Before COVID-19, my main work was across math research with a focus on network theory and education, coaching the USA International Math Olympiad team, and coming up with new ways to teach math. I am also a Hertz Foundation Fellow. Each year, the organization identifies about 15 up-and-coming researchers across science, technology, engineering, and math, and provides them with funding for their Ph.D. study in exchange for a moral commitment to assist in times of national emergency. I was activated via the foundation in mid-March to respond to the COVID-19 threat and put everything else down to focus on developing an effective and free of charge COVID-19 contact tracing solution.
I realized that my mathematical research area of network theory contained insights and approaches that could be key to controlling one of the most slippery characteristics of COVID-19: its tendency to spread before people show symptoms. Upon further investigation, I saw a way to anonymously collect the interaction network between smartphones, without linking the phones with any personal information. We could then apply methods from network theory to predict the spread of infection and intervene in advance to prevent people from getting infected. At the time, nobody was using the full power of network theory to control COVID-19, which made me decide to lead this network-empowered approach. I had the fortune of meeting several key collaborators who joined this effort early and defined the intensely committed ethos of the team that designed and built NOVID. Five months later, we are the only group left in the English-speaking world that is using network theory to fight COVID-19 at scale.
What makes NOVID different from other contact tracing solutions and why will your app solve the problem when others cannot?
The exposure notification apps you read about today are largely the same. NOVID, however, is significantly different in two ways that set it apart from every other contact tracing app in the world. And, fortunately, NOVID can coexist with the other apps. It is possible to have both apps running on your phone at the same time, without significant battery drain.
First, NOVID does not only deliver post-exposure notifications like the other apps that digitize traditional contact tracing anonymously, informing people after they have already been exposed to COVID-19-positive people, and asking them to voluntarily follow the procedures of isolation or contact tracing. Indeed, NOVID also uses network theory to deliver pre-exposure notification, informing you how far away COVID-19 has struck, measured in terms of the number of relationships from you. For example, if COVID-19 struck your coworker’s husband’s coworker, then you would find out that there was a COVID-19 case at distance three from you. This represents a new paradigm for 21st-century contact tracing. Instead of only alerting people after they have already been exposed, NOVID tells the user how far away the positive cases struck. They can then take precautionary measures as the cases strike closer to home. This is analogous to providing satellite imagery of a hurricane approaching when it is far off in the mid-Atlantic, as opposed to only finding out once it presents an imminent threat.
Importantly, this approach aligns incentives. NOVID is not only an app to protect others from you if you get sick. It also protects you from other people who are COVID-19-positive. This significantly incentivizes adoption, to the degree that in the current NOVID deployment at Georgia Tech, student leaders independently chose to promote NOVID as a solution even though it was initially introduced by the university administration.
Furthermore, NOVID also offers a new and powerful method of achieving impact. Other apps rely on voluntary compliance with anonymous requests after you are told that you were directly near a COVID-19-positive person. Since other apps do not use the network, they are unable to provide any foresight. As a result, they rely on extremely high rates of adoption, as well as high rates of compliance after receiving post-exposure notifications. In fact, with other apps, if COVID-19 reaches a person who chooses not to comply, the disease will continue to spread past the person without anyone else receiving a warning.
NOVID is different. Because it constructs a long-range network, it is useful even at the low rate of only 10% adoption. App users have commented that NOVID’s long-range network information has prompted behavior alteration, including not going into crowded places they may have visited previously and increasing health and safety behaviors such as more frequent or higher quality face covering usage. NOVID leverages selfish human instincts of self-preservation to increase precaution at scale at the critical time when COVID-19 is highly active in the app user's personal network. Therefore, NOVID works well despite low adoption rates in the same way that even if your TV image was filled with static, you would still be alerted to a hurricane coming your way a day before landfall and be in a position to take the necessary precautions.
Unfortunately, it is highly unlikely that any country or state will be able to have any app reach the very high levels of adoption that anonymous post-exposure notification apps require. Other approaches missed the value of the anonymous interaction network in substantially decreasing the adoption threshold required for effectiveness, which was NOVID’s foundational feature. Thankfully, NOVID can coexist on the same phone with a post-exposure notification app. So, even if your country or region initially supported any of the other apps, it is not too late. You can use NOVID in conjunction to provide pre-exposure notification.
NOVID’s second key differentiating feature is that we do not rely on Bluetooth for distance measurement. Instead, we use a much more accurate custom-created ultrasonic distance measurement method.
Apps have been popping up all over the world, and they all claim to be accurate enough to perform contact tracing. How do we know that NOVID is much more accurate than other apps?
It is unfortunate that misinformation has been spread by other app developers. Their announcements echoed aspirations of being able to accurately detect six-foot proximity. Essentially all other contact tracing apps rely on GPS or Bluetooth for their distance estimates. Many smartphone users know from personal experience with their apps that civilian GPS is not able to identify distance with one-foot accuracy. The laws of physics prevent Bluetooth-based methods from detecting distance with this accuracy. Indeed,
a systematic Irish study revealed that Bluetooth-based solutions are incapable of measuring distance with the necessary accuracy for contact tracing.
