arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 0707.4242stat.COstat.AP
收藏

Importance Tempering(重要性回火)

Importance Tempering

Statistical Laboratory(统计实验室);University of Cambridge(剑桥大学);CREEM;University of St Andrews(圣安德鲁斯大学)

Robert B. Gramacy, Richard J. Samworth, Ruth King

AI总结:

针对重要性回火(IT)方法因实现粗糙导致估计量方差高的问题,提出组合多个重要性采样估计量的最优方法,证明其具备与有效样本量相关的优良属性,并在可逆跳MCMC场景中验证了有效性。

摘要:

模拟回火(Simulated Tempering,ST)是一种成熟的马尔可夫链蒙特卡洛(Markov chain Monte Carlo,MCMC)方法,用于从多峰密度$\pi(\theta)$中采样。通常,ST需要引入一个辅助变量$k$,其取值为$[0,1]$的有限子集,并索引一组回火分布,即$\pi_k(\theta) \propto \pi(\theta)^k$。在这种情况下,较小的$k$值会促进更好的混合,但只有当$(\theta,k)$的联合链达到$k=1$时,才能获得来自$\pi$的样本。不过,只要计算了重要性采样(Importance Sampling,IS)权重,整个链都可用于估计感兴趣函数在$\pi$下的期望。遗憾的是,这种我们称之为重要性回火(Importance Tempering,IT)的方法效果可能不尽如人意。部分原因是最直接的实现方式很粗糙,可能导致估计量方差很高。我们推导了一种组合多个IS估计量的新最优方法,并证明所得估计量具有与有效样本量概念相关的极佳属性。我们简要报告了该最优组合方法在两个需要可逆跳MCMC的建模场景中的成功应用,而在这些场景中,粗糙方法是失效的。

英文摘要:

Simulated tempering (ST) is an established Markov chain Monte Carlo (MCMC) method for sampling from a multimodal density $π(θ)$. Typically, ST involves introducing an auxiliary variable $k$ taking values in a finite subset of $[0,1]$ and indexing a set of tempered distributions, say $π_k(θ) \propto π(θ)^k$. In this case, small values of $k$ encourage better mixing, but samples from $π$ are only obtained when the joint chain for $(θ,k)$ reaches $k=1$. However, the entire chain can be used to estimate expectations under $π$ of functions of interest, provided that importance sampling (IS) weights are calculated. Unfortunately this method, which we call importance tempering (IT), can disappoint. This is partly because the most immediately obvious implementation is naïve and can lead to high variance estimators. We derive a new optimal method for combining multiple IS estimators and prove that the resulting estimator has a highly desirable property related to the notion of effective sample size. We briefly report on the success of the optimal combination in two modelling scenarios requiring reversible-jump MCMC, where the naïve approach fails.

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。

↑