Intermediate · Research
Embedding space
Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.
TL;DR. A multi-dimensional space where learned representations (embeddings) of data are mapped to, capturing meaningful relationships.
Technical Definition
A multi-dimensional space where learned representations (embeddings) of data are mapped to, capturing meaningful relationships.
How it works
An embedding space is a vector space where data points, such as words or items, are represented as dense vectors (embeddings). The spatial arrangement within this space is designed to reflect the relationships and similarities between these data points relevant to the model's task. Distances and directions in the embedding space can reveal semantic connections.
Related Concepts
- Word2Vec — A pioneering technique that learns dense word embeddings by predicting surrounding words from large text corpora.
- Dimensionality reduction — The process of reducing the number of features or variables in a dataset while retaining essential information.
- Embedding vector — A dense array of numbers learned by a model that represents an input item in a lower-dimensional space.