Show simple item record

dc.contributor.authorGriffiths, Ryan-Rhys
dc.contributor.authorHernández-Lobato, José Miguel
dc.date.accessioned2020-04-20T01:08:29Z
dc.date.available2020-04-20T01:08:29Z
dc.date.issued2019-11-18
dc.identifier.issn2041-6520
dc.identifier.otherPMC7067240
dc.identifier.other32190274
dc.identifier.urihttps://www.repository.cam.ac.uk/handle/1810/304530
dc.description.abstractAutomatic Chemical Design is a framework for generating novel molecules with optimized properties. The original scheme, featuring Bayesian optimization over the latent space of a variational autoencoder, suffers from the pathology that it tends to produce invalid molecular structures. First, we demonstrate empirically that this pathology arises when the Bayesian optimization scheme queries latent space points far away from the data on which the variational autoencoder has been trained. Secondly, by reformulating the search procedure as a constrained Bayesian optimization problem, we show that the effects of this pathology can be mitigated, yielding marked improvements in the validity of the generated molecules. We posit that constrained Bayesian optimization is a good approach for solving this kind of training set mismatch in many generative tasks involving Bayesian optimization over the latent space of a variational autoencoder.
dc.languageeng
dc.rightsAttribution 4.0 International
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/
dc.sourceessn: 2041-6539
dc.sourcenlmid: 101545951
dc.titleConstrained Bayesian optimization for automatic chemical design using variational autoencoders.
dc.typeArticle
dc.date.updated2020-04-20T01:08:28Z
prism.endingPage586
prism.issueIdentifier2
prism.publicationNameChemical science
prism.startingPage577
prism.volume11
dc.identifier.doi10.17863/CAM.51611
rioxxterms.versionofrecord10.1039/c9sc04026a
rioxxterms.versionVoR
rioxxterms.licenseref.urihttps://creativecommons.org/licenses/by/4.0/
dc.contributor.orcidGriffiths, Ryan-Rhys [0000-0003-3117-4559]


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record

Attribution 4.0 International
Except where otherwise noted, this item's licence is described as Attribution 4.0 International