There has been a drastic growth of research in generative adversarial nets
(GANs) in the past few years. Proposed in 2014, GAN has been applied to various
applications such as computer vision and natural language processing, and
achieves impressive performance. Among the many applicati
生成对抗网络(GANs)是在多个领域中生成逼真、多样化数据的强大工具。本研究综述了 GANs 的潜在架构、验证指标和应用领域,并探讨了 GAN 与 Jensen-Shannon 散度之间的深刻联系以及 GAN 框架的最优性特征。对 GAN 变体的效率、模型架构以及训练障碍和解决方案进行了评估,并详细讨论了 GAN 与 Transformer、物理信息神经网络、大型语言模型和扩散模型等新开发的深度学习框架的整合。最后,揭示了该领域的若干问题和未来研究方向。