Knowledge Graph building from unstructured data is an attractive but also challenging task. One of the most complex aspects is relation extraction, i.e., the identification of semantic relationships between entities. This task can be accomplished either using Dependency Parsing or Large Language Models. Join Federica in this session to explore how to leverage dependency graphs, employing a rule-based approach and exploiting the Neo4j navigation experience. She will also apply GPT language models and discuss prompt engineering. Towards the end, she'll compare the two approaches on a knowledge graph concerning the Gupta Leaks scandal and highlight some interesting insights from graph data science.
Slides: https://drive.google.com/file/d/1zhTmueQJU_8v4tiWpEUu_SpLwgfNikn1/view?usp=sharing
#neo4j #graphdatabase #nodes2023 #dependency #llm