REU Site: Scientific Computing for Structure in Big or Complex Datasets

这个项目为我们的下一代提供了训练有素的研究科学家。该项目通过训练年轻科学家在理论和实践中使用计算来发现、理解和利用大型或复杂数据集的结构，从而预测科学的未来方向。大型和复杂的数据集在科学、商业和人文学科的许多领域无处不在。该项目旨在通过在研究项目比REU少的机构寻找和鼓励有才华的学生来扩大科学研究的参与。除了鼓励从事科学事业外，该项目还旨在进一步使每位参与者与科学和科学研究进行终身对话。培训、指导、演讲和同伴参与都是该项目解决的关键要素，以鼓励我们有才华的本科生成为研究科学家。该项目为本科生提供全额奖学金，让他们在为期十周的暑期课程中与导师合作，从网站的项目组合中选择一个项目。该项目从一个科学训练营开始。在整个经历中，参与者参加系列讲座，并与他们的REU参与者群体互动，进一步促进职业科学家的成长。在结束时，参与者向现场的同行和研究界展示他们的研究，其中一些项目被选中在全国会议上展示。指导项目的组合共享计算的主题，以发现结构的大型或复杂的数据集。项目的特征计算要么用于数据的探索或操作，比如用于可视化或高性能GPU计算，要么用于计算的理论框架，比如用于理解大脑过程、深度学习和神经网络。由于计算科学的跨领域性质，这些项目来自不同的领域，导师来自不同的部门，包括计算机科学、化学和心理学、计算科学中心和医学院。

This project provides for our future generation of effectively trained research scientists. The project anticipates future directions in science by training young scientists in the theoretical and practical use of computation to find, understand, and exploit structure in large or complex data sets. Large and complex data sets are ubiquitous in many areas of science, business and humanities. The project aims to broaden participation in scientific research by locating and encouraging talented students at institutions with fewer research offerings than has the REU site. Along with encouraging a career in science, the project aims to further engage each participant in a lifelong dialog with science and scientific research.Training, mentorship, presentation, and peer engagement are all key elements addressed by this project to encourage our talented undergraduates to become research scientists. The project provides full scholarships to undergraduates to work with mentors in a ten week summer session on a project from the site's portfolio of projects. The program begins with a science boot camp. Throughout the experience the participants attend lecture series and engage with their cohort of REU participants, further facilitating growth as career scientists. At the conclusion, the participants present their research to their peers and the research community at the site, with some projects selected to be presented at a national conference. The portfolio of mentored projects share the theme of computation to find structure in large or complex data sets. Projects feature computation either for the exploration or manipulation of data, such as for visualization or high performance GPU computation, or for the theoretical framework of computation, such as for understanding brain processes, deep learning, and neural nets. Because of the cross-cutting nature of computational science, the projects come from diverse fields with mentors from diverse departments, including Computer Science, Chemistry and Psychology, the Center for Computational Sciences, and the medical school.