Skip to main content
SearchLoginLogin or Signup

EVD: provenancing claims within a dgraph

Part of the "Modes of interaction between discourse graphs and AI" dgraph

Published onJun 28, 2023
EVD: provenancing claims within a dgraph

The basic structure of a discourse graph.

For attribution purposes, node metadata may be sufficient.

EVD: A claim that is not linked to evidence or a source can be identified via visual inspection of the relevant graph region.

Status: In progress

Access: R/W via this PubPub

Edge: Supports CLM: Discourse Graphs can provide a uniquely interpretable form of LLM grounding because they connect claims to their provenance in context-rich natural language

Relevant Question: How can Discourse Graphs profitably interact with AI models like LLMs?

No comments here
Why not start the discussion?