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.
Theoretical guidelines for annealed Langevin dynamics in compositional simulation-based inference
Univ. Grenoble Alpes(格勒诺布尔阿尔卑斯大学);Inria(法国国家信息与自动化技术研究所);CNRS(法国国家科学研究中心);Grenoble INP(格勒诺布尔INP);LJK(实验室);Geostatistics team(地质统计学团队);Centre for geosciences and geoengineering(地球科学与地球工程中心);Mines Paris(巴黎矿校);PSL University(巴黎 sciences et lettres 大学)
University of Cambridge(剑桥大学);Zuse Institute Berlin(柏林图灵研究所);dida Datenschmiede GmbH(dida数据安全公司);NVIDIA(NVIDIA公司);Karlsruhe Institute of Technology(卡尔斯鲁厄理工学院);California Institute of Technology(加州理工学院)