Build knowledge graphs with LLM-driven entity extraction

Embeddings databases are a union of vector indexes (sparse and dense), graph networks and relational databases. This enables vector search with SQL, topic modeling, retrieval augmented generation and more. Up until txtai 7.0, semantic graphs only sup...

neuml.hashnode.dev3 min read

#ai #llm #rag #vector-database

Responses(1)

MB

Massimo Bellino

This is a very insightful article. Two questions: 1) Can you illustrate a workflow whereby, after the LLM-driven entity extractions, the extracted topics are uploaded and linked with the existing topics present within a third-party RDF graph (in my case, GraphDB)? 2) Can the LLM-driven entity extraction classify the entity extracted (e.g. geography, industry, etc.?) so that only the entities extracted belonging to a specific entity type are selected for the next step?

Data driven analytics

Mar 11, 2024

BH

Bjorn Heijligers

Cosmologist Gone Rogue

Mar 17, 2024

Good question. Is the used model finetuned for "entity extraction"? or is that word in the prompt doing all the heavy lifting? Domain specific entity/process extraction would be a powerful thing to master.

Search Hashnode

Build knowledge graphs with LLM-driven entity extraction

Responses(1)