Label scarcity in biomedicine: Data-rich latent factor discovery enhances phenotype prediction
This work addresses data scarcity issues in disease prediction for medical applications, though it appears incremental as it builds on existing semi-supervision and VAE methods.
The authors tackled the problem of label scarcity in biomedicine by using data-rich latent factor discovery from unlabeled normal subjects to enhance phenotype prediction, achieving better scaling with increasing unlabeled data compared to PCA or Isomap.
High-quality data accumulation is now becoming ubiquitous in the health domain. There is increasing opportunity to exploit rich data from normal subjects to improve supervised estimators in specific diseases with notorious data scarcity. We demonstrate that low-dimensional embedding spaces can be derived from the UK Biobank population dataset and used to enhance data-scarce prediction of health indicators, lifestyle and demographic characteristics. Phenotype predictions facilitated by Variational Autoencoder manifolds typically scaled better with increasing unlabeled data than dimensionality reduction by PCA or Isomap. Performances gains from semisupervison approaches will probably become an important ingredient for various medical data science applications.