Early-Career and Student Support for the XX Householder Symposium

该项目将支持来自美国大学的早期职业科学家和博士生参加将于2017年6月18日至23日在弗吉尼亚州布莱克斯堡的弗吉尼亚理工大学酒店举行的“数字线性代数家庭研讨会XX”，网址为http://www.math.vt.edu/HHXX/。Householder研讨会是数值线性代数领域的首要会议，仅限申请和邀请参加。研讨会非常强调支持早期职业科学家以及这些科学家与该领域领先专家的融合。年轻研究人员的支持将增强美国在这一关键学科的持续领导地位，并将其应用于极大地造福社会的技术。数值线性代数的进步和应用是处理大型数据集的基础，其应用跨越网络搜索、医学成像、科学模拟、所有高性能计算应用以及材料和结构设计。谷歌的PageRank算法是两个数值线性代数领域，Perron-Frobenius理论/Markov链和特征值求解方法的智能组合。同样，从当今庞大的数据集（无论是来自企业、政府还是新型医疗成像设备中的传感器）中获取关键信息，也依赖于对核心数值线性代数方法奇异值分解（singular value decomposition）的重大创新。最后，确保在下一代超级计算机上进行有效的模拟，以评估国家核武器的持续安全性，需要对线性求解器（另一种普遍存在的数值线性代数方法）进行重大改变。参加家庭研讨会，其中包括讨论热门的新研究课题和新兴领域，将使年轻的研究人员接触到具有重要社会意义的问题。

This project will support the participation of early career scientists and PhD students from US universities in the "Householder Symposium XX on Numerical Linear Algebra," to be held at The Inn at Virginia Tech, June 18-23, 2017, in Blacksburg, Virginia, http://www.math.vt.edu/HHXX/. The Householder symposium is the premier meeting in the field of numerical linear algebra and participation in this symposium is by application and invitation only. The symposium places a strong emphasis on supporting early-career scientists and the intermingling of such scientists with the leading experts in the field. The support of young researchers will enhance the continued leadership of the US in this crucial discipline and its application to technologies that greatly benefit society. Advances and applications of numerical linear algebra are fundamental for handling large data sets with applications crossing web searches, medical imaging, scientific simulations, all high performance computing applications, and the design of both materials and structures.Google's PageRank algorithm is a smart combination of methods from two numerical linear algebra fields, Perron-Frobenius theory/Markov chains and eigenvalue solvers. Similarly, deriving crucial information from today's giant data sets, which dwarf anything encountered before, whether from business, government, or the sensors in new medical imaging equipment, relies on substantial innovations to a core numerical linear algebra method, the singular value decomposition. Finally, ensuring efficient simulations on the next generation of supercomputers to assess the continued safety of the nation's nuclear weapons requires drastic changes to linear solvers, another ubiquitous numerical linear algebra method. Participation in the Householder Symposium, which includes discussions of hot new research topics and emerging areas will allow young researchers to be exposed to problems of significant societal importance.