Pandemonium by Quantum Risk Analytics

We are a charitable not-for-profit corporation based in New York but with a small international team and are developing a holistic framework for modeling disease spread and risk in future pandemics and epidemics as well as COVID-19. Our framework will empower researchers to test their models and will foster larger-scale collaboration. Our publicly available COVID-19 risk assessment app is built on top of this framework.

The current pandemic revealed three challenges to effective public health responses and individual risk mitigation during a global pandemic: data sharing between governmental, academic, and industrial researchers, dissemination of accurate information from researchers to the general public, and giving the public sufficient information to make effective risk mitigation choices for themselves and their families. 

There have been more than 141 million documented COVID-19 cases worldwide, accompanied by over 3 million deaths. 1 in 11 people in the United States have knowingly had COVID-19 with >500,000 associated deaths. However, Americans are unsure of their risks of acquiring COVID-19 and its consequences, with large numbers stating that COVID-19 is not a significant threat to public health. Studies conclude that education campaigns improve public health outcomes. Also, agency and voluntariness affect public adoption of risk mitigation strategies in COVID-19.   

Our solution will enhance the public’s awareness of their infection risks, promote associated feelings of self-direction and adoption of risk reducing-behaviors, and ultimately improve health outcomes in diverse countries. These benefits will extend to public health authorities, researchers, and responses to not just the current, but also future pandemics.

Pandemonium is a framework that will enable modeling for disease transmission and spread on a completely new level. The framework is a software infrastructure that will serve as a software development kit (SDK) and an application programming interface (API), enabling disease transmission models to be developed and deployed more quickly than ever before. It will also host the most extensive collection of data possible because it allows for crowdsourced data, in addition to data from public health sources. When researchers can easily test the impact that different models would have with such extensive data, they can better advise public health strategies and work faster when a new disease is detected.

Pandemonium will also have some models and an app built in, allowing any user to assess their individual risk of infection, based on information they provide.

Pandemonium combines macro and micro models with probabilistic programming. Probabilistic programming enables the greatest inference to be drawn from limited data. The model is also being designed to be flexible in the spatial and demographic specificity, to simultaneously accommodate different levels of detail in different regions.

Pandemonium serves three stakeholder groups. Individuals and families can access a risk assessment tool that incorporates traditional and novel data for their location and demographics, and accounts for local, regional, and national travel. Public health officials can use Pandemonium to track local metrics and compare them to other regional civic units. Researchers can freely use the advanced code contained in Pandemonium to augment their studies, and potentially add to the Pandemonium core code.

One development team is working with specific test groups to assure the ease of use and completeness of Pandemonium’s web experience. Another team is reaching out to local government health agencies to assure Pandemonium aligns with appropriate standards and local requirements. Finally, our coding team is reaching out to public health and infectious disease researchers to better align the Pandemonium code interface with their needs.

In order to be most prepared for future health emergencies, there is a need for a tool to help both individuals and policymakers make more informed decisions. Pandemonium aims to solve exactly this problem by helping to respond to future health emergencies through decreasing spread and transmission of infectious diseases. Pandemonium’s publicly available tool allows individuals access to assess their risk of infection.  Pandemonium also gives public health authorities a tool to “cut through the ‘infodemic’” by giving them the information necessary to make more data-driven decisions. These predictions would ultimately allow policymakers to make more informed decisions.

Pandemonium is in the proof of concept stage. We are testing the front end/user interface to assess ease of use and functionality. The current test population includes a mix of casual computer users, web interface designers, and trained programmers/computer scientists. Currently, interface testing is by invitation only. We plan to make testing available to limited populations by June 2021. Initial populations will include high school and college students of varying social ethnic backgrounds and computer abilities, as well as senior living communities. These populations will give us a mix of use cases and allow for direct observations of user interactions. 

The analytic software is functional. Our team employs a continuous process of refinement and improvement. We are continuously incorporating additional data to expand our our regional coverage. We are building in models of vaccine use and subsequent risk mitigation as data becomes available.

• There are various open-source and free applications related to risk, but they are limited in their scope and are not designed to work together.  Our solution combines extracting data hidden in plain sight on social media, state data dashboards, while supplementing with crowd-sourced data along with combining with proprietary & confidential data (such as mobility data)  , and providing a framework to allow component models to be easily plugged into it and tested in a more holistically.  For example, if you develop a micro model, you do not need to develop all of the other components to test in a much broader context.  It makes it attractive for model developers to contribute their models to the menu of models that other researchers can select.

• Leveraging a Probabilistic Programming Language and ultimately Quantum Machine Learning is very innovative. 

• Our ability to handle mobility data in a sophisticated way that shields users from much of the complexity of implementation. 

• Our solution gives the public access to advanced analytical tools to provide individualized, granular risk analysis. We incorporate individual social history, health risk factors, and risk mitigation efforts with meso and macro infection databases to provide comprehensive single event and historical risk summaries.

