Graph Data Modeling with AI

Welcome to our website for Graph Data Modeling with AI.

This course is designed as a 10 to 14-week course for undergraduate computer science students that are interested in graph data models and how they complement the limitations of large-language models (LLMs).

I first started using LLMs to generate educational content in September of 2020. At the time, the 1.7 billion parameter GPT-3 was the the largest, most powerful LLM. However, LLMs improved both in size (number of parameters) and quality.
Now, given the right context, LLMs can generate large sections of textbooks.

Yet LLMs only make predictions of the next word or are used to convert text into images, video or speech. They have no models of the real world. They only model language. Because of this, current LLMs will forever be flawed and prone to hallucination.

That is where this book comes in. Once you have a deep understanding of graph data modeling you will understand that graph databases can complement LLMs. Where LLMs are weak, graph databases are strong. Querying graphs gives precise, reproducible and highly explainable answers. By combing LLMs with graphs you can create robust AI systems with robust features and far lower costs.

We hope you enjoy learning graph modeling with this book and we look forward to your feedback.

Please connect with me on LinkedIn if you have any questions.

• Dan McCreary