arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2010.05263cs.LGstat.ML
收藏

流形上Langevin动力学的快速收敛:测地线满足Log-Sobolev

Fast Convergence of Langevin Dynamics on Manifold: Geodesics meet Log-Sobolev

SUTD(新加坡科技设计大学);Princeton(普林斯顿大学);UC Irvine(加州大学欧文分校)

Xiao Wang, Qi Lei, Ioannis Panageas

摘要:

采样是一个基本的,可以说是非常重要的任务,在机器学习中有许多应用。一种从高维分布$e^{-f}$中对某个函数$f$进行采样的方法是朗之万算法(LA)。最近,即使在$f$是非凸的情况下,LA的快速收敛也取得了很多进展,特别是[53],[39],其中前一篇论文关注$\mathbb{R}^n$中定义的函数$f$,后一篇论文关注具有流形结构的对称性(如矩阵完成型目标)的函数。我们的工作推广了[53]的结果,其中$f$定义在流形$M$上而不是$\mathbb{R}^n$上。从技术的角度来看,我们表明,KL减少的几何速率时,分布$e^{-f}$满足一个log-Sobolev不等式$M$。

英文摘要:

Sampling is a fundamental and arguably very important task with numerous applications in Machine Learning. One approach to sample from a high dimensional distribution $e^{-f}$ for some function $f$ is the Langevin Algorithm (LA). Recently, there has been a lot of progress in showing fast convergence of LA even in cases where $f$ is non-convex, notably [53], [39] in which the former paper focuses on functions $f$ defined in $\mathbb{R}^n$ and the latter paper focuses on functions with symmetries (like matrix completion type objectives) with manifold structure. Our work generalizes the results of [53] where $f$ is defined on a manifold $M$ rather than $\mathbb{R}^n$. From technical point of view, we show that KL decreases in a geometric rate whenever the distribution $e^{-f}$ satisfies a log-Sobolev inequality on $M$.

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。

↑