This repository contains code for evaluating the semantic similarity between a sentence given in source (natural) language and another sentence given in target language using FrameNet open source technology.
0
stars
0
forks
source link
How do you create embeddings for FrameNet entities? #1
Apologies for the delayed response. Yes, entity name has been encoded along with it's lexical unit and other features as well. No, we have not explored knowledge graph embeddings in this project.
How do you create/learn embeddings for frames, frame elements, and LUs?
It looks like: embedding for an entity name is being considered as an embedding of the entity. Is it correct?
Have you tried any knowledge graph embedding algorithm for the same?