CAREER: Scalable Bayesian learning for multi-source and multi-aspect data

职业：针对多源和多方面数据的可扩展贝叶斯学习

• 批准号: 1054903
• 负责人: Yuan Qi
• 金额: $ 51.18万
• 依托单位: Purdue University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2011
• 资助国家: 美国
• 起止时间: 2011-01-01 至 2016-12-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1054903&HistoricalAwards=false
• 关键词: CAREER; Scalable; Bayesian; learning; multi

Data of growing complexity come from multiple sources with multiple aspects. These data present us with unprecedented opportunities to integrate them with predictive models to extract complex relationships among natural and man-made objects.The PI brings together models and techniques encountered in various areas, such as Bayesian statistics, computational science, and systems biology, to develop new methodologies and tools for multi-source and multi-aspect data analysis. The intellectual merit includes (i) new constrained sparse Bayesian models to make interpretable predictions in their application domains, (ii) nonparametric multi-view and multi-way models to reveal unknown complex relationships between different data sources and aspects, and (iii) scalable inference to make advanced Bayesian methods practical data analysis tools.The PI collaborates with domain experts to model online user behavior, facilitate neurologists to elucidate brain functions and help pharmaceutical researchers identify key biomarkers for drug discovery. The PI incorporates the research results into new courses he teaches, organizes workshops, and recruits graduate and undergraduate students to conduct research for this project.For further information see the project web site at the URL: http://www.cs.purdue.edu/~alanqi/projects/learning-multi-source-aspect-data

日益复杂的数据来自多个方面的多个来源。这些数据为我们提供了前所未有的机会，将它们与预测模型相结合，以提取自然和人造物体之间的复杂关系。PI汇集了贝叶斯统计、计算科学和系统生物学等各个领域遇到的模型和技术，以开发用于多源和多方面数据分析的新方法和工具。智力价值包括（i）新的约束稀疏贝叶斯模型，在其应用领域中做出可解释的预测，（ii）非参数多视图和多路模型，以揭示不同数据源和方面之间未知的复杂关系，以及（iii）可扩展推理，使先进的贝叶斯方法成为实用的数据分析工具。PI与领域专家合作，对在线用户行为进行建模，促进神经学家阐明大脑功能并帮助 制药研究人员确定药物发现的关键生物标志物。 PI 将研究成果融入到他教授的新课程中，组织研讨会，并招募研究生和本科生为该项目进行研究。有关更多信息，请参阅项目网站，网址为：http://www.cs.purdue.edu/~alanqi/projects/learning-multi-source-aspect-data