REU Site: Graph Learning and Network Analysis: from Foundations to Applications (GraLNA)

该项目建立了一个新的本科生研究经验（REU）网站，由北卡罗来纳大学格林斯博罗分校计算机科学系主办。每年夏天将有10名学生接受图机器学习和网络分析基础及其在现实网络中的具体应用的研究培训。从物联网、在线社交网络、大脑网络、分子到蛋白质-蛋白质相互作用网络，图和网络在各个科学学科中无处不在。对大规模网络的分析可以大大提高我们对复杂系统的理解。现有的方法纯粹是经验性的或缺乏深入的基础探索，因此在处理复杂的图和网络数据方面受到限制。该项目旨在为学生提供在主要研究和少数民族服务机构进行图和网络前沿研究的机会。图学习和网络分析的研究培训将有助于培养具有竞争力的下一代网络和人工智能劳动力。通过各种活动，如迎新工作坊、特邀讲座、实践项目、演讲、演示和其他专业发展机会，本科生也将提高他们的专业技能。GraLNA项目的第一个目标是为不同群体的学生提供扎实的研究经验，包括STEM中代表性不足的少数民族，特别是来自主要本科院校的学生。学生将提高研究技能以及口头和书面沟通技巧的熟练程度。第二个目标是推进对图学习和优化的理论理解，并开发处理图和网络数据中不同类型复杂性的新方法。值得注意的复杂性类型包括许多现实世界图数据的分布式性质，由图和网络中编码的敏感关系和交互引起的隐私问题，以及涉及丰富领域知识和监管约束的专门网络数据。学生参与者将参与以分布式图分析、联邦学习、优化、私有图分析、网络安全以及结构和功能脑网络分析为中心的研究项目。该项目的第三个目标是通过一系列专业发展活动为学生参与者和教师导师提供专业培训和成长，并分别为初级教师和博士生提供指导和共同指导经验。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This project establishes a new Research Experiences for Undergraduates (REU) Site hosted by the Department of Computer Science at UNC Greensboro. Ten students will receive research training each summer in the foundations of graph machine learning and network analysis and their concrete applications in real-life networks. Graphs and networks have become ubiquitous in various scientific disciplines ranging from the Internet of Things, online social networks, brain networks, and molecules to protein-protein interaction networks. Analysis of large-scale networks can bring significant advances to our understanding of complex systems. Existing methods are purely empirical or lack in-depth foundational exploration, thus limited in processing complex graph and network data. This project aims to provide students the opportunity to undertake cutting-edge research in graphs and networks at a major research and minority serving institute. The research training on Graph Learning and Network Analysis will contribute to developing a competitive next-generation network and AI workforce. Through various activities such as orientation workshops, invited lectures, hands-on projects, presentations, demos, and other professional development opportunities, undergraduate students will also enhance their professional skills.The first objective of this GraLNA project is to provide an experience of doing solid research for a diverse group of students, including underrepresented minorities in STEM, especially those from Primarily Undergraduate Institutions. Students will gain an increased proficiency in research skills as well as oral and written communication skills. The second objective is to advance the theoretical understanding of graph learning and optimization, and to also develop new approaches to handling diverse types of complexities in graph and network data. Notable types of complexities include the distributed nature of many realworld graph data, privacy concerns arising from sensitive relationships and interactions encoded in graphs and networks, and specialized network data that involves rich domain knowledge and regulatory constraints. Student participants will engage with research projects centered around distributed graph analysis, federated learning, optimization, private graph analysis, network security, and structural and functional brain network analysis. The third objective of this project is to provide for both student participants and faculty mentors professional training and growth through a series of professional development activities, and also provide junior faculty and Ph.D. students mentoring and co-advising experience respectively.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.