Machine learningGenerative models
瓦瑟施泰因生成对抗网络 (WGAN)
Wasserstein GAN (WGAN) 是 Arjovsky、Chintala 和 Bottou 于 2017 年提出的一种生成对抗网络变体,它用 Wasserstein-1(Earth Mover)距离取代了原始 GAN 中使用的 Jensen-Shannon 散度。这种替换提供了一个理论上合理的训练目标,能够实现更稳定的优化,并且损失值与生成样本的质量具有有意义的相关性,从而解决了标准 GAN 中臭名昭著的模式崩溃和梯度消失问题。
阅读完整方法
仅限会员
登录使用免费账户登录即可阅读本节。
Method map
The neighbourhood of related methods — select a node to explore.
来源
- Arjovsky, M., Chintala, S., & Bottou, L. (2017). Wasserstein generative adversarial networks. International Conference on Machine Learning (ICML), 214–223. link ↗
如何引用本页
ScholarGate. (2026, June 2). Wasserstein GAN (WGAN). ScholarGate. https://scholargate.app/zh/deep-learning/wasserstein-gan
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
Compare side by side →