Architectures · intermediate · concept 89 of 311 · Updated Nov 4, 2026

🏗️Autoencoders

A network trained to compress its input into a small code and reconstruct it, learning structure with no labels at all. The bottleneck forces the model to keep only what matters. Variants denoise images, detect anomalies, and, as VAEs, define the latent spaces diffusion models paint in.