Bridging the Gap Between $f$-GANs and Wasserstein GANs

By Jiaming Song, Stefano Ermon. 2019

46 citations recorded by the source corpus.

Generative adversarial networks (GANs) have enjoyed much success in learning high-dimensional distributions. Learning objectives approximately minimize an $f$-divergence ($f$-GANs) or an integral probability metric (Wasserstein GANs) between the model and the data distribution using a discriminator. Wasserstein GANs enjoy superior empirical performance, but in $f$-GANs the discriminator can be interpreted as a density ratio estimator which is necessary in some GAN applications. In this paper, we bridge the gap between $f$-GANs and Wasserstein GANs (WGANs). First, we list two constraints over variational $f$-divergence estimation objectives that preserves the optimal solution. Next, we minimize over a Lagrangian relaxation of the constrained objective, and show that it generalizes critic objectives of both $f$-GAN and WGAN. Based on this generalization, we propose a novel practical objective, named KL-Wasserstein GAN (KL-WGAN). We demonstrate empirical success of KL-WGAN on synthetic datasets and real-world image generation benchmarks, and achieve state-of-the-art FID scores on CIFAR10 image generation.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Ming-Hsuan Yang
2. Hung-Yu Tseng
3. Lu Jiang
4. Ce Liu
5. Weilong Yang
6. Luc Van Gool
7. Wenguan Wang
8. Dacheng Tao

Citations in the corpus

• Regularizing Generative Adversarial Networks under Limited Data (2021, methodology): “On the other hand, some models are designed to minimize the integral probability metrics (IPM) [52, 64], such as the WGAN [2, 15] frameworks.”