# Pretrained graph embeddings

Is there any precedent for DL use of pretrained graph embeddings in a similar manner to word embeddings?

Am you actually looking for graph neutral network?