Novel Analytical and Computational Approaches for Fusion and Analysis of Multi-Level and Multi-Scale Networks Data

新冠肺炎夺走了近660万人的生命，让许多运行良好的医疗体系的繁荣国家变得更加虚弱。从这次大流行中吸取的一个重要教训是，在疫情爆发之初，非药物公共卫生干预对于抑制流行曲线至关重要。对正常生活影响最小的温和干预措施仍然能够有效地减少流行病的传播是非常可取的。例如，社会疏远和病例隔离等干预措施是遏制大流行的非常有效的战略。然而，在美国，此类缓解措施依赖于个人的自我报告机制，收集这些机制既耗时又容易出错。目前的项目旨在开发更准确和计算效率更高的统计工具，以在更广泛的战线上提高缓解措施的效率。这个项目为学生提供了多种独特的机会，让他们在统计学和生物监测的界面上参与前沿和跨学科的研究。在这个项目中，研究人员通过分析流动性数据，旨在开发一套分析和计算方法，使之能够及早发现疫情爆发和准确识别感染者。与自我报告机制相比，移动性数据包含非连续的个性化信息，并且可以很容易地从公共领域获得。联系人和移动性数据都可以自然地表示为网络(图)，其中单个节点是位置或人(或一组人)，其边(连接)对应于节点之间的接触或移动性的度量。该项目将开发一系列新颖的统计和机器学习方法，用于重建伪传播时间，识别感染个体，使用大规模流动数据检测与传播路径和感染个体相关的潜在连接，以及对各种干预下网络之间的差异进行假设检验。该项目的结果将适用于广泛的生物监测任务，并将有助于我们整个社会的福祉。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

COVID-19 has claimed nearly 6.6 million lives and made many prosperous nations with well-run healthcare systems weaker. One important lesson learned from this pandemic is that non-pharmaceutical public health interventions are critical to suppress the epidemic curve at the beginning of the epidemic breakout. Mild interventions with minimal impact on normal life that are still capable to effectively reduce the epidemic spread are highly desirable. Such interventions as, for example, social distancing and case isolation are very effective strategies to suppress the pandemic. However, in the U.S., such mitigation measures rely on individuals' self-reporting mechanisms, which are time-consuming to collect and error-prone. The current project aims to develop more accurate and computationally efficient statistical tools to enhance efficiency of mitigation measures at a broader front. This project offers multiple unique opportunities for students to participate in cutting-edge and interdisciplinary research at the interface of statistics and bio-surveillance.In this project, by analyzing mobility data, the investigators aim to develop a suite of analytical and computational approaches that enables the early detection of the epidemic outbreak and accurate identification of infected individuals. Compared to self-reporting mechanisms, mobility data contains non-continuous individualized information and can be easily obtained from the public domain. Both the contact and mobility data can be naturally represented as networks (graphs), where the individual node is a location or a person (or a group of people), and its edges (connections) correspond to measures of contact or mobility between the nodes. The project will develop a series of novel statistical and machine learning methods for reconstructing pseudo-transmission time, identifying the infected individuals, detecting potential connections related to transmission pathways and infectious individuals using large-scale mobility data, as well as hypothesis testing for the differences between networks under various interventions. The results of the project will be applicable to a wide range of bio-surveillance tasks and will contribute to the wellbeing of our society as a whole.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.