Part 5 Information retrieval with knowledge graphs and LLMs
The integration of KGs with LLMs reaches its practical culmination in this final part of the book, where we explore how to use these combined technologies for accurate and reliable information retrieval. We’ll focus on the practical implementation of systems that use KGs as ground truth to enhance LLM capabilities while preventing hallucinations.
Chapter 13 explores the integration of KGs with LLMs through retrieval augmented generation (RAG), showing how graph RAG systems can use structured data and language understanding to provide more accurate and transparent responses.
Chapter 14 illustrates how to build sophisticated question-answering systems that emulate domain expert reasoning to create contextually aware solutions. We demonstrate systematic approaches to intent detection, schema translation, and expert knowledge embedding.
Chapter 15 brings together all these concepts in a complete, working implementation using LangGraph and Streamlit. We demonstrate how to build a modular pipeline in which specialized agents handle different aspects of question processing while maintaining system observability and extensibility.
These chapters provide you with the practical knowledge to implement production-ready systems that combine the strengths of KGs and LLMs, serving as a comprehensive guide for organizations looking to deploy reliable, knowledge-grounded AI solutions.