• The people we directly serve will be through our Personal Risk Assement App.

• We currently have only test users as the solution is in the development & testing phase, and those users are cautioned not to rely on the results so we are not actually serving anyone yet, but we are preparing to enter the pilot phase by this Fall.

• Depending on the state of this or the next pandemic or epidemic at the time and local conditions,  in one year, we intend to serve up to 3% of the US population and have a significant user-base outside the US.  In 5 years, we hope to serve up to 10% of the US population and have very substantial global usage.  But even when there is not an active pandemic nor epidemic, our tool can still be used for other infectious diseases like the seasonal influenza.

• The number of people that we impact indirectly could be far higher by making our framework and models available to public health officials, researchers and even other developers that want to build apps upon our framework.

We have the following key indicators that will be measured as follows: 

1. Coverage ratio of people at-risk: vulnerable people accessing the app/total vulnerable population, measures the degree to which the project reaches the population most in need.

2. Effect of the population using our app in a given county: usage of our app will be measured by the number using the app divided by the total population of the state. 

3. Age- and sex-specific mortality rate caused by a pathogen.

4. Performance evaluation of app: So far we have not had significant results; but we are working toward them.

Project duration: 8  months 

• Activities: Platform design (periodical database acquisition and real time updating, updating knowledge through scientific articles), periodic training of assigned personnel, platform upgrade.

• Measurement of indicators within the established time frame.

• Time: It will depend on the amount of work and human resources.

• Currently, we have 9 people total working on the solution, all volunteers, including 4 UC-Berkeley interns. 

• 1 working mostly full-time now on it.

• Various advisory contributors, including 2 advisory board members (not counted above).

As a global solution team, we have a diverse skill set ranging from management to information technology to healthcare. Specifically, we bring skills and experience such as in life science industry, strategy and innovation, both business management and not-for-profit corporate management, privacy, GIS, data science, software development, artificial intelligence, including quantum machine learning, scheduling & combinatorial optimization, as well as numerical modeling and fluid mechanics. The team is led by a board of directors of MIT alumni; they bring strong technical experience and background to the team, and have expanded the team and skill sets beyond those. Our team has also adopted a strategic approach to build upon the existing solutions to lead to complex analyses to revolutionize the management of epidemics & pandemics like COVID19.

The team behind Quantum Risk Analytics, Inc. is highly diverse in terms of both talent and network connections. Most namely, we have roots with MIT alumni and MIT academic figures. Moreover, having a team of UC Berkeley interns, we have ties to the Berkeley community of researchers and technology firms, all of which are at the forefront of not only Silicon Valley but across the world. 

o    Recognition - We are  essentially completely unknown. Winning any prize regardless of the monetary amount would bring very helpful prestige, that would cause some to take a look at us and what we are doing who might otherwise disregard us out of hand.

o    Resources - Seed funding

§  We need to hire some part- & full-time employees in:

· Administrative

· Strategic Partnerships

· Development

· Marketing & Fundraising

· Grant writing

· Legal

§  Data - We may also need to purchase some proprietary data (The best offer we have so far £40,000 for a limited data set, representing a 90% discount.  But this probably would not suffice.)

§  Marketing & Operating expenses

o    Partners - In order to offer a platform & app, we need to work with tech. companies who have the hardware & cloud-computing infrastructure in place.  (Otherwise, we will be limited to offering an open-source software framework.)  Social media companies have pertinent data and infrastructure on which to run some of our ML models, e.g. social distancing & face masks recognition.  Other companies with proprietary data, such as vaccine manufacturers for efficacy.

We are developing a solution that employs various AI & ML technologies for the benefit of humanity. Our primary goal is to leverage these and other advanced technologies that would normally be out of reach for most people to utilize directly for assessing their risks for COVID-19 with the necessary specificity to allow them to optimize their choices to minimize COVID-19 risks weighed against other priorities. Simply staying home all the time indefinitely is not something that works for most people. So how do people understand what their true risks are in the various real-life scenarios that people face? We have formed a not-for-profit corporation to both develop the technology and make it available for free to help people get those specific answers.

The AI related technologies that we employ include Quantum Machine Learning, Deep Learning and other Machine Learning, probabilistic programming languages (such as Pyro), Bayesian inference, Markov Chain Monte Carlo (MCMC), Motion Tracking, Image Analysis and Classification, etc.

We would use the AI for Humanity Prize money, if we are so fortunate to receive the high honor of winning any, primarily to hire more highly-skilled technical staff to develop our solution more quickly and secondly to scale deployment, so it can help more people sooner, and expand to serve more countries sooner.

Pandemonium will allow health officials at all levels of government identify localities at greatest risk of infection. With this information, officials can allocate supplies to the areas where they will do the most good.