Do you have a penchant for building artificial intelligence (AI) models or exploring the potential harnessed within machine learning? Are you interested in helping change the current pandemic-induced global landscape? If so, the latest XPRIZE competition may be calling your name.
About 8 months ago, XPRIZE created a data driven alliance focused on more effectively using AI, machine learning and data to respond to COVID and pandemics – whether it's identifying better ways to do contract tracing or finding better drugs through simulations or predictive capabilities. "We've been doing this work for a few months with a number of our partners (including Anthem, Veterans Affairs, Intel, Microsoft, Google and Cognizant),” says Amir Banifatemi, chief innovation and growth officer of XPRIZE. “We were looking to identify how we can help policymakers and public health authorities with more of an augmented capability based on facts and evidence."
The outcome? XPRIZE decided to deploy a challenge for engineers and problem solvers to figure out how to get a better handle on what needs to be done to protect people and reopen economies, while minimizing potential outbreaks. “The idea is to create a public utility service that could operate in a similar fashion to how someone consults the Weather Channel to inform planning decisions,” says Banifatemi. "How can we give a number of organizations and those who have to protect people valuable suggestions and mitigations about policies on being safe, going to work, wearing masks or opening schools.”
Banifatemi tells IndustryWeek, all those decisions should come from a numerous gathered data points based on the number of people who have contracted the virus, those currently under treatment, the level of access to health logistics as well as other statistical data elements. "This type of statistical data is available, but not in a way that helps us be more granular with a better resolution at the community level," he says.
Down to details
Based on technology and AI models developed by Cognizant, and using data compiled by the Oxford COVID-19 Government Response Tracker (based at Oxford University’s Blavatnik School of Government), competing teams will build data-driven AI models that predict local COVID-19 transmission rates and prescribe intervention and mitigation measures that, with testing in “what-if” scenarios, are shown to minimize infection rates as well as negative economic impacts. Successful models may also serve as a roadmap for future crises.
According to the official XPRIZE release, a total purse of $500K will be awarded at the conclusion of the challenge, which will close in February 2021. Unlike previous competitions, the urgency to end the COVID-19 pandemic and its devastating effects calls for an accelerated four-month timeframe for the Pandemic Response Challenge. Teams must register to join the competition by December 8, 2020. In Phase 1, teams will provide accurate predictions of COVID-19 transmission based on local data, unique intervention strategies, and mitigation policies and practices by December 22, 2020. In Phase 2, the top 50 teams will provide prescriptor models that will be evaluated against minimizing the number of cases and minimizing the stringency (i.e. cost) of the Intervention Plans (IPs). Phase 2 will conclude in February 2021. Throughout each phase, teams will be provided with cloud and compute services, courtesy of supporting partner, AWS, to facilitate development of their proposed solutions.
“Unfortunately, there is a lot of uncertainty and misinformation about what needs to be done, so having factual data is better,” he says. “As vaccines are going to be coming, and gradually be made available, the same mitigation needs to be happening. We don't pretend that AI will have the answer for you, it's just going to be a suggestion, and should help support better decision making.”
Harnessing technology and collaboration
This is an AI/machine learning in action style competition, explains Banifatemi. “Competing teams will need to have a notion about data science, machine learning, but also immunology, public health statistics or social protection or health care mitigation or logistics in health or health coverage,” he says. Many teams could include one or two people in one discipline working with others possessing complementary key competencies. The XPrize platform allows for teams to meet one another, merge, complement or recruit to add new people.
“It is important to note that collaboration matters. A number of barriers to collaboration were broken because of the first wave of COVID when there was a dire need to actively collaborate and share knowledge,” he says. “Collaborating to create something that can serve as a public utility is another part of the fight. We can create something amazing through the capabilities of machine learning and AI.”
Banifatemi notes many teams will have commercial content that will find amazing opportunities coming out of this competition because they will have demonstrated their capabilities and the robustness of their technology or innovation.
“We hope that the connection between AI, machine learning, public health and resilience will be a starting point for a number of conversations and opportunities for more data sharing,” he says.