[论文解读] CovidSens: A Vision on Reliable Social Sensing for COVID-19

Md Tahmid Rashid, Dong Wang|arXiv (Cornell University)|Apr 9, 2020

Data-Driven Disease Surveillance参考文献 99被引用 7

一句话总结

CovidSens 提出了一种由人工智能驱动、集成边缘计算的社会感知框架，利用移动用户产生的实时社交媒体数据，以高时效性和可靠性检测并预测 COVID-19 的传播。通过结合设备端人工智能、联邦学习以及与无人机和车辆等物理传感器的集成，该框架能够在保障隐私的同时实现实时风险预警和态势感知，并解决虚假信息和数据可靠性方面的挑战。

ABSTRACT

With the spiraling pandemic of the Coronavirus Disease 2019 (COVID-19), it has becoming inherently important to disseminate accurate and timely information about the disease. Due to the ubiquity of Internet connectivity and smart devices, social sensing is emerging as a dynamic AI-driven sensing paradigm to extract real-time observations from online users. In this paper, we propose CovidSens, a vision of social sensing based risk alert systems to spontaneously obtain and analyze social data to infer COVID-19 propagation. CovidSens can actively help to keep the general public informed about the COVID-19 spread and identify risk-prone areas. The CovidSens concept is motivated by three observations: 1) people actively share their experience of COVID-19 via online social media, 2) official warning channels and news agencies are relatively slower than people reporting on social media, and 3) online users are frequently equipped with powerful mobile devices that can perform data processing and analytics. We envision unprecedented opportunities to leverage posts generated by ordinary people to build real-time sensing and analytic system for gathering and circulating COVID-19 propagation data. Specifically, the vision of CovidSens attempts to answer the questions: How to distill reliable information on COVID-19 with prevailing rumors and misinformation? How to inform the general public about the state of the spread timely and effectively? How to leverage the computational power on edge devices to construct fully integrated edge-based social sensing platforms? In this vision paper, we discuss the roles of CovidSens and identify potential challenges in developing reliable social sensing based risk alert systems. We envision that approaches originating from multiple disciplines can be effective in addressing the challenges. Finally, we outline a few research directions for future work in CovidSens.

研究动机与目标

• 应对快速蔓延的 COVID-19 大流行期间对及时、准确信息传播的迫切需求。
• 通过利用实时、用户生成的社交媒体内容，克服官方报告延迟，实现疫情暴发的早期检测。
• 开发一种可扩展的、保护隐私的系统，利用边缘设备在本地处理数据，减少对集中式基础设施的依赖。
• 将社会感知与物理感知（例如无人机、车载网络）相结合，验证社会报告并提高数据可靠性。
• 通过人工智能驱动的信息提炼和约束优化技术，减轻社交媒体中的虚假信息和数据不一致问题。

提出的方法

• 利用社交媒体帖子（例如推文）作为实时数据源，推断区域感染趋势和高风险区域。
• 在设备端实施人工智能和边缘计算，执行本地数据处理，降低延迟，保护用户隐私。
• 应用联邦学习（FL）在去中心化设备上训练全局模型，同时处理设备流失和异步更新问题。
• 将社会感知与无人机（UAVs）和车辆传感器网络（VSNs）集成，验证报告并引导物理传感器前往高风险或高活动区域。
• 结合无人机采集的数据，利用计算建模（例如疾病传播模型）提升检测准确性。
• 采用约束优化和估计理论，量化不确定性，提高社交媒体信号的可靠性。

实验结果

研究问题

• RQ1 如何从嘈杂且充斥谣言的社交媒体数据中提取可靠、实时的 COVID-19 信息？
• RQ2 如何利用基于边缘的人工智能和联邦学习，实现可扩展、低延迟且保护隐私的社会感知？
• RQ3 如何有效整合社会感知与物理感知（例如无人机、车辆）以验证并增强数据可靠性？
• RQ4 应采用何种机制，根据社会信号引导自主系统（例如无人机）前往感兴趣区域，同时遵守物理约束？
• RQ5 如何通过跨学科的人工智能与优化技术，减轻社交媒体中的虚假信息和数据不一致问题？

主要发现

• 社交媒体用户报告健康经历和症状的速度快于官方渠道，从而实现疫情暴发的更早检测。
• 边缘设备（如智能手机和物联网传感器）可执行设备端人工智能推理，降低延迟，提升实时数据处理中的隐私保护。
• 联邦学习可在去中心化设备上实现可扩展的模型训练，但设备流失和异步更新问题仍待解决。
• 将社会感知与无人机和车辆传感器网络集成，可通过物理传感器数据验证社会报告，提高数据可靠性。
• 社会信号可引导无人机和车辆前往高风险区域，实现对疫情暴发和人群行为的快速现场验证。
• 计算建模结合无人机观测数据，可提升疾病传播预测的准确性，并支持实时预警系统。