Intermediate · Research
Embedding vector
Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.
TL;DR. A dense array of numbers learned by a model that represents an input item in a lower-dimensional space.
Technical Definition
A dense array of numbers learned by a model that represents an input item in a lower-dimensional space.
How it works
An embedding vector is a fixed-size array of floating-point numbers that serves as a learned representation of an input. These vectors are typically generated by an embedding layer and are designed to capture the essential characteristics or meaning of the input in a way that is useful for a machine learning model. They are not random but are adjusted during training to optimize performance.
Related Concepts
- Embedding layer — A hidden neural network layer that learns lower-dimensional representations (embeddings) for high-dimensional categorical data.
- Representation Learning — The process of automatically discovering meaningful representations of data from raw inputs.