PIPP Phase 1; PILOT: Predictive Intelligence for Limiting Outbreak Threats

近几十年来，在气候变化、城市化和冲突的推动下，新的传染病在世界各地迅速出现。为应对未来的传染病危机，迫切需要设计新的数据驱动工具，用于大流行监测、预测和缓解。COVID-19大流行表明，这些工具不仅必须纳入我们对新传染病基础科学的理解，还必须纳入社会对这些疾病的反应如何影响其传播。为了解决这一迫切需求，PIPP第一阶段PILOT（限制疫情威胁的预测智能）规划项目将汇集来自广泛相关学科的专业知识，包括公共卫生、临床生物医学、计算机科学、人工智能和社会科学。试点调查人员将通过多个圆桌研讨会与学术界、从业者和决策者社区合作，以确定现有的知识差距，并建立大流行监测、预测和缓解方面的最佳做法。该项目还将通过教育和培训研究生和博士后来吸引下一代流行病学者，重点是传播科学以产生社会影响。学生和研究员将协助调查员小组将研讨会成果提炼成白皮书和政策简报，并通过开放获取的在线知识门户和虚拟市政厅会议与公众分享。此外，为了让市民体验学习，我们将邀请市民参与全球直播的社区传染病危机模拟。该计划第一阶段规划项目的成功实施将导致在该计划第二阶段为未来的大流行病威胁开发和部署新的数据驱动建模管道。PILOT建模管道将新型数字数据源与上述学科的方法相结合，以解决三个相互关联的科学挑战：(1)理解和建模疾病监测的大流行潜力，(2)理解和建模疾病预测干预措施的影响，以及(3)理解和建模接受（和拒绝）疾病缓解的干预措施。在公私合作方案第一阶段规划项目期间，第一个挑战的进展将涉及确定在单一时间点（即在特定情况下病原体出现或重新出现后立即出现）评估特定病原体的大流行潜力时应优先利用哪些信息源和计算方法。同样，第二个挑战的进展将涉及探索现有的知识差距，通过基于智能体和多智能体决策的博弈论模型来模拟干预，特别是在信息有限的条件下（即，可能需要基于模拟的场景分析）。最后，第三个挑战的进展将涉及建立旨在鼓励采取干预措施的社会传染模型的最佳做法（即，侧重于复杂的传染和识别跨社会网络的影响者）。因此，PIPP第一阶段PILOT项目的总体目标将是通过确定哪种数据类型和方法选择最适合开发和部署PILOT建模管道（即，考虑到现有的知识差距和最佳实践）来规划一个中心规模的工作。管道的实施将在今后的中心规模工作中进行。该奖项由流行病预防的跨部门预测情报第一阶段（PIPP）计划提供支持，该计划由生物科学（BIO）、计算机信息科学与工程（CISE）、工程（ENG）和社会、行为和经济科学（SBE）委员会共同资助。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。

In recent decades, new infectious diseases have emerged at a rapid rate around the world—driven by climate change, urbanization, and conflict. To prepare for future infectious disease crises, there is an urgent need to devise new data-driven tools for pandemic surveillance, prediction, and mitigation. The COVID-19 pandemic has demonstrated that such tools must incorporate not only our understanding of the science that underpins new infectious diseases, but also how society’s responses to them can affect their propagation. To address this urgent need, the PIPP Phase I PILOT (Predictive Intelligence for Limiting Outbreak Threats) planning project will bring together the expertise from a wide range of relevant disciplines, including public health, clinical biomedicine, computer science, artificial intelligence, and social science. PILOT Investigators will collaborate with academic, practitioner, and decision-maker communities through multiple roundtable workshops to determine existing knowledge gaps and establish best practices in pandemic surveillance, prediction, and mitigation. This project will also engage the next generation of pandemic scholars by educating and training graduate students and postdoctoral fellows, with an emphasis on communicating science for societal impact. Students and fellows will assist the Investigator team in distilling workshop outcomes into white papers and policy briefs, which will be shared with the public via an open access online knowledge portal and virtual town hall meetings. Moreover, to enable experiential learning, members of the public will be invited to participate in a globally-broadcast, community-wide infectious disease crisis simulation. Success in operationalizing this PIPP Phase I planning project will lead to the development and deployment of a new data-driven modeling pipeline for future pandemic threats during PIPP Phase II.The PILOT modeling pipeline will combine novel digital data sources with methods from the aforementioned disciplines to address three interconnected scientific challenges: (1) understanding and modeling pandemic potential for disease surveillance, (2) understanding and modeling the impact of interventions for disease prediction, and (3) understanding and modeling intervention acceptance (and refusal) for disease mitigation. During the PIPP Phase I planning project, progress towards the first challenge will involve determining which information sources and computational approaches should be preferentially leveraged when assessing a given pathogen’s pandemic potential at a single point in time (i.e., immediately following its emergence or re-emergence in a given context). Likewise, progress towards the second challenge will involve exploring existing knowledge gaps in simulating interventions via agent-based and game-theoretic models of multi-agent decision-making, particularly under conditions with limited information (i.e., wherein simulation-based scenario analyses may be necessary). Finally, progress towards the third challenge will involve establishing best practices for social contagion models that aim to encourage intervention uptake (i.e., with a focus on complex contagion and identification of influencers across social networks). Thus, the overarching goal of the PIPP Phase I PILOT project will be to plan for a center-scale effort by ascertaining which data types and methodological choices are most appropriate for the development and deployment of the PILOT modeling pipeline (i.e., given existing knowledge gaps and best practices). Implementation of the pipeline will be pursued in a future center-scale effort.This award is supported by the cross-directorate Predictive Intelligence for Pandemic Prevention Phase I (PIPP) program, which is jointly funded by the Directorates for Biological Sciences (BIO), Computer Information Science and Engineering (CISE), Engineering (ENG) and Social, Behavioral and Economic Sciences (SBE).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.