The power of collaboration and teamwork in solving scientific challenges is well recognized. From research labs to technology companies, teamwork is often touted as the key to success. However, harnessing the power of competition—seemingly the opposite of collaboration—can also serve as a powerful approach to finding answers to important problems. Specifically, crowdsourcing contests are an effective approach where large numbers of people are invited to compete to provide solutions to challenging problems, with winning entries being awarded prizes.
This approach has been used effectively in research related to the COVID-19 pandemic. For example, when the pandemic hit the United States in March 2020, a collaborative effort was initiated between various institutions to create a database of knowledge on COVID-19 (led by the Allen Center for AI along with Georgetown University, the Chan Zuckerberg Initiative, and the White House).
This database, known as the COVID-19 Open Research Dataset (CORD-19) [1], consolidated a wide range of research studies examining various aspects of the disease. With this information, users could easily perform their own analyses of COVID-19, including searching in the text of papers for keywords that give insight into the behavior of the virus.
The crowdsourcing aspect of this effort was the CORD-19 Challenge [2], which was managed by Kaggle, a community of data scientists and machine learning practitioners. Through the CORD-19 Challenge, researchers were tasked with developing data mining tools to answer key COVID-19 questions such as, “What do we know about non-pharmaceutical interventions?” and “What do we know about vaccines and therapeutics?”
The CORD-19 Challenge was run as a contest, with rewards of $1000 for the best answer to each of these questions. This format helps generate a large number of responses for each question to select the ideal solution. By creating a competitive environment that pushes participants to create better solutions than others, it is possible to come up with a wide range of potential answers to the problems presented by COVID-19.
How has the CORD-19 Challenge assisted in the process of coming up with answers? For one, it has resulted in a vast collection of data on research studies relating to various aspects of the COVID-19 pandemic. It has also resulted in hundreds of thousands of research papers being analyzed through machine learning and text mining algorithms to come to various conclusions on the questions mentioned above.
The contest format for the CORD-19 Challenge has many benefits. One advantage is that different individuals will have different approaches and skill sets to solving a particular challenge. The competition format allows organizers to cost-effectively look at a range of solutions and pick the best. Why do the individuals participate in the contests, knowing that their efforts can potentially go uncompensated if they do not come out on top? The human motivation for this comes from not only the desire to contribute and be a part of a solution, but also from the thrill of competing against others. By tapping into this inherent desire to win, crowdsourcing contests like the CORD-19 Challenge have found a powerful way of identifying the best answers to a variety of questions.
[1] Lu Wang, Lucy et al. “CORD-19: The Covid-19 Open Research Dataset.” ArXiv arXiv:2004.10706v2. 22 Apr. 2020 Preprint.
[2] “COVID-19 Open Research Dataset Challenge (CORD-19).” Kaggle, 2020, www.kaggle.com/allen-institute-for-ai/CORD-19-research-challenge.