BriefGPT.xyz
Jul, 2024
扩散模型分离潜变空间的等距表示学习
Isometric Representation Learning for Disentangled Latent Space of Diffusion Models
HTML
PDF
Jaehoon Hahm, Junho Lee, Sunghyun Kim, Joonseok Lee
TL;DR
使用等距扩散进行几何规则化，使扩散模型学习到一个几何上完整的训练数据流形的潜空间，实现了更加解耦的潜空间、更平滑的插值、更准确的反演以及对潜空间中属性的更精确控制。
Abstract
The
latent space
of
diffusion model
mostly still remains unexplored, despite its great success and potential in the field of generative modeling. In fact, the
→