Reinterpreting Importance-Weighted Autoencoders

Chris Cremer, Quaid Morris, David Duvenaud

Abstract: The standard interpretation of importance-weighted autoencoders is that they maximize a tighter lower bound on the marginal likelihood. We give an alternate interpretation of this procedure: that it optimizes the standard variational lower bound, but using a more complex distribution. We formally derive this result, and visualize the implicit importance-weighted approximate posterior.

TL;DR: IWAE optimizes the standard variational lowerbound, but using a more complex variational distribution

Keywords: Unsupervised Learning

Conflicts: cs.toronto.edu, harvard.edu, cam.ac.uk