Improving Variational Inference with Inverse Autoregressive Flow

Understand

The framework of normalizing flows provides a general strategy for flexible variational inference of posteriors over latent variables.

• We propose a new type of normalizing flow, inverse autoregressive flow (IAF), that, in contrast to earlier published flows, scales well to high-dimensional latent spaces.
• The proposed flow consists of a chain of invertible transformations, where each transformation is based on an autoregressive neural network.
• In experiments, we show that IAF significantly improves upon diagonal Gaussian approximate posteriors.