GAN and VAE from an Optimal Transport Point of View
arXiv:1706.01807
Abstract
This short article revisits some of the ideas introduced in arXiv:1701.07875 and arXiv:1705.07642 in a simple setup. This sheds some lights on the connexions between Variational Autoencoders (VAE), Generative Adversarial Networks (GAN) and Minimum Kantorovitch Estimators (MKE).
References in corpus (2)
Cited by in corpus (18)
- Defense Against Adversarial Attacks Using Feature Scattering-based Adversarial Training
- A survey on GANs for computer vision: Recent research, analysis and taxonomy
- Improving GANs Using Optimal Transport
- Monge-Ampère Flow for Generative Modeling
- Autoregressive Quantile Networks for Generative Modeling
- Regularized Optimal Transport and the Rot Mover's Distance
- On parameter estimation with the Wasserstein distance
- Linear Time Sinkhorn Divergences using Positive Features
- On Wasserstein Reinforcement Learning and the Fokker-Planck equation
- Asymptotic Guarantees for Learning Generative Models with the Sliced-Wasserstein Distance
- Sliced-Wasserstein Flows: Nonparametric Generative Modeling via Optimal Transport and Diffusions
- Conditional deep surrogate models for stochastic, high-dimensional, and multi-fidelity systems
- A gradual, semi-discrete approach to generative network training via explicit Wasserstein minimization
- SDE approximations of GANs training and its long-run behavior
- k-GANs: Ensemble of Generative Models with Semi-Discrete Optimal Transport
- A Chain Graph Interpretation of Real-World Neural Networks
- Progressive VAE Training on Highly Sparse and Imbalanced Data
- Loosely Coupled Federated Learning Over Generative Models