Statically bounded-memory delayed sampling for probabilistic streams

Statically bounded-memory delayed sampling for probabilistic streams
复制标题

概率流的静态有界内存延迟采样

DOI:
10.1145/3485492
复制
发表时间:
2021
影响因子:
--
通讯作者:
Carbin, Michael
Carbin, Michael
中科院分区:
--
文献类型:
--
作者:
Atkinson, Eric;Baudart, Guillaume;Mandel, Louis;Yuan, Charles;Carbin, Michael

文献摘要

参考文献

被引文献

相似文献

概率编程语言帮助开发人员执行贝叶斯推理。这些语言为概率建模和自动推理提供了编程结构和工具。之前的工作引入了概率编程语言ProbZelus,将概率编程功能扩展到无界数据流。这项工作表明,延迟采样推理算法可以扩展到工作在流上下文中。ProbZelus表明,虽然延迟采样可以有效地部署在一些程序,根据考虑的概率模型,延迟采样是不能保证使用有限的内存量在程序的执行过程中,在本文中,我们目前的条件下,概率程序的执行延迟采样将在有限的内存中执行。这两个条件是延迟采样核心运算的两个基本性质:它们被消耗的性质和不可分离的路径性质。当且仅当一个程序满足它们消耗和不可分离路径的性质时,它才能在延迟采样下在有界内存中执行。我们提出了一个静态分析,抽象这些属性,以确保任何程序,通过分析满足这些属性,从而在有限的内存延迟采样下执行。
Probabilistic programming languagesaid developers performing Bayesian inference. These languages provide programming constructs and tools for probabilistic modeling and automated inference. Prior work introduced a probabilistic programming language, ProbZelus, to extend probabilistic programming functionality to unbounded streams of data. This work demonstrated that thedelayed samplinginference algorithm could be extended to work in a streaming context. ProbZelus showed that while delayed sampling could be effectively deployed on some programs, depending on the probabilistic model under consideration, delayed sampling is not guaranteed to use a bounded amount of memory over the course of the execution of the program.In this paper, we the present conditions on a probabilistic program’s execution under which delayed sampling will execute in bounded memory. The two conditions are dataflow properties of the core operations of delayed sampling: them-consumed propertyand theunseparated paths property. A program executes in bounded memory under delayed sampling if, and only if, it satisfies them-consumed and unseparated paths properties. We propose a static analysis that abstracts over these properties to soundly ensure that any program that passes the analysis satisfies these properties, and thus executes in bounded memory under delayed sampling.
DOI: --
发表时间: 1997
期刊: International Conference on Concurrency Theory
影响因子: --
作者:
Vineet Gupta;R. Jagadeesan;V. Saraswat
通讯作者: V. Saraswat
DOI: 10.1145/2737924.2737982
发表时间: 2015
期刊: Proceedings of the 36th ACM SIGPLAN Conference on Programming Language Design and Implementation
影响因子: --
作者:
A. Nori;Sherjil Ozair;S. Rajamani;Deepak Vijaykeerthy
通讯作者: Deepak Vijaykeerthy
概率编程语言 Anglican 中的延迟采样
DOI: --
发表时间: 2017
期刊:
影响因子: --
作者:
Daniel Lundén
通讯作者: Daniel Lundén
Hakaru 中程序转换的概率推理(系统描述)
DOI: 10.1007/978-3-319-29604-3_5
发表时间: 2016
期刊: IEEE INFOCOM 2009
影响因子: --
作者:
P. Narayanan;J. Carette;Wren Romano;Chung;R. Zinkov
通讯作者: R. Zinkov
DOI: --
发表时间: 2017-01
期刊: ArXiv
影响因子: --
作者:
Dustin Tran;M. Hoffman;R. Saurous;E. Brevdo;K. Murphy;D. Blei
通讯作者: Dustin Tran;M. Hoffman;R. Saurous;E. Brevdo;K. Murphy;D. Blei