Wasserstein Divergence for GANs
This addresses a key bottleneck in generative adversarial networks for computer vision by improving stability and performance, though it is an incremental advancement over existing Wasserstein GANs.
The paper tackles the challenge of approximating the Lipschitz constraint in Wasserstein GANs by proposing a novel Wasserstein divergence (W-div) that relaxes this requirement, resulting in a stable method (WGAN-div) that shows superior performance on image synthesis benchmarks compared to state-of-the-art methods.
In many domains of computer vision, generative adversarial networks (GANs) have achieved great success, among which the family of Wasserstein GANs (WGANs) is considered to be state-of-the-art due to the theoretical contributions and competitive qualitative performance. However, it is very challenging to approximate the $k$-Lipschitz constraint required by the Wasserstein-1 metric~(W-met). In this paper, we propose a novel Wasserstein divergence~(W-div), which is a relaxed version of W-met and does not require the $k$-Lipschitz constraint. As a concrete application, we introduce a Wasserstein divergence objective for GANs~(WGAN-div), which can faithfully approximate W-div through optimization. Under various settings, including progressive growing training, we demonstrate the stability of the proposed WGAN-div owing to its theoretical and practical advantages over WGANs. Also, we study the quantitative and visual performance of WGAN-div on standard image synthesis benchmarks of computer vision, showing the superior performance of WGAN-div compared to the state-of-the-art methods.