latent

Definition

Describes variables or representations that are not directly observed but inferred inside the model. Compressing inputs into a lower-dimensional latent space makes generation, retrieval, and interpolation far cheaper, which is why latent diffusion and variational autoencoders are built around it. The word always points at the intermediate state between visible input and visible output.

Explainers using this term

• Paper Walkthrough — Beyond Retrieval: LatentStream Turns Retrieved Video Into Latent MemoryBeyond Retrieval: Progressive Latent Memory Evolution for Streaming Video Understanding
• Paper Explained — GenFirst: Let Generation Shape the Latent Space First, Reconstruction SecondGenFirst: Generation Before Reconstruction for Stable End-to-End Latent Generative Modeling
• Paper Walkthrough: AURORA-LM — A Diffusion Language Model That Keeps Its Text Latent Full-Width Instead of Compressing ItAURORA-LM: Autoencoding Unified Representation for Continuous-Latent Diffusion Language Modeling