Teaching Assistant of Artificial Intelligence course at The University of Alabama at Birmingham

This is a description of my teaching experience at The University of Alabama at Birmingham for CS660/CS760, the Artificial Intelligence course.

You can view the course syllabus here

In this course, I prepare the materials including documents, slides, and demo for GraphRAG lecture. The lecture delves into the innovative approaches of Graph Retrieval-Augmented Generation (GraphRAG), focusing on the synergy between graph-based data structures and generative models. The lecture primarily draws from the comprehensive survey by Peng et al. [1], as well as practical insights from the Neo4j Developer Blog on integrating Neo4j with LangChain [2].

Lecture Structure

1. Motivation
   • An effort to reach human-level reasoning ability
   • Limitations of RAG in real-world scenarios
2. Preliminaries
   • Text-Attributed Graphs
   • Graph Neural Networks
3. Overview of GraphRAG
   • Problem definition and conceptual framework
4. Graph-Guided Indexing
   • Graph data
   • Indexing
5. Graph-Guided Retrieval
   • Retriever
   • Retrieval Paradigm
   • Retrieval Granularity
   • Retrieval Enhancement
6. Graph-Enhanced Generation
   • Generation Techniques
   • Graph Formats
   • Generation Enhancement

7. Applications : Examples of industries and domains where GraphRAG is being or can be applied, such as healthcare, finance, and legal tech.

8. Related Works
   • GraphRAG by NebulaGraph
   • GraphRAG by Microsoft
   • GRAG
9. Demo: Practical insights on implementing GraphRAG with Neo4j and LangChain, offering a hands-on perspective

References

[1] : Peng et al., “Graph Retrieval-Augmented Generation: A Survey” - A comprehensive survey of the GraphRAG model, its applications, and theoretical underpinnings.

[2]: Neo4j Developer Blog - Global GraphRAG: Neo4j + LangChain - Practical insights on implementing GraphRAG with Neo4j and LangChain, offering a hands-on perspective.