Сравнение методов

	Wasserstein GAN (WGAN)	CycleGAN	Диффузионная модель
Область	Глубокое обучение	Глубокое обучение	Глубокое обучение
Семейство	Machine learning	Machine learning	Machine learning
Год появления	2017	2017	2020
Автор метода	Martín Arjovsky, Soumith Chintala & Léon Bottou	Jun-Yan Zhu et al.	Ho, J., Jain, A. & Abbeel, P.
Тип	Generative adversarial network variant	Unsupervised image-to-image translation	Generative deep learning (denoising diffusion)
Основополагающий источник	Arjovsky, M., Chintala, S., & Bottou, L. (2017). Wasserstein generative adversarial networks. International Conference on Machine Learning (ICML), 214–223.	Zhu, J.-Y., Park, T., Isola, P., & Efros, A. A. (2017). Unpaired image-to-image translation using cycle-consistent adversarial networks. IEEE International Conference on Computer Vision (ICCV), 2242–2251.	Ho, J., Jain, A. & Abbeel, P. (2020). Denoising Diffusion Probabilistic Models. NeurIPS.
Другие названия	WGAN, Earth-Mover GAN, Wasserstein Generative Adversarial Network, Wasserstein-GAN	Cycle-Consistent Adversarial Networks, Unpaired Image-to-Image Translation, Cycle-GAN, Çevrimsel Tutarlı GAN	Difüzyon Modeli (DDPM / Stable Diffusion), difüzyon modeli, denoising diffusion model, DDPM
Связанные	3	3	4
Сводка	Wasserstein GAN (WGAN) is a generative adversarial network variant introduced by Arjovsky, Chintala, and Bottou in 2017 that replaces the Jensen-Shannon divergence used in the original GAN with the Wasserstein-1 (Earth Mover) distance. This substitution provides a theoretically grounded training objective that yields more stable optimization and a loss value that correlates meaningfully with generated sample quality, addressing the notorious mode collapse and vanishing gradient problems of standard GANs.	CycleGAN, introduced by Zhu et al. at ICCV 2017, learns to translate images between two visual domains without requiring paired training examples. It trains two generators and two discriminators simultaneously, enforcing a cycle-consistency constraint so that an image translated from domain X to Y and back again recovers the original. This makes it applicable whenever large aligned datasets are unavailable, such as converting photographs to artwork styles, turning summer landscapes into winter scenes, or mapping satellite imagery to map tiles.	A diffusion model is a generative deep-learning method, introduced by Ho, Jain and Abbeel in 2020 (DDPM), that learns to produce high-quality images, audio and molecular structures by reversing a step-by-step noising process. It has largely displaced GANs as the current state of the art in generative modelling.