GraphRAG Essential Training

Course details

This beginner-friendly course introduces the fundamentals of GraphRAG (Graph Retrieval-Augmented Generation), a cutting-edge technique that combines knowledge graphs with generative AI to enhance contextual relevance and precision. Designed for professionals and students new to GraphRAG, the course covers key concepts, including graph structures, nodes, edges, and relationships, as well as practical skills in building and configuring GraphRAG models. Explore how to integrate GraphRAG into existing workflows to create enriched, data-driven AI applications, through hands-on exercises and real-world examples. By the end of the course, you will be able to identify use cases and implement GraphRAG effectively in a generative AI pipeline.

Learning objectives
Understand and apply essential graph principles and structures, such as nodes, edges, and relationships, to enhance retrieval in AI applications.
Identify practical use cases for GraphRAG such as the minimization of hallucinations and determine how to leverage it effectively in projects.
Write basic GraphRAG software to combine graph-based knowledge with generative AI utilizing basic Python and Neo4j.
Integrate GraphRAG with other NLP tools to create dynamic and contextually enriched AI responses utilizing standard large language models (LLMs) such as ChatGPT, Mistral, etc.

Concepts

0. Introduction

01 - Prevent GenAI apps from hallucinating
02 - An example of knowledge graphs at work in AI
03 - Codespaces explainer video

1. Graphs Made Simple - Understanding the Basics

04 - What is a graph
05 - Recognizing graphs in everyday life
06 - Why are graphs useful
07 - Nodes, relationships, and properties
08 - Graphs vs. traditional data structures
09 - Knowledge graphs - Connecting information
10 - How graphs power AI

2. Getting Started with Graph Tools

11 - What is a graph database
12 - Introduction to Neo4j
13 - Setting up Neo4j
14 - Exploring the Neo4j browser
15 - Cypher basics - Writing your first queries
16 - More cypher - Retrieving nodes and relationships
17 - Connecting Python to Neo4j
18 - Testing your setup with Cypher queries

3. Building Your First Knowledge Graph for Graph Retrieval-Augmented Generation (GraphRAG)

19 - Introduction to Retrieval-Augmented Generation (RAG)
20 - How RAG works with vector embeddings
21 - Improving your RAG with graphs - GraphRAG
22 - Overview of LangChain
23 - Key concepts in LangChain for graph workflows
24 - Populating a knowledge graph into Neo4j using LangChain
25 - Challenge - Query your knowledge graph with Cypher
26 - Solution - Query your knowledge graph with Cypher

4. Connecting Knowledge Graphs to Generative AI

27 - Creating a GraphRAG pipeline with LangChain to query your data
28 - Enhancing your knowledge graph with richer data
29 - Using knowledge graphs in a GraphRAG pipeline
30 - Comparing the GraphRAG results to a traditional vector-based RAG
31 - Evaluating your GraphRAG pipeline
32 - Challenge - Evaluate your GraphRAG application
33 - Solution - Evaluate your GraphRAG application

5. Putting It All Together - Creating a Question-Answering Bot with GraphRAG

34 - Introduction to capstone project
35 - Walkthrough of capstone solution

Conclusion

36 - Continuing on with knowledge graphs, GraphRAG, and GenAI