Linked Data and AI: Retrieval Augmented Generation (RAG)

This course provides a systematic introduction to the technical principles and application scenarios of RAG (Retrieval-Augmented Generation). Through hands-on demonstrations and guided practical sessions, participants will learn how to build RAG systems that integrate Large Language Models (LLMs) with local knowledge bases to empower personal research or library services. We will begin with constructing document-based RAG systems, followed by an exploration of RAG systems built upon Relational Databases, linked data, and Knowledge Graphs. The technical depth and complexity of the implementation will be adjusted in real-time based on the participants’ learning progress.

Note

Although no prior programming experience is required, a basic understanding of computer programming and databases will be helpful. This course is intended for those who have social sciences and humanities background and are interested in learning linked data and AI and gaining hands-on experience.

Recommended Preparation

It is recommended to install the following components before the course begins:

Local LLM Deployment

• Install Ollama
• Download at least one model (e.g., DeepSeek-R1:7B)
• Install AnythingLLM

AI Programming Environment & Tools

• VS Code with GitHub Copilot or Lingma
• Python 3.9 or higher
• You may also register an account in Kaggle or similar platforms (such as Colab and Jupyter) to run Python codes

Knowledge Graph Storage & Retrieval

• Install Neo4j Desktop (v5.20.x)