Unsupervised Discovery of Interpretable Directions in the GAN Latent Space

TL;DR

This paper presents an unsupervised approach to discover interpretable semantic directions in GAN latent spaces, enabling controllable image manipulations without requiring labels or pretraining, and demonstrates practical applications like saliency detection.

Contribution

The authors introduce a simple, model-agnostic unsupervised method to identify meaningful directions in GAN latent spaces, expanding the scope of discoverable transformations.

Findings

Identified directions for background removal and other manipulations.

Achieved competitive results in weakly-supervised saliency detection.

Revealed non-trivial semantic directions without supervision.

Abstract

The latent spaces of GAN models often have semantically meaningful directions. Moving in these directions corresponds to human-interpretable image transformations, such as zooming or recoloring, enabling a more controllable generation process. However, the discovery of such directions is currently performed in a supervised manner, requiring human labels, pretrained models, or some form of self-supervision. These requirements severely restrict a range of directions existing approaches can discover. In this paper, we introduce an unsupervised method to identify interpretable directions in the latent space of a pretrained GAN model. By a simple model-agnostic procedure, we find directions corresponding to sensible semantic manipulations without any form of (self-)supervision. Furthermore, we reveal several non-trivial findings, which would be difficult to obtain by existing methods, e.g., a direction corresponding to background removal. As an immediate practical benefit of our work, we show how to exploit this finding to achieve competitive performance for weakly-supervised saliency detection.