Parallel Computation and Learning

并行计算与学习

• 批准号: 8902500
• 负责人: Leslie Valiant
• 金额: $ 34万
• 依托单位: Harvard University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1989
• 资助国家: 美国
• 起止时间: 1989-08-01 至 1992-12-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=8902500&HistoricalAwards=false
• 关键词: Parallel; Computation; Learning

The project is concerned with developing and applying ideas from complexity theory to three distinct but related areas of computation: Under continued investigation are the theoretical possibilities for general purpose parallel computers. The problem is to devise idealised high level models that are both convenient for the programmer as well as efficiently simulated in realistic technologies. As a distinct topic in parallel computation the computational capabilities of brainlike models of computation are being studied. Such neural models appear to become more distinct from other complexity models if biologically plausible restrictions on time and space resources are imposed. Thirdly, partly in support of this latter work, the PI will continue to investigate the limits of what can be learned or induced, as imposed by such criteria as polynomial time computability.

该项目涉及将复杂性理论的思想开发和应用到三个不同但相关的计算领域：正在继续研究通用并行计算机的理论可能性。 问题是设计理想化的高级模型，既方便程序员又可以在现实技术中进行有效模拟。 作为并行计算中的一个独特主题，类脑计算模型的计算能力正在被研究。 如果对时间和空间资源施加生物学上合理的限制，这样的神经模型似乎与其他复杂模型更加不同。 第三，部分为了支持后一项工作，PI 将继续研究可以学习或归纳的限制，正如多项式时间可计算性等标准所强加的那样。