Maximally Robust Satisficing Bayesian Optimization

Samuli Kinnunen, Petrus Mikkola, Antti Niskanen, Arto Klami
Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, PMLR 337:3117-3142, 2026.

Abstract

Many design tasks can be cast as black-box function optimization, enabling use of {Bayesian} optimization to find an ideal design with minimal number of trials. However, often we do not actually need the optimum but instead a sufficiently good solution is enough, for instance a material that is durable enough for its intended use.In most cases there are multiple satisfactory solutions, forming a superlevel set of the function, raising a key question of which one to prefer. We answer this by explaining why robustness to input perturbations that may occur when the solution is deployed is a good criterion and by introduce a {Bayesian} optimization method that efficiently finds satisficing solutions that are robust to maximally large perturbations. In contrast to previous works, we assume the inputs can be accurately controlled during optimization, but will be perturbed after the deployment.

Cite this Paper

BibTeX
@InProceedings{pmlr-v337-kinnunen26a, title = {Maximally Robust Satisficing {Bayesian} Optimization}, author = {Kinnunen, Samuli and Mikkola, Petrus and Niskanen, Antti and Klami, Arto}, booktitle = {Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence}, pages = {3117--3142}, year = {2026}, editor = {Perković, Emilija and Malinsky, Daniel}, volume = {337}, series = {Proceedings of Machine Learning Research}, month = {17--21 Aug}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v337/main/assets/kinnunen26a/kinnunen26a.pdf}, url = {https://proceedings.mlr.press/v337/kinnunen26a.html}, abstract = {Many design tasks can be cast as black-box function optimization, enabling use of {Bayesian} optimization to find an ideal design with minimal number of trials. However, often we do not actually need the optimum but instead a sufficiently good solution is enough, for instance a material that is durable enough for its intended use.In most cases there are multiple satisfactory solutions, forming a superlevel set of the function, raising a key question of which one to prefer. We answer this by explaining why robustness to input perturbations that may occur when the solution is deployed is a good criterion and by introduce a {Bayesian} optimization method that efficiently finds satisficing solutions that are robust to maximally large perturbations. In contrast to previous works, we assume the inputs can be accurately controlled during optimization, but will be perturbed after the deployment.} }
Endnote
%0 Conference Paper %T Maximally Robust Satisficing Bayesian Optimization %A Samuli Kinnunen %A Petrus Mikkola %A Antti Niskanen %A Arto Klami %B Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2026 %E Emilija Perković %E Daniel Malinsky %F pmlr-v337-kinnunen26a %I PMLR %P 3117--3142 %U https://proceedings.mlr.press/v337/kinnunen26a.html %V 337 %X Many design tasks can be cast as black-box function optimization, enabling use of {Bayesian} optimization to find an ideal design with minimal number of trials. However, often we do not actually need the optimum but instead a sufficiently good solution is enough, for instance a material that is durable enough for its intended use.In most cases there are multiple satisfactory solutions, forming a superlevel set of the function, raising a key question of which one to prefer. We answer this by explaining why robustness to input perturbations that may occur when the solution is deployed is a good criterion and by introduce a {Bayesian} optimization method that efficiently finds satisficing solutions that are robust to maximally large perturbations. In contrast to previous works, we assume the inputs can be accurately controlled during optimization, but will be perturbed after the deployment.
APA
Kinnunen, S., Mikkola, P., Niskanen, A. & Klami, A.. (2026). Maximally Robust Satisficing Bayesian Optimization. Proceedings of the 42nd Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research 337:3117-3142 Available from https://proceedings.mlr.press/v337/kinnunen26a.html.

Related Material