arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1708.07367math.STcs.LGmath.PRstat.MLstat.TH
收藏

单样本路径下可逆马尔可夫链的混合时间估计

Mixing time estimation in reversible Markov chains from a single sample path

Columbia University(哥伦比亚大学);Ben-Gurion University(本-古里安大学);University of Oregon(俄勒冈大学);Microsoft Research(微软研究院);University of Alberta(阿尔伯塔大学)

Daniel Hsu, Aryeh Kontorovich, David A. Levin, Yuval Peres, Csaba Szepesvári

摘要:

有限、遍历、可逆马尔可夫链的谱间隙$γ$是衡量渐近收敛速度的一个重要参数。在应用中,转移矩阵$P$可能是未知的,但可以观察到链的一个样本直到固定时间$n$。我们在这里考虑的问题估计$γ$从这些数据。令$π$为$P$和$π_\star = \min_x π(x)$的平稳分布。我们表明,如果$n = \tilde{O}\bigl(\frac{1}{γπ_\star}\bigr)$,那么$γ$可以估计到乘法常数内的概率很高。当$π$在$d$状态上是均匀的时,这匹配(直到对数校正)精确估计$γ$所需的$\tildeΩ\bigl(\frac{d}γ\bigr)$步骤的下限。此外,我们提供了第一个程序,用于计算一个完全依赖于数据的间隔,从一个单一的有限长度的轨迹链,该链的混合时间$t_{\text{mix}}$的陷阱在规定的置信水平。该区间不需要知道链的任何参数。这与以前的方法形成对比,以前的方法要么只提供点估计,要么需要重置机制或额外的先验知识。该区间是围绕弛豫时间$t_{\text{relax}} = 1/γ$构建的,该弛豫时间与混合时间密切相关,并且区间的宽度以$1/\sqrt{n}$的速率大致收敛到零,其中$n$是样本路径的长度。

英文摘要:

The spectral gap $γ$ of a finite, ergodic, and reversible Markov chain is an important parameter measuring the asymptotic rate of convergence. In applications, the transition matrix $P$ may be unknown, yet one sample of the chain up to a fixed time $n$ may be observed. We consider here the problem of estimating $γ$ from this data. Let $π$ be the stationary distribution of $P$, and $π_\star = \min_x π(x)$. We show that if $n = \tilde{O}\bigl(\frac{1}{γπ_\star}\bigr)$, then $γ$ can be estimated to within multiplicative constants with high probability. When $π$ is uniform on $d$ states, this matches (up to logarithmic correction) a lower bound of $\tildeΩ\bigl(\frac{d}γ\bigr)$ steps required for precise estimation of $γ$. Moreover, we provide the first procedure for computing a fully data-dependent interval, from a single finite-length trajectory of the chain, that traps the mixing time $t_{\text{mix}}$ of the chain at a prescribed confidence level. The interval does not require the knowledge of any parameters of the chain. This stands in contrast to previous approaches, which either only provide point estimates, or require a reset mechanism, or additional prior knowledge. The interval is constructed around the relaxation time $t_{\text{relax}} = 1/γ$, which is strongly related to the mixing time, and the width of the interval converges to zero roughly at a $1/\sqrt{n}$ rate, where $n$ is the length of the sample path.

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。

↑