Density estimation using Real NVP

27 May 2016Laurent Dinh • Jascha Sohl-Dickstein • Samy Bengio

Unsupervised learning of probabilistic models is a central yet challenging problem in machine learning. Specifically, designing models with tractable learning, sampling, inference and evaluation is crucial in solving this task. We extend the space of such models using real-valued non-volume preserving (real NVP) transformations, a set of powerful invertible and learnable transformations, resulting in an unsupervised learning algorithm with exact log-likelihood computation, exact sampling, exact inference of latent variables, and an interpretable latent space.

Full paper

Evaluation


Task Dataset Model Metric name Metric value Global rank Compare
Image Generation CIFAR-10 PixelRNN Model Entropy 3.0 # 14
Image Generation CIFAR-10 Real NVP Model Entropy 3.49 # 15