High-performance Computational Bayesian Inference for Astronomy

天文学的高性能计算贝叶斯推理

• 批准号: 0611948
• 负责人: Martin Weinberg
• 金额: $ 77.19万
• 依托单位: University of Massachusetts Amherst
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2006
• 资助国家: 美国
• 起止时间: 2006-08-01 至 2010-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0611948&HistoricalAwards=false
• 关键词: performance; Computational; Bayesian; Inference; Astronomy

Bayesian analysis has the power to combine information efficiently from diverse sources, helping to provide confidence bounds on theoretical models and methods for both model selection and hypothesis testing. This project brings state-of-the-art statistical and computational algorithms to bear on a set of data and theory comparison in astronomy. The wok bears on a variety of problems in other sciences, helping to fill the need for optimized data-mining tools. The system can be controlled and operated remotely and collaboratively. The introduction of a persistence system will permit saving of both data and metadata, making it accessible for future use. The statistical approach make the most efficient use of all the available information and permit the testing of hypotheses directly. Bayesian statistical models provide maximum flexibility in incorporating and using all the information, making a virtual observatory test-bed. Multi-disciplinary training opportunities are provided for graduate students and the work disseminated at both meetings, training sessions and for individual users through downloads.

贝叶斯分析有能力有效地将来自不同来源的联合收割机信息结合起来，有助于为模型选择和假设检验提供理论模型和方法的置信界限。该项目带来了最先进的统计和计算算法，以承担一组天文学数据和理论比较。 这项工作涉及其他科学中的各种问题，有助于满足对优化数据挖掘工具的需求。 该系统可以远程和协作地控制和操作。 采用持久性系统将能够保存数据和元数据，以便将来使用。 统计方法最有效地利用了所有可用的信息，并允许直接检验假设。 贝叶斯统计模型在整合和使用所有信息方面提供了最大的灵活性，从而形成了一个虚拟观测台测试平台。 为研究生提供多学科培训机会，并在会议和培训班上传播工作成果，并通过下载向个人用户传播。