GANspection

TL;DR

This paper investigates how GANs learn and encode data distributions, demonstrating they capture meaningful semantic information rather than memorizing data, through advanced latent space analysis on LSUN and CelebA datasets.

Contribution

It advances understanding of GANs' learned manifolds by applying interpolation, extrapolation, and vector arithmetic to reveal semantic encoding in the latent space.

Findings

GANs learn a true data probability distribution

Semantic information is encoded in the learned manifold

GANs do not simply memorize data samples

Abstract

Generative Adversarial Networks (GANs) have been used extensively and quite successfully for unsupervised learning. As GANs don't approximate an explicit probability distribution, it's an interesting study to inspect the latent space representations learned by GANs. The current work seeks to push the boundaries of such inspection methods to further understand in more detail the manifold being learned by GANs. Various interpolation and extrapolation techniques along with vector arithmetic is used to understand the learned manifold. We show through experiments that GANs indeed learn a data probability distribution rather than memorize images/data. Further, we prove that GANs encode semantically relevant information in the learned probability distribution. The experiments have been performed on two publicly available datasets - Large Scale Scene Understanding (LSUN) and CelebA.