« Back to Glossary Index

In generative AI, latent space refers to a compressed, lower-dimensional representation of data that captures its essential features. Models like Variational Autoencoders (VAEs) and Generative Adversarial Networks (GANs) map complex, high-dimensional data—such as images or text—into this latent space. By manipulating points within this space, these models can generate new, realistic data samples that resemble the original dataset. This approach enables the creation of diverse and novel content, making latent space a fundamental concept in generative AI.

« Back to Glossary Index