Mila > News > An interdisciplinary team led by Mila researchers reaches final round of global Pandemic Response Challenge

11 Mar 2021

An interdisciplinary team led by Mila researchers reaches final round of global Pandemic Response Challenge

M-montreal-quebec, a team led by Mila researchers Marc-Andre Rousseau and Brady Neal, among the finalists of the $500K Pandemic Response Challenge launched by XPRIZE and Cognizant. Winners of the competition were announced earlier this week.

Launched in November 2020, the Pandemic Response Challenge invited teams from anywhere on the planet to develop data-driven AI systems to predict COVID-19 infection rates and prescribe intervention plans that regional governments, communities, and organizations could implement to contain the pandemic and safely reopen. M-montreal-quebec was among a pool of 48 finalists from 17 countries selected for the final round out of a total of 104 teams from 28 countries.

Rousseau, with research expertise in immunology and microbiology, partnered up with Neal, a machine learning (ML) researcher, recruiting 10 additional researchers and three advisors from Mila and around the world to form a cross-disciplinary research team at the intersection of life sciences, ML and economics (see the full list of participants and affiliations below).

“Our team was full of global talent,” said Neal, a PhD student at Mila and UdeM who helped recruit team members from Luxembourg, Portugal, Australia, and India.

M-montreal-quebec team members, first row, from left to right: Marc-Andre Rousseau*, Brady Neal*, Marharyta Aleksandrova, Thin Nguyen, Diogo Pernes, Chen-Yang Su*. Second row: Sai Aravind Sreeramadas*, Prabodh Wankhede*, Baihan Lin, Olexa Bilaniuk*, Andrew Williams*, Makesh Narsimhan*. Team Advisors, third row: Irina Rish*, Djallel Bounneffouf and Yang Zhang. *Mila members

The competition consisted of two phases. In Phase 1, teams were tasked with analyzing COVID-19 data, intervention strategies, and mitigation policies to develop and test a prediction model that could anticipate global infection rates. During the prediction phase, M-montreal-quebec set themselves apart with their temperature hypothesis, namely that discomfort caused by high or low temperatures tends to cause people to flock indoors, where air is often recirculated through HVAC systems. Such indoor conditions could aggregate levels of aerosolized particles of a respiratory virus like SARS-CoV-2 and lead to a higher risk of new infections.

“Phase 2 was more complex, we had to recommend optimal government interventions while taking into account the associated costs,” explained Neal. Action plans put forward by participating teams were evaluated against key benchmarks, including minimizing the number of cases and mitigating the severity of intervention plans and confinement measures.

M-montreal-quebec pursued this stage of the challenge by employing evolutionary methods for evolving neural networks, as well as framing the problem as a contextual combinatorial bandit problem in order to consider a wide range of variables.

“The finalists in the Pandemic Response Challenge have demonstrated incredible innovation in their efforts to help the world emerge from the COVID-19 pandemic,” said Brian Humphries, CEO of Cognizant, when the 48 finalists were selected. “Advancements these teams [have made] can have far-reaching implications—empowering policy-makers and business leaders globally with data-driven tools […] and proving the value of AI and collaboration in addressing future humanitarian crises.”


For more information:

To learn more about the project, visit: http://m-montreal-quebec.ca

XPRIZE Pandemic Response Challenge Winners Announcement: https://news.cognizant.com/2021-3-9-XPRIZE-and-Cognizant-Announce-Grand-Prize-Winners-in-Pandemic-Response-Challenge

M-montreal-quebec Team Members:

Team Leaders

  • Marc-Andre Rousseau, Intern, UdeM, Mila
  • Brady Neal, PhD Student, UdeM, Mila

Team Members

  • Diogo Pernes, PhD Student, University of Porto, INESC TEC
  • Thin Nguyen, Senior Research Fellow, Applied AI Institute, Deakin University
  • Marharyta Aleksandrova, Postdoctoral Researcher, University of Luxembourg
  • Andrew Williams, MSc Student, UdeM, Mila-LITQ
  • Chen-Yang Su, MSc Student, McGill, Mila, Lady Davis Institute (Jewish General Hospital)
  • Olexa Bilaniuk, Research Software Developer, Innovation, Development and Technologies, Mila
  • Makesh Narsimhan, MSc Student, UdeM, Mila
  • Baihan Lin, PhD Student, Center for Theoretical Neuroscience and Zuckerman Mind Brain and Behavior Institute, Columbia University
  • Prabodh Wankhede, MSc Student, UdeM, Mila
  • Sai Aravind Sreeramadas, MSc Student, UdeM, Mila

Advisors

  • Irina Rish, Associate Professor, UdeM, Mila
  • Djallel Bounneffouf, Team Leader and Research Staff Member, AI Foundations Group of the IBM Thomas J. Watson Research Center
  • Yang Zhang, Director of Model Development, Canadian Economic Analysis Department at the Bank of Canada