Scalable Monte Carlo Methods for large-scale infererence

用于大规模推理的可扩展蒙特卡罗方法

• 批准号: 1656790
• 金额: --
• 依托单位: University of Oxford
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2015
• 资助国家: 英国
• 起止时间: 2015 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-1656790
• 关键词: Scalable; Monte; Carlo; Methods; large

We are in the era of Big Data and big models. In this context, standard Monte Carlo methods such as Markov Chain Monte Carlo methods do not scale, ie the computational complexity is far too high. The aim of the DPhil is to develop, analyse and apply new scalable Monte Carlo methods, which are particularly suitable for an implementation on large-scale distributed systems.The project will fall within the EPSRC statistics and applied probability research area

我们正处于大数据和大模型的时代。在这种情况下，标准蒙特卡罗方法，如马尔可夫链蒙特卡罗方法不缩放，即计算复杂度太高。本项目的目标是开发、分析和应用新的可扩展蒙特卡罗方法，特别适合在大规模分布式系统上实现。该项目将属于EPSRC统计和应用概率研究领域

项目成果

Large Sample Asymptotics of the Pseudo-Marginal Method

伪边际方法的大样本渐近

• DOI: 10.48550/arxiv.1806.10060
• 发表时间: 2018
• 期刊: arXiv e-prints
• 作者: Schmon Sebastian M.