Informative GANs via Structured Regularization of Optimal Transport

TL;DR

This paper introduces a novel GAN framework that uses structured regularization of optimal transport to learn disentangled, interpretable latent representations, demonstrating improved interpretability and stability in training.

Contribution

It proposes a new informative OT-based GAN that regularizes the transportation plan to encourage informative latent subspaces, advancing disentangled representation learning.

Findings

Regularizations lead to disentangled, interpretable latent representations

Stable training algorithm for the proposed GAN

Effective discovery of structured latent spaces

Abstract

We tackle the challenge of disentangled representation learning in generative adversarial networks (GANs) from the perspective of regularized optimal transport (OT). Specifically, a smoothed OT loss gives rise to an implicit transportation plan between the latent space and the data space. Based on this theoretical observation, we exploit a structured regularization on the transportation plan to encourage a prescribed latent subspace to be informative. This yields the formulation of a novel informative OT-based GAN. By convex duality, we obtain the equivalent view that this leads to perturbed ground costs favoring sparsity in the informative latent dimensions. Practically, we devise a stable training algorithm for the proposed informative GAN. Our experiments support the hypothesis that such regularizations effectively yield the discovery of disentangled and interpretable latent representations. Our work showcases potential power of a regularized OT framework in the context of generative modeling through its access to the transport plan. Further challenges are addressed in this line.