Encoding Invariances in Deep Generative Models

06/04/2019 ∙ by Viraj Shah, et al. ∙ 5 ∙ share

Reliable training of generative adversarial networks (GANs) typically require massive datasets in order to model complicated distributions. However, in several applications, training samples obey invariances that are a priori known; for example, in complex physics simulations, the training data obey universal laws encoded as well-defined mathematical equations. In this paper, we propose a new generative modeling approach, InvNet, that can efficiently model data spaces with known invariances. We devise an adversarial training algorithm to encode them into data distribution. We validate our framework in three experimental settings: generating images with fixed motifs; solving nonlinear partial differential equations (PDEs); and reconstructing two-phase microstructures with desired statistical properties. We complement our experiments with several theoretical results.

READ FULL TEXT

Enforcing Statistical Constraints in Generative Adversarial Networks for Modeling Chaotic Dynamical Systems

Simulating complex physical systems often involves solving partial diffe...
05/13/2019 ∙ by Jin-Long Wu, et al. ∙ 6 ∙ share

read it
Inverse Estimation of Elastic Modulus Using Physics-Informed Generative Adversarial Networks

While standard generative adversarial networks (GANs) rely solely on tra...
05/20/2020 ∙ by James E. Warner, et al. ∙ 0 ∙ share

read it
Physics-informed deep generative models

We consider the application of deep generative models in propagating unc...
12/09/2018 ∙ by Yibo Yang, et al. ∙ 0 ∙ share

read it
Generative Adversarial Networks

We propose a new framework for estimating generative models via an adver...
06/10/2014 ∙ by Ian J. Goodfellow, et al. ∙ 0 ∙ share

read it
On Unifying Deep Generative Models

Deep generative models have achieved impressive success in recent years....
06/02/2017 ∙ by Zhiting Hu, et al. ∙ 0 ∙ share

read it
Analyzing and Improving Generative Adversarial Training for Generative Modeling and Out-of-Distribution Detection

Generative adversarial training (GAT) is a recently introduced adversari...
12/11/2020 ∙ by Xuwang Yin, et al. ∙ 12 ∙ share

read it
Generative adversarial training of product of policies for robust and adaptive movement primitives

In learning from demonstrations, many generative models of trajectories ...
11/06/2020 ∙ by Emmanuel Pignat, et al. ∙ 0 ∙ share

read it

Encoding Invariances in Deep Generative Models

Authors

Enforcing Statistical Constraints in Generative Adversarial Networks for Modeling Chaotic Dynamical Systems

Inverse Estimation of Elastic Modulus Using Physics-Informed Generative Adversarial Networks

Physics-informed deep generative models

Generative Adversarial Networks

On Unifying Deep Generative Models

Analyzing and Improving Generative Adversarial Training for Generative Modeling and Out-of-Distribution Detection

Generative adversarial training of product of policies for robust and adaptive movement primitives

This week in AI