« Back to Glossary Index

In machine learning, latent space refers to a compressed, abstract representation of data created by a model, such as a neural network. In this space, similar data points are positioned closer together, capturing the underlying patterns or features.