Skip to content

RAG & Search

Bi-encoder

An AI setup that prepares two pieces of information separately before comparing them.

Example

A search service prepares the question and each document separately before comparing them.

Why people use it

It speeds up large searches by doing some document preparation before a question arrives.

What you'll hear

“Can we prepare the documents before people search?”

What this means for you

Check whether the faster comparisons still find the details your task needs.

Can you control it?

Developer-only

The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.

Common questions

What happens when a stored document changes?
Its prepared description may need updating so comparisons reflect the new version.
Can documents be prepared before the question exists?
Yes. Stored numerical descriptions can be compared with a question that arrives later.
What does this speed trade away?
Separately processing the question and document can miss details that become clearer when they are read together.

Related terms

Still have questions?

Up to 500 characters.

Ask LATHIC about AI. Relevant glossary entries may be included.

Your question, the glossary entries it matches, and a rotating pseudonymous identifier go to Microsoft Azure’s OpenAI service through Vercel AI Gateway to generate an answer. Zero retention and no training are required of the provider, and LATHIC does not save your question or answer. Privacy Notice