I am looking into utilizing spacy's custom dependency parser (https://spacy.io/usage/training#tagger-parser - "Training a parser for custom semantics").
It seems as if you would have to manually label the heads and dependencies. Does anyone know of a more efficient way to do this?
I plan on using Snorkel to create labeling functions for other things, but given that you have to essentially label each word is seems as if that would be tough to apply here.