REU Site: Interdisciplinary Integration in Statistical Learning and Data Mining

这个由美国国家科学基金会支持的REU网站将在2017-2019年的每个夏天为北卡罗来纳大学威尔明顿分校（UNCW）的本科生提供统计学习和数据挖掘方面的跨学科研究经验，并应用于计算机视觉和模式识别。UNCW通过应用学习致力于本科教育和研究，非常适合为参与者提供完整的REU体验。这个项目的动机是缺乏具有分析技能的数据科学家，最近学生对数据科学的兴趣激增，以及学术界、工业界和政府对数据科学职业选择的认识。该计划旨在通过应用学习让学生参与本科研究经验，并提供发展定量和批判性思维技能的机会，以及提高与其他学科专业人员有效沟通的机会。该课程的智力重点是介绍当代统计学习理论和数据挖掘技术，并应用于分析人类面部特征。学生将学习计算机视觉和模式识别的重要影响、人类图像分析的挑战、数学和统计学基础的回顾、图像预处理、当代统计学习理论和数据挖掘技术。将详细讨论以下主题：数据清洗和可视化、降维、回归和分类、软件工程、高性能计算等。研究项目是这些技术的应用，强调跨学科的实际应用。问题的价值、背景、建模假设、统计理论、数值解、计算技术的可视化及其解释将在参与者中得到很好的阐述。在导师的适当干预下，通过以团队为基础的协作和合作、小组会议和每周介绍和报告的反馈，鼓励批判性和反思性思维。

This NSF supported REU Site will provide undergraduate students interdisciplinary research experience in statistical learning and data mining with applications in computer vision and pattern recognition at the University of North Carolina Wilmington (UNCW), for ten weeks during each summer of 2017-2019. UNCW has an institutional commitment to undergraduate education and research through applied learning, and is ideally suited to provide a complete REU experience for the participants. This project is motivated by the shortage of data scientists with analytical skills, recent surge of interests from students, and bringing in awareness of data science career options in academics, industry, and government. The program is designed to involve students in undergraduate research experiences through applied learning and to provide opportunities to develop quantitative and critical thinking skills, and opportunities to improve effective communication skills with professionals from other disciplines. The intellectual focus of the program is to introduce contemporary statistical learning theory and data mining techniques, with applications in analyzing human facial features. Students will be given lectures on the significant impact of computer vision and pattern recognition, challenges in human image analysis, review of fundamentals in mathematics and statistics, image preprocessing, and contemporary statistical learning theory and data mining techniques. The following topics will be discussed in detail: data cleaning and visualization, dimension reduction, regression and classification, software engineering, high performance computing, etc. Research projects are applications of these techniques and emphasize on real world application with interdisciplinary integration. The values of the problem, background, modeling assumptions, statistical theory, numerical solutions, and visualization with computational technology and its interpretation will be well articulated among the participants. Critical and reflective thinking are encouraged, under proper intervenes by mentors, through team-based collaboration and cooperation, group meetings and feedback from weekly presentation and reports.