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 2605.07950cs.LG
收藏

缓慢退火 Langevin 动力学:理论及其在无训练引导生成中的应用

Slowly Annealed Langevin Dynamics: Theory and Applications to Training-Free Guided Generation

Nanyang Technological University(南洋理工大学);City University of Hong Kong(香港城市大学)

Atsushi Nitanda, Dake Bu, Yueming Lyu, Tanya Veeravalli

AI总结:

本文提出缓慢退火 Langevin 动力学(SALD)理论,用于跟踪移动的目标分布路径并通过时间减缓近似终端目标,引入了Velocity-Aware SALD(VA-SALD)以提升无训练引导生成的性能,提供了收敛保证。

摘要:

我们研究了缓慢退火 Langevin 动力学(SALD),这是一种用于跟踪移动目标分布路径并通过时间减缓近似终端目标的采样器。我们通过KL微分不等式建立了非渐近收敛保证,证明了减缓通过中间目标的收缩和路径的复杂性来改进跟踪。受无训练引导生成与预训练分数生成模型的启发,我们进一步引入了Velocity-Aware SALD(VA-SALD),它显式地结合了预训练模型的底层边际分布,并利用减缓来校正引导引起的额外偏差。这为扩散型及相关生成模型家族的无训练引导生成提供了系统框架,并提供了收敛保证,澄清了中间函数不等式和引导偏差的作用。代码可在https://github.com/anitan0925/sald获得。

英文摘要:

We study Slowly Annealed Langevin Dynamics (SALD), a sampler for tracking a path of moving target distributions and approximating the terminal target through time slowdown. We establish non-asymptotic convergence guarantees via a KL differential inequality, showing that slowdown improves tracking through contraction of intermediate targets and the complexity of the path. Motivated by training-free guided generation with pretrained score-based generative models, we further introduce Velocity-Aware SALD (VA-SALD), which explicitly incorporates the underlying marginal distributions of the pretrained model and uses slowdown to correct the additional deviation induced by guidance. This yields a principled framework for training-free guided generation for diffusion-based and related generative model families, together with convergence guarantees that clarify the roles of intermediate functional inequalities and guidance bias. Code is available at https://github.com/anitan0925/sald.

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。

↑