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 2411.18702cs.CVcs.AIcs.LGeess.IV
收藏

带Tweedie的随机游走:基于分数的扩散模型的统一视角

Random Walks with Tweedie: A Unified View of Score-Based Diffusion Models

Washington University in St. Louis(华盛顿大学圣路易斯分校);Los Alamos National Laboratory(洛斯阿拉莫斯国家实验室)

Chicago Y. Park, Michael T. McCann, Cristina Garcia-Cardona, Brendt Wohlberg, Ulugbek S. Kamilov

AI总结:

该研究借助少量基础结论给出基于分数的扩散模型的简洁自洽推导,提出通用训练与采样算法模板,验证了不同模型对应模板内的特定选择,且无需似然近似即可实现条件采样。

摘要:

我们仅借助少数基础教材结论,对若干具有影响力的基于分数的扩散模型给出了简洁推导。扩散模型近年来已成为生成逼真合成信号(尤其是自然图像)的强大工具,在图像处理逆问题的SOTA算法中常发挥作用。这类算法往往出奇地简单,但其背后的理论却并不浅显,现有文献中存在多种复杂的理论阐释。本文面向信号处理领域研究者,为基于分数的扩散模型提供了一套简单且基本自洽的理论依据。该方法可得到用于扩散模型训练与样本生成的通用算法模板。我们证明,若干有影响力的扩散模型均对应这些模板内的特定选择,同时验证了更直接的替代算法选择能取得相当的效果。该方法的额外优势是无需任何似然近似即可实现条件采样。

英文摘要:

We present a concise derivation for several influential score-based diffusion models that relies on only a few textbook results. Diffusion models have recently emerged as powerful tools for generating realistic, synthetic signals -- particularly natural images -- and often play a role in state-of-the-art algorithms for inverse problems in image processing. While these algorithms are often surprisingly simple, the theory behind them is not, and multiple complex theoretical justifications exist in the literature. Here, we provide a simple and largely self-contained theoretical justification for score-based diffusion models that is targeted towards the signal processing community. This approach leads to generic algorithmic templates for training and generating samples with diffusion models. We show that several influential diffusion models correspond to particular choices within these templates and demonstrate that alternative, more straightforward algorithmic choices can provide comparable results. This approach has the added benefit of enabling conditional sampling without any likelihood approximation.

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。

↑