This repository contains data and source code for a prototypical Knowledge Extraction application for the manuscripts of Leopardi preserved at the Cambridge University Digital Library.
The automatically extracted Knowledge Graph is available in this Turtle File.
The Knowledge Graph was extracted by using Babelscape's REBEL jointly with ChatGPT, which was used to pre-process the Italian text in natural language triples. An illustration of our pipeline is presented below.
The source code and data here published are the result of research studies for the digitization project related to the manuscripts of Giacomo Leopardi and carried out by the National Center of Leopardian Studies and the University of Macerata.
For additional information, you can contact:
- Cristian Santini ([email protected])