arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2308.06342cs.LG
收藏

Mirror Diffusion Models(镜像扩散模型)

Mirror Diffusion Models

Yale University(耶鲁大学)

Jaesung Tae

AI总结:

针对离散类别数据及受约束域扩散建模难题,该理论报告受 mirror Langevin 算法启发提出 MDMs,并在单纯形扩散中验证,同时扩展至图像与文本生成。

摘要:

扩散模型已成功应用于多种连续域中的生成任务。然而,将扩散应用于离散类别数据仍是一项非平凡任务。此外,连续域中的生成在实践中通常需要裁剪,这促使人们需要一种将扩散适配到受约束域的理论框架。受用于受约束采样问题的 mirror Langevin 算法启发,在这份理论报告中,我们提出 Mirror Diffusion Models(MDMs,镜像扩散模型)。我们在单纯形扩散的背景下展示 MDMs,并提出向图像和文本生成等流行域的自然扩展。

英文摘要:

Diffusion models have successfully been applied to generative tasks in various continuous domains. However, applying diffusion to discrete categorical data remains a non-trivial task. Moreover, generation in continuous domains often requires clipping in practice, which motivates the need for a theoretical framework for adapting diffusion to constrained domains. Inspired by the mirror Langevin algorithm for the constrained sampling problem, in this theoretical report we propose Mirror Diffusion Models (MDMs). We demonstrate MDMs in the context of simplex diffusion and propose natural extensions to popular domains such as image and text generation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