RPGAN: GANs Interpretability via Random Routing
This provides a tool for interpretability in generative models, which is an incremental improvement for researchers and practitioners in machine learning.
The paper tackles the problem of interpreting GANs by introducing RPGAN, which uses random paths in the generator network as its latent space to analyze factors of variation across layers, and demonstrates competitive generation quality and efficient incremental learning on new data.
In this paper, we introduce Random Path Generative Adversarial Network (RPGAN) -- an alternative design of GANs that can serve as a tool for generative model analysis. While the latent space of a typical GAN consists of input vectors, randomly sampled from the standard Gaussian distribution, the latent space of RPGAN consists of random paths in a generator network. As we show, this design allows to understand factors of variation, captured by different generator layers, providing their natural interpretability. With experiments on standard benchmarks, we demonstrate that RPGAN reveals several interesting insights about the roles that different layers play in the image generation process. Aside from interpretability, the RPGAN model also provides competitive generation quality and allows efficient incremental learning on new data.