Tim Toohill - AI-Generated Graph DB Queries - Developing an algorithm that uses an LLM to spt NLQs
AI-Generated Graph Database Queries - Developing an algorithm that uses an LLM to support Natural Language Queries (NLQs) by Tim Toohill
Visit https://d4con.io to learn more.
Abstract: We start by reviewing different types of databases, such as SQL, NoSQL, and graph databases, and we discuss the advantages of each for NLQ. We then identify a dataset that we can use to answer questions. We build a relatively simple ETL pipeline to load the dataset into a property graph. Using Python, we walk through the implementation of an algorithm to align a Large Language Model (LLM) to the property graph schema and process its output. Finally, we will see how I implemented NLQ in a full stack property graph application.
Twitter: / groupcertus
Presented at the 2023 D4 Conference (August 24, 2023)