Autocodificadores variacionales (VAE)
Definición
Los VAEs aprenden un espacio latente entrenando un encoder-decoder with a variational (reparameterized) objective. They support generation and smooth interpolation in latent space.
Se diferencian de GANs (adversariales) y difusión (eliminación de ruido): el espacio latente está regularizado (KL hacia un prior) so it is smooth and interpretable. Generation can be blurrier than GANs/diffusion, but VAEs are useful for representation learning, anomaly detection, and when a low-D latent is desired.
Cómo funciona
Input se pasa a un encoder que produce parameters of a latent distribution (por ej. mean and log-variance for Gaussian). A z vector is sampled (reparameterization trick: z = mean + std * epsilon) and fed to the decoder, which reconstructs the input. Loss = reconstruction loss (por ej. MSE or cross-entropy) + KL divergence from the latent to a prior (por ej. standard normal). The KL term regularizes the latent space; the reconstruction term keeps it informative. At generation time, sample z from the prior and run the decoder.
Casos de uso
VAEs suit tasks that need a continuous latent space: smooth generation, anomaly detection, or learned representations.
- Generative modeling with smooth latent interpolation
- Anomaly detection via reconstruction error
- Learned representations for downstream tasks