Start: 01.01.2026
End: 01.07.2026
Type: Master Thesis
Student: Subitha Murugesan
Supervisor: Prof. Dr. Andreas Harth, Rene Dorsch
Abstract:
In recent years, Retrieval-Augmented Generation (RAG) has emerged as a powerful method to enhance the performance of large language models (LLMs) by grounding their outputs in external knowledge sources. Traditional RAG systems typically rely on unstructured document retrieval, which can limit their ability to reason over complex relationships or perform multi-hop queries. Graph RAG is a novel extension of this paradigm that incorporates structured knowledge graphs into the retrieval process, enabling more nuanced and context-aware generation.
This thesis aims to explore how Graph RAG approaches can be effectively integrated with existing structured databases such as relational databases or domain-specific systems like FHIR or Siemens Healthineers databases (e.g., cPDM) and to evaluate their performance, feasibility, and practical implications in real-world enterprise environments.