AIGlossaryLab
⌕
← Back
DataIntermediate2013

Embedding

Embedding

An embedding turns meaning into a list of numbers, allowing an AI system to compare how similar things are.

Explained simply

An embedding turns meaning into a list of numbers, allowing an AI system to compare how similar things are.

1Text or image
→
2Convert to numbers
→
3Compare distance
→
4Find similarity

At a glance

Category
Data
Difficulty
Intermediate
Introduced
2013

Real example

A search tool finds “annual leave policy” when you search for “holiday entitlement.”

Why it matters

Embeddings power semantic search, recommendations and RAG systems.

Timeline

Origins

The ideas behind Embedding begin developing.

2013

Embedding becomes a recognised term or technique.

Wider use

Research and practical applications increase.

Modern AI

Embedding becomes connected to newer AI systems and products.

Today

Embedding remains relevant in the data area of AI.

Dates for broad concepts may describe the period when the idea emerged or became widely used, rather than one exact invention date.

Learn next

Continue with these connected terms:

Vector SearchVector search finds items whose embeddings are close together, which usually means their meaning is similar.→Vector DatabaseA vector database stores embeddings so a system can quickly find items with similar meaning.→RAGRAG lets an AI search trusted information before it writes an answer. RAG stands for Retrieval-Augmented Generation.→

Simple infographic

A quick visual way to understand Embedding.

1Text or image
→
2Convert to numbers
→
3Compare distance
→
4Find similarity
1 of 6

Was this helpful?

Your feedback helps improve this explanation.