Dataset Open Access
Up-to-date information and pre-trained embeddings can be found here.
In order to support NLP research efforts related to COVID-19, we have developed resources for training vector-valued embeddings of COVID-19 related medical concepts, primarily using the CORD-19 dataset.
This resource includes only concept embeddings. Download of the full sets of concept, term, and word embeddings from here requires a valid UMLS Terminology Services account, in order to validate licensed access to SNOMED-CT.