Primal-Dual Wasserstein GAN
arXiv:1805.09575
Abstract
We introduce Primal-Dual Wasserstein GAN, a new learning algorithm for building latent variable models of the data distribution based on the primal and the dual formulations of the optimal transport (OT) problem. We utilize the primal formulation to learn a flexible inference mechanism and to create an optimal approximate coupling between the data distribution and the generative model. In order to learn the generative model, we use the dual formulation and train the decoder adversarially through a critic network that is regularized by the approximate coupling obtained from the primal. Unlike previous methods that violate various properties of the optimal critic, we regularize the norm and the direction of the gradients of the critic function. Our model shares many of the desirable properties of auto-encoding models in terms of mode coverage and latent structure, while avoiding their undesirable averaging properties, e.g. their inability to capture sharp visual features when modeling real images. We compare our algorithm with several other generative modeling techniques that utilize Wasserstein distances on Frechet Inception Distance (FID) and Inception Scores (IS).
arXiv comment14 pages, 16 figures
References in corpus (7)
- Auto-Encoding Variational Bayes
- Stochastic Backpropagation and Approximate Inference in Deep Generative Models
- Improved Training of Wasserstein GANs
- Adversarial Feature Learning
- Adversarially Learned Inference
- Adversarial Autoencoders
- Many Paths to Equilibrium: GANs Do Not Need to Decrease a Divergence At Every Step
Cited by in corpus (8)
- Sorting out Lipschitz function approximation
- Adversarial Lipschitz Regularization
- Generative Minimization Networks: Training GANs Without Competition
- A Wasserstein GAN model with the total variational regularization
- k-GANs: Ensemble of Generative Models with Semi-Discrete Optimal Transport
- Wasserstein GANs with Gradient Penalty Compute Congested Transport
- Pairwise Augmented GANs with Adversarial Reconstruction Loss
- Input Convex Gradient Networks