paper-with-me

홈 › Papers

Generalization and Stability of GANs: A theory and promise from data augmentation

2021-01-01 · Khoat Than, Nghia Vu

The instability when training generative adversarial networks (GANs) is a notoriously difficult issue, and the generalization of GANs remains open. In this paper, we will analyze various sources of instability which not only come from the discriminator but also the generator. We then point out that the requirement of Lipschitz continuity on both the discriminator and generator leads to generalization and stability for GANs. As a consequence, this work naturally provides a generalization bound for a large class of existing models and explains the success of recent large-scale generators. Finally, we show why data augmentation can ensure Lipschitz continuity on both the discriminator and generator. This work therefore provides a theoretical basis for a simple way to ensure generalization in GANs, explaining the highly successful use of data augmentation for GANs in practice.

📄 PDF Abstract BibTeX

Code (0)

등록된 구현이 없습니다.

Tasks

Data Augmentation

Similar Papers 제목 키워드 기반

CHAIN: Enhancing Generalization in Data-Efficient GANs via lipsCHitz continuity constrAIned Normalization

2024-03-31 · CVPR 2024 1 · Yao Ni, Piotr Koniusz

Generative Adversarial Networks (GANs) significantly advanced image generation but their performance heavily depends on abundant training data. In scenarios with limited data, GANs often struggle with discriminator overf…

Image Generation

Generalization of GANs and overparameterized models under Lipschitz continuity

2021-04-06 · Khoat Than, Nghia Vu

Generative adversarial networks (GANs) are so complex that the existing learning theories do not provide a satisfactory explanation for why GANs have great success in practice. The same situation also remains largely ope…

Data AugmentationGeneralization Bounds

(q,p)-Wasserstein GANs: Comparing Ground Metrics for Wasserstein GANs

2019-02-10 · Anton Mallasto, Jes Frellsen, Wouter Boomsma, Aasa Feragen

Generative Adversial Networks (GANs) have made a major impact in computer vision and machine learning as generative models. Wasserstein GANs (WGANs) brought Optimal Transport (OT) theory into GANs, by minimizing the $1$-…

Improving Generalization and Stability of Generative Adversarial Networks

2019-02-11 · ICLR 2019 5 · Hoang Thanh-Tung, Truyen Tran, Svetha Venkatesh

Generative Adversarial Networks (GANs) are one of the most popular tools for learning complex high dimensional distributions. However, generalization properties of GANs have not been well understood. In this paper, we an…

Understanding and Stabilizing GANs' Training Dynamics with Control Theory

2019-09-29 · Kun Xu, Chongxuan Li, Jun Zhu, Bo Zhang

Generative adversarial networks (GANs) are effective in generating realistic images but the training is often unstable. There are existing efforts that model the training dynamics of GANs in the parameter space but the a…

Image GenerationL2 Regularization