While many articles have raised accuracy issues about Bluetooth, it is telling that there is not a single article that demonstrates the successful use of Bluetooth to reliably measure distance. Because it is important to be forthcoming about accuracy, NOVID took the bold step of displaying its real-time distance estimates from nearby devices on its main screen for its users to see. This allows everyone to know just how accurate the distance measurement is. Our team also conducted a systematic experiment with nearly 600 measurements using five-year-old iPhones and
publicly shared the positive results.
NOVID was able to achieve this by using ultrasound to estimate distance instead of Bluetooth. This is similar to how dolphins and bats navigate, or to how most people estimate how far away lightning has struck by assessing how long it takes the sound of the thunder to reach ones ears. Although this sounds easy, it represents a significant technological challenge because we need to estimate the time for sound to travel with sub-millisecond accuracy. Consequently, there were no “off-the-shelf” application components that could be assembled into an ultrasonic distance sensing smartphone app. The NOVID team developed this new technology on its own. As a result, NOVID is the only contact tracing app in the world that includes ultrasonic distance measurement.
We hope that every Bluetooth-based app that seeks mass adoption will also be upfront about its accuracy. If every contact tracing app were to be as forthcoming as NOVID and display its estimated distances to nearby devices on the main screen for every user to see and test for themselves, people would be able to make better informed decisions.
Why did you choose a solution that focused so heavily on anonymity and really how anonymous is this data?
When I turned my full attention to fighting COVID-19, my goal was not to just make an app, but to help end COVID-19. Therefore, the technology of the app was only one part of the equation. The other part concerned incentives for mass adoption. We did not want to build an app whose only pathway to mass adoption was by force, whether through mandatory installation or by suppressing better approaches. Consequently, anonymity was essential. We needed to build an app with no downside to installation.
Fortunately, because our starting point was network theory, we saw that we could achieve our objectives without collecting personal information. It is not important exactly who has COVID-19 as long as the app has a way of informing devices that had been nearby. Also, because our distance measurement method used ultrasound instead of GPS, we did not need to collect sensitive, personal data.
We protect the anonymity of the data by not asking for any information about the user in the first place. When new users install the app, they do not provide their phone number, email address, or ZIP code. We instead ask for sensor permissions and use them responsibly.
Georgia Tech Research Institute conducted an independent assessment and reported no evidence that sound collected via the microphone is stored or transmitted. We continue to send the independent researchers the source code for our app as it is approved for distribution by Apple and Google.
The bottom line is this: NOVID does not intentionally collect any personal information. People are used to apps collecting personal information to sell advertising as that is the typical model of an app that is run as a business. NOVID has no interest in collecting personal information because our goal is to do our part to eliminate COVID-19 in the most effective, accurate, and anonymous way possible.
How do you see this technology affecting the world long term and what does the future of NOVID look like?
We have introduced a new and distinctly 21st-century angle on contact tracing, analogous to providing satellite tracking of hurricanes. Before hurricane tracking, people had little to no warning as hurricanes approached. Now, we take it for granted that hurricanes are tracked, giving people time to adequately prepare for impact. NOVID brings this paradigm-shifting approach to controlling infectious disease, using transmitters and sensors that only became ubiquitously available on smartphones in the recent past. We provide everyone with a smartphone around the world with their own independent capability to see COVID-19 coming, so they can respond with the level of caution they deem appropriate. This puts control into the hands of the app users who can alter their behavior to protect themselves. We seek to significantly boost awareness of this new approach and tool as rapidly as possible in order to help restore our world to normalcy.
NOVID’s first deployments happen on university campuses, but we are already talking to school districts and cities about how we can assist at wider scales. We aim to work with city, county, state, and national governments to help bring our new approach to as many people as possible.
Our work will likely have a significant impact on the way people around the globe are going to respond to future pandemics. It is our hope that when the next novel virus arises, significantly larger numbers of people in various parts of the world will be able to immediately install the appropriate variant of NOVID on their phones and help stamp out the disease before it spreads worldwide. Unfortunately, disease outbreaks occur frequently. After controlling COVID-19, the NOVID team will hence turn its attention to other diseases as they arise. That is the reason why we are in the process of accumulating expertise in network-based analysis of infectious diseases. And through the focused fight against COVID, we aim to develop a veteran team with substantial breadth and depth of knowledge and experience. Our team's goal is to help protect the world on an ongoing basis.
The NOVID team sincerely thanks C4CS
® for the assistance to increase awareness of the many important but nuanced issues in this space, and helping the general public to interpret the characteristics that set NOVID apart.
C4CS® has been assisting NOVID pro bono. Recently, Our Region's Business host Bill Flanagan interviewed NOVID founder Dr. Po-Shen Loh and C4CS® President & CEO Oliver S. Schmidt about crisis management, COVID-19 response, and effective contact tracing. The interview can be accessed via this link.
The free NOVID app is available for Apple and Android devices. If you would like to visit NOVID's website, please click here.
