Сравнение методов

	Нормализующие потоки	Диффузионная модель	Вариационный автокодировщик
Область	Глубокое обучение	Глубокое обучение	Глубокое обучение
Семейство	Machine learning	Machine learning	Machine learning
Год появления	2015	2020	2014
Автор метода	Danilo Rezende & Shakir Mohamed	Ho, J., Jain, A. & Abbeel, P.	Kingma, D. P. & Welling, M.
Тип	Generative model via invertible transformations	Generative deep learning (denoising diffusion)	Deep generative latent-variable model (encoder–decoder)
Основополагающий источник	Rezende, D. J., & Mohamed, S. (2015). Variational inference with normalizing flows. International Conference on Machine Learning (ICML), 1530–1538.	Ho, J., Jain, A. & Abbeel, P. (2020). Denoising Diffusion Probabilistic Models. NeurIPS.	Kingma, D. P. & Welling, M. (2014). Auto-Encoding Variational Bayes. International Conference on Learning Representations (ICLR).
Другие названия	Flow-Based Generative Models, Invertible Neural Networks, Exact Likelihood Models, Akışa Dayalı Üretici Modeller	Difüzyon Modeli (DDPM / Stable Diffusion), difüzyon modeli, denoising diffusion model, DDPM	Değişkensel Otokodlayıcı (VAE), VAE, auto-encoding variational Bayes, deep latent variable model
Связанные	2	4	5
Сводка	Normalizing flows are a class of generative models that learn a complex probability distribution by applying a sequence of invertible, differentiable transformations to a simple base distribution such as a standard Gaussian. Introduced by Rezende and Mohamed (2015) in the context of variational inference, they enable exact likelihood computation and efficient sampling, making them a principled alternative to VAEs and GANs for density estimation and generation tasks.	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.	The Variational Autoencoder (VAE) is a deep generative latent-variable model, introduced by Diederik Kingma and Max Welling in 2014, that encodes data as a probability distribution in a latent space and samples from that distribution to generate new examples. It is used for data generation, anomaly detection, and feature learning.