Log In Sign Up

Refining Deep Generative Models via Wasserstein Gradient Flows

by   Abdul Fatir Ansari, et al.

Deep generative modeling has seen impressive advances in recent years, to the point where it is now commonplace to see simulated samples (e.g., images) that closely resemble real-world data. However, generation quality is generally inconsistent for any given model and can vary dramatically between samples. We introduce Discriminator Gradient flow (DGflow), a new technique that improves generated samples via the gradient flow of entropy-regularized f-divergences between the real and the generated data distributions. The gradient flow takes the form of a non-linear Fokker-Plank equation, which can be easily simulated by sampling from the equivalent McKean-Vlasov process. By refining inferior samples, our technique avoids wasteful sample rejection used by previous methods (DRS MH-GAN). Compared to existing works that focus on specific GAN variants, we show our refinement approach can be applied to GANs with vector-valued critics and even other deep generative models such as VAEs and Normalizing Flows. Empirical results on multiple synthetic, image, and text datasets demonstrate that DGflow leads to significant improvement in the quality of generated samples for a variety of generative models, outperforming the state-of-the-art Discriminator Optimal Transport (DOT) and Discriminator Driven Latent Sampling (DDLS) methods.


page 7

page 18

page 21

page 22

page 23

page 24


Latent Space Refinement for Deep Generative Models

Deep generative models are becoming widely used across science and indus...

Deep Generative Learning via Variational Gradient Flow

We propose a general framework to learn deep generative models via Varia...

On Unifying Deep Generative Models

Deep generative models have achieved impressive success in recent years....

Simultaneous Multiple-Prompt Guided Generation Using Differentiable Optimal Transport

Recent advances in deep learning, such as powerful generative models and...

Monge-Ampère Flow for Generative Modeling

We present a deep generative model, named Monge-Ampère flow, which build...

Spectral Distribution aware Image Generation

Recent advances in deep generative models for photo-realistic images hav...

StrokeGAN: Reducing Mode Collapse in Chinese Font Generation via Stroke Encoding

The generation of stylish Chinese fonts is an important problem involved...