SHF: Small: Statistical Analysis of Software

SHF：小型：软件统计分析

• 批准号: 0915766
• 负责人: Alexander Aiken
• 金额: $ 50万
• 依托单位: Stanford University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2009
• 资助国家: 美国
• 起止时间: 2009-08-01 至 2013-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0915766&HistoricalAwards=false
• 关键词: SHF; Small; Statistical; Analysis; Software

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).The project investigates statistical software analysis, which infersrelationships among program components by using statistical propertiesderived from multiple program executions.To motivate statistical techniques, it is useful to draw analogies tostatic analysis methods. Static analysis is about inferringdependencies between program components: If a value is changed in onecomponent, how does that affect a value in a different component?Static analysis tends to work best for properties that are local,meaning the pieces of the program we are trying to relate are notseparated by a great deal of other computation. The statistical analogof dependencies is correlation. Instead of proving definitively viastatic reasoning the presence or absence of dependencies, we canobserve at run-time that some properties of two components have highor low correlation. Importantly, correlation is not affected bysyntactic or even dynamic locality: if two components have acorrelation, regardless of how much time or computation passes betweenthe execution of one component and the execution of the other, thiscorrelation can be detected if the appropriate statistical question isasked.The initial focus is on using cross-correlation, which which computesthe maximum correlation between two sequences of observations, toformalize statistical correlation between software components thathave a direction in time. This idea gives rise to a natural graph thatcaptures the strength and direction of statistical influence onecomponent has upon another; these graphs are analogous to traditionaldependency graphs, but have unique and useful properties.

该奖项是根据2009年的《美国回收与再投资法》（公共法第111-5号）资助的。该项目调查了统计软件分析，该统计软件分析通过使用从多个程序执行中衍生的统计属性来提高程序组件之间的互相关系。为了激励统计技术，它对绘制类似物分析方法很有用。静态分析是关于程序组件之间的推断依赖性：如果在Oneconent中更改了值，则该值如何影响不同组件中的值？静态分析倾向于最适合本地属性的属性，这意味着我们试图关联的程序的段落被大量的其他计算所指出。统计类似物的依赖性是相关性。我们没有通过静态推理确定证明存在或不存在依赖性，而是在运行时划定了两个组件的某些属性具有较低的相关性。重要的是，相关性不会受到宪法甚至动态位置的影响：如果两个组成部分具有超相关性，无论是在一个组件执行的时间或计算中，无论是在一个组件执行的时间或计算是否通过，并且可以检测到此相关性，那么是否可以检测到适当的统计问题。是否均在均等的统计范围上。软件组件之间的统计相关性，这些组件已在时间上的方向。这个想法产生了一个自然图，可以遵守统计影响的强度和方向对另一个组成部分的强度和方向。这些图类似于传统依赖性图，但具有独特且有用的属性。