Repository logo
 

Constrained Bayesian optimization for automatic chemical design using variational autoencoders.

Published version
Peer-reviewed

Type

Article

Change log

Authors

Griffiths, Ryan-Rhys  ORCID logo  https://orcid.org/0000-0003-3117-4559
Hernández-Lobato, José Miguel 

Abstract

Automatic 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.

Description

Keywords

34 Chemical Sciences

Journal Title

Chem Sci

Conference Name

Journal ISSN

2041-6520
2041-6539

Volume Title

11

Publisher

Royal Society of Chemistry (RSC)