# Improving GAN Training with Probability Ratio Clipping and Sample Reweighting

arXiv: https://arxiv.org/abs/2006.06900 · HTML: https://mlsanity.com/paper/2006.06900/

## Key facts

- Cited by 18 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Jun 2020.

## Built on by (most cited first, 18 of 18)

- [Knowledge Distillation in Iterative Generative Models for Improved Sampling Speed](https://mlsanity.com/paper/2101.02388/) (2021)
- [Better Generative Replay for Continual Federated Learning](https://mlsanity.com/paper/2302.13001/) (2023)
- [Math Word Problem Generation with Mathematical Consistency and Problem Context Constraints](https://mlsanity.com/paper/2109.04546/) (2021)
- [Self-Diagnosing GAN: Diagnosing Underrepresented Samples in Generative Adversarial Networks](https://mlsanity.com/paper/2102.12033/) (2021)
- [Unaligned Image-to-Image Translation by Learning to Reweight](https://mlsanity.com/paper/2109.11736/) (2021)
- [Federated Few-Shot Learning with Adversarial Learning](https://mlsanity.com/paper/2104.00365/) (2021)
- [Unified Generation, Reconstruction, and Representation: Generalized Diffusion with Adaptive Latent Encoding-Decoding](https://mlsanity.com/paper/2402.19009/) (2024)
- [Toward a `Standard Model' of Machine Learning](https://mlsanity.com/paper/2108.07783/) (2021)
- [Fine-tuning Language Models with Generative Adversarial Reward Modelling](https://mlsanity.com/paper/2305.06176/) (2023)
- [EXoN: EXplainable encoder Network](https://mlsanity.com/paper/2105.10867/) (2021)
- [Cooperative Distribution Alignment via JSD Upper Bound](https://mlsanity.com/paper/2207.02286/) (2022)
- [Qffusion: Controllable Portrait Video Editing via Quadrant-Grid Attention Learning](https://mlsanity.com/paper/2501.06438/) (2025)
- [Towards Practical Non-Adversarial Distribution Matching](https://mlsanity.com/paper/2310.19690/) (2023)
- [Generating Realistic X-ray Scattering Images Using Stable Diffusion and Human-in-the-loop Annotations](https://mlsanity.com/paper/2408.12720/) (2024)
- [Information-Theoretic Generative Clustering of Documents](https://mlsanity.com/paper/2412.13534/) (2024)
- [GDA-AM: On the effectiveness of solving minimax optimization via Anderson Acceleration](https://mlsanity.com/paper/2110.02457/) (2021)
- [Securing Transfer-Learned Networks with Reverse Homomorphic Encryption](https://mlsanity.com/paper/2505.14323/) (2025)
- [Expressive Score-Based Priors for Distribution Matching with Geometry-Preserving Regularization](https://mlsanity.com/paper/2506.14607/) (2025)

## Abstract (arXiv)

Despite success on a wide range of problems related to vision, generative adversarial networks (GANs) often suffer from inferior performance due to unstable training, especially for text generation. To solve this issue, we propose a new variational GAN training framework which enjoys superior training stability. Our approach is inspired by a connection of GANs and reinforcement learning under a variational perspective. The connection leads to (1) probability ratio clipping that regularizes generator training to prevent excessively large updates, and (2) a sample re-weighting mechanism that improves discriminator training by downplaying bad-quality fake samples. Moreover, our variational GAN framework can provably overcome the training issue in many GANs that an optimal discriminator cannot provide any informative gradient to training generator. By plugging the training approach in diverse state-of-the-art GAN architectures, we obtain significantly improved performance over a range of tasks, including text generation, text style transfer, and image generation.