NEtwork-based Data-driven Approach (NEDA) for Outbreak Monitoring

The tool provides certified indicators for systematically monitoring the progress of an outbreak in a metapopulation network that can be used for what-if tradeoffs and decision making. We seek to create a situational awareness/decision tool for real-time monitoring of outbreak both spatially and temporarily. 

Professor Milad Siami (Northeastern University)

We seek to create a NEtwork-based Data-driven Approach (NEDA) for outbreak monitoring to enable the early mitigation of a pandemic disaster. A pandemic can come in a single wave or several waves separated by several months or years, and future waves may be more severe than the initial one. If our work succeeds, the proposed tool will save thousands of lives. The project forms an interdisciplinary worldwide research collaboration between Co-PIs, with expertise in control theory, ML, systems biology, and transport planning to perform research and develop tools for timely COVID-19 challenges.

This solution directly addresses solution challenge 2b (Respond: Cut through the "infodemic") through developing an ideal outbreak monitoring system aiding policymakers with a systematic information repository incorporating certified indicators that represent both short-term and long-term outlook for outbreaks. It can save lives and mitigate the negative health impact of many for the current and future outbreaks. We have started to work and apply our tool on different cities like Tehran (Iran), Calgary (Canada), and Boston (USA).   

The tool will support decision-makers and authorities for more informed decision making regarding actions to be taken (lockdown, travel restrictions, national and international travels, remote working orders, aggressive testing, etc) based on rigorous data and certified (is that a good word) indicators of current and future outbreaks. Authorities can take the optimum actions depending on the current and projected level of the certified indicators supported by the historical data by several cities around the world. Ultimately this solution will serve people with saving lives and negative health impacts and anxiety of outbreaks. People can daily check the NEDA score and plan their day accordingly. For example, they can decide to work/shop online or in-person, follow social distancing, or any other non-pharmaceutical interventions. 

• Direct Social and Health Impact:

The proposed research, for the first time, will design a reliable monitoring framework with guaranteed properties for interconnected dynamical networks with time delays (e.g., pandemics), which will directly impact the social and economic efficacy and sustainability of the mitigation plans.

• Recent COVID-related Education Activities:

PI Siami developed “EECE7398 ST: Intro Dist. Intelligence - Spring 2020” where graduate students have had the opportunity to work on COVID-19 related problems as early January 23, and the students performed a data-driven approach to examine the spread of infectious diseases using epidemic networks (cf. selected student projects [2], [3], [4],[5]).The students have become actively involved in this research through class projects.

The solution will aid authorities and people to make more informed decisions regarding outbreak risks and actions required. As the result, implementing this solution will directly impact the number of people affected by the outbreak and the associated cost with respect to the number of people affected both at the local (e.g. school board decision on being open or closed, or parents decision sending their kids to school on week by week basis), municipal (e.g. residents planning for social activity, use of public transportation, and cities taking actions regarding public service delivery), provincial/sate (e.g. authorities decision on lockdowns and restrictions on mobility), and federal(e.g. authorities decision on restriction/ease of international travels, mandatory isolations at home or hotels, etc).

We will begin with designing our solution for one country (national level) and starting with focusing on a single city (local level) during the first year. Then we scale up the solution to implement for at least three regions during the second year. We hope to develop a transformative and universal solution being available for as many countries and regions during the third year. This will still require partnerships and interests from the new regions. 

During the first year, while designing our solution for a certain city and a country for the national level, we will collect the required data for the scale-up of the solution for the other 2 regions. The initial partnership with each country is already developed by the team. Moving to the third year, we will explore further partnerships with other regions. This will be made possible through communicating this solution through website development, videos, presentations at conferences and symposiums, TED Talks, course developments during the first and second year. 

The key measure of success especially during the first year is the exposure of our solution through different means (social media, publications, website visits, news articles and media). The number of confirmed partners (municipalities, schools, universities and research institutes) that are aware of our solution and consider implementing it will be our longer-term measure of success.  

The team has already shown progress through this objective by forming official partnerships with the City of Calgary (Canada), and Tehran Municipality (through a partnership with Sharif University of Technology). Both are directly involved with the solutions being developed by this team and consider the implementation of the solution when developed. 

The major steps toward realizing NEDA stem from the need to address the following barriers:

•  Lack of a framework that allows for real-time learning, control, and sensing in complex dynamical systems.
•  Lack of tools for provably correct inference in “data deluged” scenarios.
•  The pathogen underlying a pandemic can have extremely differential effects that are in some ways unique and can impact the spread of disease. For example, COVID-19: 
  • is having less of an impact on the health of adolescent populations who tend to be risk-seeking and social interaction seeking (due to various developmental factors),
  • can infect people who remain seemingly asymptomatic while others have severe symptoms,
  • can have incubation periods of up to 2 weeks

Addressing these challenges requires an approach that cuts across engineering, computer and social sciences, and psychology. 

We aim to develop and validate this system for several cities around the globe. Currently, we have been successful to obtain all required data and partially develop the model for two cities. However given limited resources, we are not currently able to scale up. The support from this initiative will help us to perform more rigorous model development and initiate more partnerships with other cities and regions.

• University of Calgary
• Northeastern University
• Sharif University of Technology 

This will give us an opportunity to perform a collaborative research between different universities around the globe and support underrepresented graduate students from different countries. Moreover, if our work succeeds, the NEDA tool will save thousands of lives In Tehran, Calgary, and Boston. 

Northeastern University

NU can help/support our team to expedite the process of developing the NEDA tool.