Workshop on Data-Enabled Science

数学和物理科学(MPS)社区产生了大量的科学数据。主要的实验和设施现在每年产生数PB的数据，这些数据必须在全球范围内分发以进行分析。已经在开发中的项目将以更快的速度产生更大的数据量，接近每周1艾字节，执行分析需要exaflop计算能力。除了越来越多的巨大数据生成器之外，几乎所有的科学都变得数据密集型，规模和/或复杂性不断增加，甚至在个别实验室的PI水平上也是如此。这一趋势超越了MPS学科，扩展到：生物数据；金融、商业和零售数据；视听数据；数据同化和数据融合；以及人文和社会科学数据。如果要实现科学进步，几乎所有学科都需要潜在的激进的新的数学和统计方法来处理未来的数据集。拟议的MPS数据使能科学讲习班将提供对MPS社区需求的高级别评估，包括预期的数据生成、为科学挖掘数据的能力和能力、当前努力的优势和劣势，以及开发新算法和数学方法的工作。讲习班还将评估在今后五年内满足这些需求所需的资源。

The Mathematical and Physical Sciences (MPS) community generates much of the data in science. Major experiments and facilities are now generating petabytes of data per year that must be distributed globally for analysis. Projects already in development will generate much largervolumes at faster rates, approaching an exabyte per week, with exaflop computing capacity needed to perform the analysis. In addition to this growing number of prodigious data generators, virtually all of science is becoming data-intensive, with increasing size and/or complexity, even at the level of PIs in individual labs. This trend extends beyond MPS disciplines to:biological data; financial, commercial, and retail data; audio and visual data; data assimilation and data fusion; and data in the humanities and social sciences. Virtually all disciplines need potentially radical new mathematical and statistical ways to handle future data sets if scientific advances are to be realized. The proposed MPS Workshop on Data-Enabled Science will provide a high-level assessment of the needs of the MPS communities, including anticipated data generation, capability and inability to mine the data for science, strengths and weaknesses of current efforts, and work on developing new algorithms and mathematical approaches. The workshop will also provide an assessment of the resource requirements for addressing these needs over the next five years.