Exploring and Exploiting Hubness Priors for High-Quality GAN Latent Sampling

TL;DR

This paper introduces a novel method for sampling high-quality images from GANs by leveraging the hubness priors in the latent space, providing a more efficient and theoretically grounded approach than existing tricks.

Contribution

The paper proposes a new a priori sampling method based on hubness priors, offering a theoretical understanding and practical improvement over traditional truncation tricks.

Findings

Hub latents have higher sampling densities and contribute to better image quality.

The truncation trick is an approximation of the central clustering of hub latents.

Experimental results validate the effectiveness of the proposed method.

Abstract

Despite the extensive studies on Generative Adversarial Networks (GANs), how to reliably sample high-quality images from their latent spaces remains an under-explored topic. In this paper, we propose a novel GAN latent sampling method by exploring and exploiting the hubness priors of GAN latent distributions. Our key insight is that the high dimensionality of the GAN latent space will inevitably lead to the emergence of hub latents that usually have much larger sampling densities than other latents in the latent space. As a result, these hub latents are better trained and thus contribute more to the synthesis of high-quality images. Unlike the a posterior "cherry-picking", our method is highly efficient as it is an a priori method that identifies high-quality latents before the synthesis of images. Furthermore, we show that the well-known but purely empirical truncation trick is a naive approximation to the central clustering effect of hub latents, which not only uncovers the rationale of the truncation trick, but also indicates the superiority and fundamentality of our method. Extensive experimental results demonstrate the effectiveness of the proposed method.