Unsupervised Panoptic Interpretation of Latent Spaces in GANs Using Space-Filling Vector Quantization
This addresses the challenge of interpretability in GAN latent spaces for researchers and practitioners, offering an unsupervised method that improves over supervised approaches, though it is incremental as it builds on existing vector quantization techniques.
The paper tackles the problem of interpreting latent spaces in GANs by proposing space-filling vector quantization (SFVQ) to capture morphological structures, enabling identification of latent space parts corresponding to specific generative factors and controllable image transformations.
Generative adversarial networks (GANs) learn a latent space whose samples can be mapped to real-world images. Such latent spaces are difficult to interpret. Some earlier supervised methods aim to create an interpretable latent space or discover interpretable directions, which requires exploiting data labels or annotated synthesized samples for training. However, we propose using a modification of vector quantization called space-filling vector quantization (SFVQ), which quantizes the data on a piece-wise linear curve. SFVQ can capture the underlying morphological structure of the latent space, making it interpretable. We apply this technique to model the latent space of pre-trained StyleGAN2 and BigGAN networks on various datasets. Our experiments show that the SFVQ curve yields a general interpretable model of the latent space such that it determines which parts of the latent space correspond to specific generative factors. Furthermore, we demonstrate that each line of the SFVQ curve can potentially refer to an interpretable direction for applying intelligible image transformations. We also demonstrate that the points located on an SFVQ line can be used for controllable data augmentation.