GraphRAG – a joint research initiative

Our LAB team is always delighted to collaborate with clients’ AI Digital and Innovation team on research projects and publications.

🙌 This research publication was the result of a joint initiative with BNP Paribas, Neo4j and Lingua Custodia

👩‍🔬 The research project titled ‘GraphRAG: Leveraging Graph-Based Efficiency to Minimize Hallucinations in LLM-Driven RAG for Finance Data’ has been accepted for publication for the GenAIK 2025 workshop at Abu Dhabi in January 2025

ℹ️ The research explores different ways to introduce structured knowledge from Knowledge Graphs (KG) in a RAG pipeline.

💁 While it is not straightforward to convert graph-based knowledge to text leverageable by an LLM, we show that, if done correctly, it can help to reduce hallucinations and the cost of inference by drastically reducing tokens consumption. 💪

Well done to all the authors! This is a fantastic achievement 🤩