← Dictionary
Dictionary · AI

What is Embedding?

It is the conversion of words or data into numerical strings so that they can be understood by artificial intelligence.

Overview

Embedding is a digital mapping method that allows computers to understand the relationship between words. Thanks to this process, words are converted into coordinates in a multidimensional space according to their semantic proximity.

Analogy: Imagine stacking the books in a library side by side not just by their names, but by the similarity of the topics they cover. In this way, the computer understands that the words 'king' and 'queen' are close to each other and that they are in a different world than 'apple' and 'car'.

How it works

By reading huge texts, the artificial intelligence model learns in what contexts words are used. Then, it assigns to each word a long series of numbers, that is, a vector, that represents its world of meaning. These numbers determine the semantic distance of the word from all other words.

Where it is used

It is used in search engines to bring results that are close in meaning even if they do not exactly match the word the user is looking for. It is also the fundamental building block in any application where artificial intelligence models need to make sense of text.

Commonly confused with

It is often thought of as just a translation method, but this is not a translation but a mathematical placement of meaning in space.

Frequently asked questions

Why don't we just use words?

Computers don't know what words mean, they can only process numbers.

Is embedding permanent?

Yes, numerical vectors created once are stored in the database and reused whenever needed.

Related terms

Related tools

This explanation was written in plain language for TreScout and machine-translated from the Turkish original · the Turkish version prevails. If something looks wrong or missing, write to hello@trescout.com. Read in Turkish →