HPC Response to the COVID-19 Pandemic

Description

This panel brought together multiple experts across a vast range of scientific disciplines from numerical simulation to data-informed diagnosis in order to feature the contributions made over the past 12+ months by the HPC community to provide a multimodal response to the world-wide emergence and spread of SARS-CoV-2 coronavirus and the ensuing global COVID-19 pandemic. Their recent efforts highlighted the importance of computational science as a tool that helps uncover the structure and mechanics of the viral spread and drastically reduced the number of candidate treatments eligible for laboratory testing. From the transformative medical advances to drug discovery driven by supercomputing, this panel spaned a range of approaches to the ongoing crisis and the upcoming challenges of mutating virus variants and vaccine hesitancy. By casting our panelist selection net widely, we facilitated for a vigorous discussion and productive interaction toward comprehensive conclusions to be drawn by the attendees.

The emergence of SARS-CoV-2 virus around December 2019 and the ensuing pandemic in the following months created unprecedented circumstances and the aftermath would be felt for years to come as we created the new normal in our professional and personal lives.

Organizers:

Panelists:

David Keyes
King Abdullah University of Science and Technology (KAUST) [ slides ]
Rick Stevens
Argonne National Laboratory (ANL), University of Chicago
Florent Duchaine
CERFACS, France
Makoto Tsubokura
Kobe University, Japan [ slides ]
Daniel Jacobson
Oak Ridge National Laboratory (ORNL)
Patricia Kovatch
Icahn School of Medicine at Mount Sinai, Columbia University [ slides ]

Sample Panelists' Questionnaire

  1. What is the most interesting update from your presentation at SIAM CSE21?
  2. Which new COVID-related issue emerged as the active area for HPC to tackle?
  3. Do the COVID variants represent a game changer in the current HPC fight of the global pandemic?
  4. What are the specific HPC-COVID issues in your country/continent/region?
  5. What HPC collaborations do you consider to be important to leverage interdisciplinary and interna- tional HPC community?
  6. What is the greatest opportunity going forward for HPC involvement in COVID response?

Perspectives Represented by the Panelists

We believe that a successful panel should include a variety of approaches and diversity of perspectives. This avoids single-mindedness and group-think and allows for cross-pollination of ideas that leverage the network effects of many-to-many interactions across a broad scope of contributions. This mitigates the risk that the panel ends up being a boring repetition of similar views. To that end, we gathered a set of panelists that guarantee the required variety and diversity across the following approaches:

  1. Viral droplet propagation with dynamic fluid transport models including High-Fidelity Large-Eddy Simulations -- some the world's largest.
  2. Drug lifecycle identification and maintenance. AI-based diagnostics based on a range of medical tests.
  3. Study of symptomatology and mechanisms in COVID-19 immunological response.
  4. Detection and treatment for inflammatory markers of COVID-related lung disease in, for example, CT scans.
  5. Involvement of the HPC consortium of 40 members for a unified mechanism for COVID-related HPC funding.
  6. DOE sponsored National Virtual BioTechnology Laboratory (NVBTL) and its role in coordinating research across institutional silos.
  7. panelists are recognized experts in their respective fields of science and hold a multitude of positions that intersect HPC and the COVID-19 task forces. They have been selected across world-wide locations and from different areas of societal engagement including biology, medicine, CFD simulations, AI-methods, and governmental funding agencies. We believe that this diversity will benefit and enable broader participation by the audience.

  1. Visualization -- (virus geometric features)
  2. Molecular dynamics -- (viral mechanisms)
  3. Fluid-dynamics and transport -- (spread, safety, protocols, and flow obstacles)
  4. Diagnostics -- (clinical predictive accuracy)
  5. Drug selection and repurposing -- (trial selection and efficacy)
  6. Tracking spread and mutation rate -- (understanding variant emergence)
  7. Forecasting spread -- (resource planning and allocation across international boundaries)
  8. Optimizing response -- (policy setting and sun-setting)
  9. Vaccination logistics -- (development/boosters/manufacture/resource allocation)
  10. Social (mis)information dynamics -- (beyond sentiment analysis)
  11. Economical consequences -- (macro-economics modeling of quarantine policies)

Audience Participation

Due to SC21' strong hybrid format for the event, we were prepared to offer fully interactive panel experience for both in-person and online particpants. We offered the audience an opportunity to move to a fully interactive experience without the awkward cross-talk of online meetings. We took advantage of the SC21 audio-visual aides to engage the interaction between the panelists and the audience on the HUBB platform.

Related Event from the Past

A related event was held earlier this year at the SIAM CSE21 conference in March 2021: SIAM CSE21 minisympsium with two organizers: Piotr Luszczek and Hatem Ltaief (moderators on this panel).

This served as a backdrop for the panel by providing reference material for the audience and the starting point for the panelists to discuss the new updates and argue where the HPC should focus next regarding the COVID19 response.

The past SIAM event could have been considered a sequence of mini-presentations and had limited interaction between panelists due to the tight schedule imposed by the organizers. The SC21 panel was informed by the interactions that were spurred in-between these presentations and was more interactive.