研究人员发布了关于组合式基于仿真的推理中退火Langevin动力学的理论指导，旨在通过提供超参数的明确决策规则来提高采样精度。另一篇论文通过大偏差理论为加速Langevin蒙特卡洛采样变体提供了一种统一的研究方法。第三项研究分析了预处理退火Langevin动力学（特别是针对多模态高斯混合模型）的维度均匀离散化，并展示了不同的离散化方案如何影响稳定性和准确性。 AI
影响 这些论文推进了对对训练和评估AI模型至关重要的采样方法的理论理解。
排序理由 该集群包含多篇详细介绍统计和机器学习方法理论进展和分析的学术论文。
AI 生成摘要 · Google Gemini · 来自 5 个来源。 我们如何撰写摘要 →
arXiv:2605.21253v1 Announce Type: new Abstract: Compositional score-based approaches to simulation-based inference (SBI) approximate the posterior over a shared parameter given $n$ independent observations by aggregating individually learned posterior scores: currently, there are…
arXiv:2503.19066v2 Announce Type: replace-cross Abstract: Langevin algorithms are popular Markov chain Monte Carlo methods that are often used to solve high-dimensional large-scale sampling problems in machine learning. The most classical Langevin Monte Carlo algorithm is based o…
Compositional score-based approaches to simulation-based inference (SBI) approximate the posterior over a shared parameter given $n$ independent observations by aggregating individually learned posterior scores: currently, there are two main propositions of such methods (Geffner …
arXiv:2605.16473v1 Announce Type: new Abstract: Obtaining stable diffusion-based samplers in high- and infinite-dimensional settings is challenging because errors can accumulate across high-frequency coordinates and make the dynamics unstable under refinement of the finite-dimens…
Obtaining stable diffusion-based samplers in high- and infinite-dimensional settings is challenging because errors can accumulate across high-frequency coordinates and make the dynamics unstable under refinement of the finite-dimensional approximation of the underlying function-s…