A neural network, introduced by Diederik Kingma and Max Welling in 2013, that compresses data into a smooth latent space and generates new examples by decoding points from it. It was one of the first deep generative models.
