In this paper, we propose a new approach to train Generative Adversarial Networks (GANs) where we deploy a double-oracle framework using the generator and discriminator oracles. GAN is essentially a two-player zero-sum game between the generator and the discriminator. Training GANs is challenging as a pure Nash equilibrium may not exist and even finding the mixed Nash equilibrium is difficult as GANs have a large-scale strategy space. In DO-GAN, we extend the double oracle framework to GANs. We first generalize the players' strategies as the trained models of generator and discriminator from the best response oracles. We then compute the meta-strategies using a linear program. For scalability of the framework where multiple generators and discriminator best responses are stored in the memory, we propose two solutions: 1) pruning the weakly-dominated players' strategies to keep the oracles from becoming intractable; 2) applying continual learning to retain the previous knowledge of the networks. We apply our framework to established GAN architectures such as vanilla GAN, Deep Convolutional GAN, Spectral Normalization GAN and Stacked GAN. Finally, we conduct experiments on MNIST, CIFAR-10 and CelebA datasets and show that DO-GAN variants have significant improvements in both subjective qualitative evaluation and quantitative metrics, compared with their respective GAN architectures.

中文翻译：

---

DO-GAN：用于生成对抗网络的双重Oracle框架

在本文中，我们提出了一种训练生成对抗网络（GAN）的新方法，其中我们使用生成器和鉴别器Oracle部署双Oracle框架。GAN本质上是生成器和鉴别器之间的两人零和游戏。训练GAN具有挑战性，因为可能不存在纯粹的Nash平衡，而且由于GAN具有大规模的战略空间，因此甚至很难找到混合的Nash平衡。在DO-GAN中，我们将双Oracle框架扩展到GAN。首先，我们将玩家的策略概括为训练有素的生成器和甄别器与最佳响应预言家的模型。然后，我们使用线性程序计算元策略。为了在内存中存储多个生成器和鉴别器最佳响应的框架的可伸缩性，我们提出两种解决方案：1）修剪弱势玩家的策略，以防止神谕变得棘手；2）应用持续学习以保留网络的先前知识。我们将框架应用于已建立的GAN体系结构，例如香草GAN，深度卷积GAN，频谱归一化GAN和堆叠式GAN。最后，我们对MNIST，CIFAR-10和CelebA数据集进行了实验，结果表明，与各自的GAN架构相比，DO-GAN变体在主观定性评估和定量指标方面均具有显着改善。频谱归一化GAN和堆叠GAN。最后，我们对MNIST，CIFAR-10和CelebA数据集进行了实验，结果表明，与各自的GAN架构相比，DO-GAN变体在主观定性评估和定量指标方面均具有显着改善。频谱归一化GAN和堆叠GAN。最后，我们对MNIST，CIFAR-10和CelebA数据集进行了实验，结果表明，与各自的GAN架构相比，DO-GAN变体在主观定性评估和定量指标方面均具有显着改善。