Variational Autoencoder

Models that learn to generate samples resembling training data.

Two-network setup where generator fools a discriminator.

Generator produces limited variety of outputs.

Exact likelihood generative models using invertible transforms.

Learns the score (∇ log p(x)) for generative sampling.

Controls amount of noise added at each diffusion step.

Diffusion performed in latent space for efficiency.

Model that compresses input into latent space and reconstructs it.