Density estimation using Real NVP

TL;DR

This paper introduces Real NVP, a set of invertible transformations enabling efficient density estimation, sampling, and inference for natural images, with exact likelihood computation and an interpretable latent space.

Contribution

The paper presents Real NVP, a novel class of invertible transformations that extend density estimation capabilities with exact likelihood, sampling, and inference.

Findings

Effective modeling of natural images demonstrated

Exact likelihood and sampling achieved

Interpretable latent space developed

Abstract

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. We demonstrate its ability to model natural images on four datasets through sampling, log-likelihood evaluation and latent variable manipulations.