What is a Vector Database? Explained in Everyday Language for AI Beginners
In the last two articles of this series, I introduced RAG and Embedding. This time, let’s discuss another highly popular term in the world of AI: Vector Database.
Vector Database Before AI
First thing first, if you’ve been hearing this buzzword more frequently lately, you might already know that vector databases aren’t new. Before their widespread use in Large Language Models (LLM), vector databases have been used in areas such as:
- Recommendation Systems
- Search Engines
- Medical Imaging
- Fraud Detection
- Speech Recognition
With that being said, it means that to understand “what is a vector database,” we need to know what all these examples have in common:
It is “Similarity”.
All of the examples mentioned above are systems driven by the need to find similar items based on input. For instance, they might associate “restaurants” with “what to eat” in a Google Search or suggest similar cat videos to what you’ve been watching on TikTok.
Traditional databases, however, are typically not structured to store data in this way. This is where vector databases come…