ฐานข้อมูลเวกเตอร์คืออะไร

Meet the vector database — the technology behind much of today's AI. You'll learn what it is, why it exists, and how it helps AI find meaning in data.

SQL databases shine on structured data like IDs and prices. But how do you search a 'happy dog' image or find an article semantically similar to another?

At the heart of it all is the vector — simply a list of numbers representing the features of some data, like a point in multi-dimensional space.

AI models turn complex data into numeric vectors, a process called embedding. The magic: similar things end up as vectors that sit close together in that space.

Here's what a tiny vector looks like in code — imagine each number standing for a feature like furry, four-legged, or purrs.

This is the core idea: vectors that are numerically close represent data that's semantically similar. 'Apple' lands nearer 'banana' than 'car'.

Ordinary databases can't find close vectors fast across hundreds of dimensions. Vector databases are purpose-built for lightning-fast nearest-neighbor search.

The main job is similarity search: hand it a query vector and it instantly returns the closest matches — like finding a song by how it feels, not its title.

Vector databases power recommendations: watching a movie, the system finds others with similar meaning vectors you'll likely enjoy. 'Users who bought this also bought...'

They also enable semantic search: ask 'healthy breakfast options?' and get results by meaning, not exact keywords — perfect for chatbots and knowledge bases.

Vector databases are specialized tools. What key advantages do they offer over traditional databases for AI-driven applications?

Recap: vector databases store numeric representations of meaning, enable fast similarity search by finding close vectors, and power recommendations and semantic search. Next: embeddings.