VinBrain reaches final round of global competition “COVID-19 Pandemic Response Challenge”

The VinBrain team in the US has reached the final
round of the Pandemic Response Challenge held by XPRIZE and Cognizant.

This competition, which has prize money of
$500,000, aims to harness the power of data and artificial intelligence to
predict COVID-19 infection rates, and propose Intervention Plans for
governments, communities and regional institutions to control COVID-19 Pandemic
when reopening the economy.

HANOI, VIETNAM – Media OutReach – 24 February 2021 – The Pandemic Response Challenge co-organised by XPRIZE and Cognizant was
launched in November 2020. Participating teams will develop models based on
artificial intelligence to make recommendations to achieve the goal
of reopening society safely and limit negative impacts on the economy
while minimising the SARS-CoV-2 virus outbreak.

Based on the jury’s independent assessment of the AI prediction of the
COVID-19 trend and the infection patterns of the participating teams, the jury
selected 48 teams from 17 countries for the final round out of a total of 104
teams from 28 countries in the semi-finals.

Thanks to the guidance by Mr. Mudit Jain — the Deputy General Director of
VinBrain Product Development in the United States, the Vinteam team has
developed a solution applying artificial intelligence to the accurate diagnosis
of COVID-19 infections and provides intervention options.

The infection patterns vary by country and territory. As a result,
interventions may work in one country or territory but be less effective in
another.

Members of VinBrain in the US have come up with an extensive solution to
this problem by developing the region-specific AI models to predict the
COVID-19 infection time series in those regions. The Vinteam developed this
solution based on a classical epidemic model called the SEIR model,  which
was then expanded to integrate with artificial intelligence to turn it into a
dynamic model. In order to train its AI models, the team uses available
features/data such as demographics, medical infrastructure, and economic
indicators. A special feature used by Vinteam is Google Maps mobile metrics
collected from worldwide Android phones in an anonymous and secure way,
measuring the movements of people in public areas such as parks, hospitals, workplaces,
grocery stores, stations, airports, etc. Another unique feature used by Vinteam
is Google’s COVID-19 infection trend data which is aggregated anonymously and
contains popular COVID-19 Symptom-Related Google Searches.

The jury selected Vinteam for the final round based on the accuracy of the
AI ​​models and its ability to easily interpret the predictions of the
solution. In addition, this approach can be used for any future pandemic
response model.

Mr. Amir Banifatemi, General Manager for Innovation and Growth, said: “We
aim to maximise the power of collaboration, competition and innovation to
accelerate possible solutions for COVID-19 and future pandemics. The
competition has shown encouraging results in a very short time, reflecting the
application of artificial intelligence to the society.”

The Pandemic Response Challenge is sponsored by XPRIZE, which designs and
operates incentive competitions to solve humanity’s grand challenges, in
partnership with Cognizant, which is considered one of 15 iconic companies in
the US, operating extensively in the field of information technology,
developing business processes and consulting cooperation for some of the most
successful units worldwide. The organisers also hope the competition will
promote the use of artificial intelligence and data in tackling other
humanitarian challenges. The competition’s results will be announced at the end
of February 2021.