Show simple item record

dc.contributor.authorGriffiths, Ryan-Rhys
dc.contributor.authorA Aldrick, Alexander
dc.contributor.authorGarcia-Ortegon, Miguel
dc.contributor.authorLalchand, Vidhi
dc.contributor.authorLee, Alpha A
dc.date.accessioned2021-11-23T18:12:46Z
dc.date.available2021-11-23T18:12:46Z
dc.date.issued2021-11-24
dc.date.submitted2021-01-08
dc.identifier.othermlstac298c
dc.identifier.otherac298c
dc.identifier.othermlst-100414.r1
dc.identifier.urihttps://www.repository.cam.ac.uk/handle/1810/330992
dc.description.abstractAbstract: Bayesian optimisation is a sample-efficient search methodology that holds great promise for accelerating drug and materials discovery programs. A frequently-overlooked modelling consideration in Bayesian optimisation strategies however, is the representation of heteroscedastic aleatoric uncertainty. In many practical applications it is desirable to identify inputs with low aleatoric noise, an example of which might be a material composition which displays robust properties in response to a noisy fabrication process. In this paper, we propose a heteroscedastic Bayesian optimisation scheme capable of representing and minimising aleatoric noise across the input space. Our scheme employs a heteroscedastic Gaussian process surrogate model in conjunction with two straightforward adaptations of existing acquisition functions. First, we extend the augmented expected improvement heuristic to the heteroscedastic setting and second, we introduce the aleatoric noise-penalised expected improvement (ANPEI) heuristic. Both methodologies are capable of penalising aleatoric noise in the suggestions. In particular, the ANPEI acquisition yields improved performance relative to homoscedastic Bayesian optimisation and random sampling on toy problems as well as on two real-world scientific datasets. Code is available at: https://github.com/Ryan-Rhys/Heteroscedastic-BO
dc.languageen
dc.publisherIOP Publishing
dc.subjectPaper
dc.subjectBayesian optimisation
dc.subjectGaussian processes
dc.subjectheteroscedasticity
dc.titleAchieving robustness to aleatoric uncertainty with heteroscedastic Bayesian optimisation
dc.typeArticle
dc.date.updated2021-11-23T18:12:46Z
prism.issueIdentifier1
prism.publicationNameMachine Learning: Science and Technology
prism.volume3
dc.identifier.doi10.17863/CAM.78437
dcterms.dateAccepted2021-09-23
rioxxterms.versionofrecord10.1088/2632-2153/ac298c
rioxxterms.versionVoR
rioxxterms.licenseref.urihttp://creativecommons.org/licenses/by/4.0
dc.contributor.orcidGriffiths, Ryan-Rhys [0000-0003-3117-4559]
dc.identifier.eissn2632-2153


Files in this item

Thumbnail
Thumbnail

This item appears in the following Collection(s)

Show simple item record