Discriminating Against Unrealistic Interpolations in Generative Adversarial Networks

TL;DR

This paper proposes a method to improve latent space interpolations in GANs by leveraging the discriminator to avoid unrealistic samples, resulting in more meaningful and high-quality generated interpolations.

Contribution

It introduces a lightweight approach that reuses the discriminator to modify the latent space metric, enhancing the realism of interpolations in pre-trained GANs.

Findings

Discriminator-based metric effectively avoids unrealistic interpolations.

Improved smoothness and realism in generated sample paths.

Method is lightweight and applicable to pre-trained models.

Abstract

Interpolations in the latent space of deep generative models is one of the standard tools to synthesize semantically meaningful mixtures of generated samples. As the generator function is non-linear, commonly used linear interpolations in the latent space do not yield the shortest paths in the sample space, resulting in non-smooth interpolations. Recent work has therefore equipped the latent space with a suitable metric to enforce shortest paths on the manifold of generated samples. These are often, however, susceptible of veering away from the manifold of real samples, resulting in smooth but unrealistic generation that requires an additional method to assess the sample quality along paths. Generative Adversarial Networks (GANs), by construction, measure the sample quality using its discriminator network. In this paper, we establish that the discriminator can be used effectively to avoid regions of low sample quality along shortest paths. By reusing the discriminator network to modify the metric on the latent space, we propose a lightweight solution for improved interpolations in pre-trained GANs.