Probability Theory and Statistics in High and Infinite Dimensions: Empirical Processes Theory and Beyond

高维和无限维中的概率论和统计：经验过程理论及其他

• 批准号: 1407649
• 负责人: Vladimir Koltchinskii
• 金额: $ 2.3万
• 依托单位: Georgia Tech Research Corporation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2014
• 资助国家: 美国
• 起止时间: 2014-05-01 至 2015-04-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1407649&HistoricalAwards=false
• 关键词: Probability; Theory; Statistics; Infinite; Dimensions

The conference on "Probability Theory and Statistics in High and Infinite Dimensions: Empirical Processes Theory and Beyond" will focus on empirical processes and other methods of high-dimensional probability and their role in a variety of problems related to statistical inference for high-dimensional data as well as to nonparametric inference. The conference will be held on June 23-25, 2014 at the Center for Mathematical Sciences, University of Cambridge, UK. The list of topics includes concentration inequalities, exponential and moment bounds for Gaussian, empirical and other related processes, nonasymptotic bounds for random matrices, complexity penalization methods and oracle inequalities in high-dimensional inference, sparse recovery and low rank matrix recovery, high-dimensional problems in machine learning, nonparametric estimation and hypotheses testing, Bayesian nonparametrics.High-dimensional probability and statistics have had impact far beyond mathematical sciences. The methods of high-dimensional probability have been crucial in understanding the performance of new machine learning algorithms and in the design of contemporary methods of analysis of big data. Bringing together researchers who have made and are currenly making important contributions to this area would facilitate the exchange of ideas leading to much better understanding of the role of high-dimensional phenomena in statistical inference and machine learning. The conference would be especially beneficial for junior participants entering these research areas. This award will support the participation of approximately twenty US-based participants in the conference.Conference web site: www.statslab.cam.ac.uk/~nickl/Site/2014.html

关于“概率论和统计在高和无限维：经验过程理论和超越”的会议将集中在经验过程和高维概率的其他方法，以及它们在各种问题中的作用，有关统计推断高维数据以及非参数推断。会议将于2014年6月23日至25日在英国剑桥大学数学科学中心举行。主题列表包括浓度不等式，高斯，经验和其他相关过程的指数和矩界，随机矩阵的非渐近界，高维推理中的复杂性惩罚方法和oracle不等式，稀疏恢复和低秩矩阵恢复，机器学习中的高维问题，非参数估计和假设检验，贝叶斯非参数化。高维概率和统计学的影响远远超出了数学科学。高维概率方法在理解新机器学习算法的性能和设计当代大数据分析方法方面至关重要。将已经和正在为这一领域做出重要贡献的研究人员聚集在一起，将有助于交流思想，从而更好地理解高维现象在统计推断和机器学习中的作用。会议将特别有利于进入这些研究领域的初级参与者。 该奖项将支持大约20名美国与会者参加会议。会议网站：www.statslab.cam.ac.uk/~nickl/Site/2014.html