Early Career Support for the Householder XVII Symposium

为美国早期职业科学家和博士生参加第十七届数值线性代数家庭研讨会提供资金支持。研讨会将于2008年6月1日至6日在德国苏黎世举行。矩阵计算是数值线性代数中的一个核心主题，在科学计算中起着核心作用，是许多偏微分方程和优化数值过程的核心。最近，矩阵方法在计算生物学和许多信息科学应用中找到了自己的方式。住户专题讨论会汇集了来自世界各地的学者，无论老少，进行为期五天的非正式研究交流，为这一领域的活力做出了贡献。多核技术对可扩展的方法提出了更高的要求，因此毫不奇怪，模型约简、低秩张量近似和数据压缩将成为研讨会上反复出现的主题。在某些应用中，矩阵维度是如此之大，以至于需要全新的方法。基于采样和数据结构的矩阵分解，这些结构与多级存储系统相匹配。Householder XVII的与会者将带着对“大n”矩阵计算的强烈感觉离开会议。研讨会之所以成为年轻与会者“启动”职业生涯的特别有效的工具，有几个原因。重点是正在进行的研究和邀请过程，积极支持崭露头角的研究人员。通过具体的“前瞻性思维”、开放的麦克风、小组讨论，有意建立一种公共意识，让人们了解该领域的发展方向。由于它是该领域的顶级会议，研讨会吸引了来自世界各地的顶尖高级研究人员；这是一个“必须参加”的活动。由于会议规模不大（约150人），每位与会者都会发表演讲，因此资深研究人员和新博士的参与是有保证的。使美国的早期职业科学家能够参加数字线性代数家庭研讨会，确保了这个对美国科学和工程至关重要的领域的健康。与扩大参与计算科学的目标一致。

Financial support is provided for the participation of USA-based early career scientists and PhD students at the Householder Symposium XVII on Numerical Linear Algebra. The Symposium is to be held in June 1-6, 2008 in Zeuthen, Germany. Matrix computations, a central topic in numerical linear algebra, plays a central role in scientific computing and is at the heart of many numerical procedures for partial differential equations and optimization. More recently, matrix methods have found their way in computational biology and a host of information science applications. The Householder symposia contribute to the vibrancy of this field by bringing together scholars from all over the world, both young and old, for five days of informal research exchanges. Multicore technologies have placed a premium on methods that scale and so it is no surprise that model reduction, low rank tensor approximation, and data compression, will be recurring themes at the Symposium. Matrix dimensions are so large in certain applications that radically new approaches are required,e.g., matrix factorizations that are based on sampling and data structuresthat match multilevel memory systems. Attendees of Householder XVII will exit the meeting with a sharpened sense of ``large n'' matrix computations. There are several reasons why the Symposium is a particularly effective vehicle for ``jump starting'' the careers of its young attendees. There is an emphasis on ongoing research and an invitation process that proactively favors budding researchers. There is a deliberate attempt to build a communal sense of where the field is headed through specific ``forward thinking'', open microphone, panel discussions. Because it is the premier conference in the area, the Symposium attracts the top senior researchers from around the world; it is a ``must attend'' event. The intermingling of established researchers with the fresh PhDs is guaranteed because the conference is modest in size (about 150) and everyone who attends gives a talk. Enabling USA-based early career scientists to attend the Householder Symposium on Numerical Linear Algebra ensures the well being of a field that is critical to American Science and Engineering. Consistent with the goal of broadening participation in computational science.