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v139/jeong21a/jeong21a.pdf}, url = {https://proceedings.mlr.press/v139/jeong21a.html}, abstract = {A Gaussian process is a standard surrogate model for an unknown objective function in Bayesian optimization. In this paper, we propose a new surrogate model, called the objective bound conditional Gaussian process (OBCGP), to condition a Gaussian process on a bound on the optimal function value. The bound is obtained and updated as the best observed value during the sequential optimization procedure. Unlike the standard Gaussian process, the OBCGP explicitly incorporates the existence of a point that improves the best known bound. We treat the location of such a point as a model parameter and estimate it jointly with other parameters by maximizing the likelihood using variational inference. Within the standard Bayesian optimization framework, the OBCGP can be combined with various acquisition functions to select the next query point. In particular, we derive cumulative regret bounds for the OBCGP combined with the upper confidence bound acquisition algorithm. Furthermore, the OBCGP can inherently incorporate a new type of prior knowledge, i.e., the bounds on the optimum, if it is available. The incorporation of this type of prior knowledge into a surrogate model has not been studied previously. We demonstrate the effectiveness of the OBCGP through its application to Bayesian optimization tasks, such as the sequential design of experiments and hyperparameter optimization in neural networks.} }
Endnote
%0 Conference Paper %T Objective Bound Conditional Gaussian Process for Bayesian Optimization %A Taewon Jeong %A Heeyoung Kim %B Proceedings of the 38th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2021 %E Marina Meila %E Tong Zhang %F pmlr-v139-jeong21a %I PMLR %P 4819--4828 %U https://proceedings.mlr.press/v139/jeong21a.html %V 139 %X A Gaussian process is a standard surrogate model for an unknown objective function in Bayesian optimization. In this paper, we propose a new surrogate model, called the objective bound conditional Gaussian process (OBCGP), to condition a Gaussian process on a bound on the optimal function value. The bound is obtained and updated as the best observed value during the sequential optimization procedure. Unlike the standard Gaussian process, the OBCGP explicitly incorporates the existence of a point that improves the best known bound. We treat the location of such a point as a model parameter and estimate it jointly with other parameters by maximizing the likelihood using variational inference. Within the standard Bayesian optimization framework, the OBCGP can be combined with various acquisition functions to select the next query point. In particular, we derive cumulative regret bounds for the OBCGP combined with the upper confidence bound acquisition algorithm. Furthermore, the OBCGP can inherently incorporate a new type of prior knowledge, i.e., the bounds on the optimum, if it is available. The incorporation of this type of prior knowledge into a surrogate model has not been studied previously. We demonstrate the effectiveness of the OBCGP through its application to Bayesian optimization tasks, such as the sequential design of experiments and hyperparameter optimization in neural networks.
APA
Jeong, T. & Kim, H.. (2021). Objective Bound Conditional Gaussian Process for Bayesian Optimization. Proceedings of the 38th International Conference on Machine Learning, in Proceedings of Machine Learning Research 139:4819-4828 Available from https://proceedings.mlr.press/v139/jeong21a.html.

Related Material