Explore modules

Variational Autoencoder

Visualize the encoding of data into a probabilistic latent space and its reconstruction.

Free theory article

Visualize the encoding of data into a probabilistic latent space and its reconstruction.

Concept & interpretation

A Variational Autoencoder (VAE) is a powerful generative model that extends the standard autoencoder architecture. Instead of encoding an input as a single point in a latent space, a VAE encodes it as a probability distribution (typically a Gaussian). This allows the model to map inputs to a continuous, structured latent space. By sampling from this space and decoding the result, the VAE can generate novel, plausible data that resembles the training set.

Read the full free theory article