Artificial Intelligence

    

Deterministic Degradation Process for Diffusion GAN and Its Inversion

Authors: Jeongik Cho

Recently, diffusion models have shown impressive generative performance. However, they have the disadvantage of having a high latent dimension and slow sampling speed. To increase the sampling speed of diffusion models, diffusion GANs have been proposed. But the latent dimension of diffusion GANs using non-deterministic degradation is still high, making it difficult to invert the generative model. In this paper, we introduce an invertible diffusion GAN that uses deterministic degradation. Our proposed method performs inverse diffusion using deterministic degradation without a model, and the generator of the GAN is trained to perform the diffusion process with the latent random variable. The proposed method uses deterministic degradation, so the latent dimension is low enough to be invertible.

Comments: 10 Pages.

Download: PDF

Submission history

[v1] 2023-02-25 22:10:48

Unique-IP document downloads: 285 times

Vixra.org is a pre-print repository rather than a journal. Articles hosted may not yet have been verified by peer-review and should be treated as preliminary. In particular, anything that appears to include financial or legal advice or proposed medical treatments should be treated with due caution. Vixra.org will not be responsible for any consequences of actions that result from any form of use of any documents on this website.

Add your own feedback and questions here:
You are equally welcome to be positive or negative about any paper but please be polite. If you are being critical you must mention at least one specific error, otherwise your comment will be deleted as unhelpful.

comments powered by Disqus