Variational autoencoders learn transferrable representations of metabolomics data

Abstract

Variable autoencoders offer an alternative way to interrogate metabolomic data and identify meaningful, non-linear relationships.