A team of Hopkins undergraduate students is participating in ProjectX, a machine learning competition hosted by the University of Toronto Undergraduate Artificial Intelligence Group (UofT AI Group). Teams from 23 universities are competing in this three-month research-based competition for a prize of $70,000.
Matthew Figdore, a senior Computer Science major, is one of the undergraduates on the team. He is joined by Sally Cao and Daniel Weber, who are studying Computer Science and Applied Mathematics and Statistics. Daniel Borders, a senior Biomedical Engineering and Applied Mathematics and Statistics major, and Gary Yang, a Biomedical Engineering and Computer Science major, are also on the team.
Figdore discovered the competition on the Computer Science Department's Slack channel. He forwarded the event’s information to the mailing list of all undergraduates studying Computer Science — Cao, Borders, Yang and Weber responded and became his new teammates.
This year the competition is focused on mitigating climate change. On Sept. 1, ProjectX assigned focus areas related to climate change to students, who began working with experts in the field and research mentors.
The competition is broken down into several tracks, such as weather disaster early prediction and emissions and energy efficiency. The Hopkins ProjectX team is joining six other teams in the infectious disease detection and prevention track of the competition, which Figdore explained is an important angle.
“There are numerous links between climate change and increasing risk of infection,” Figdore wrote in an email to The News-Letter. “For example, climate change can result in higher risks of malaria and superbugs and the rise of new fungal diseases.”
The convergence of viruses and climate change has been an important area of research in light of the current pandemic. A study published in Nature reported that about 30% of emerging diseases are transmitted through vectors — living organisms that can transmit infectious disease such as mosquitoes, birds, flies and lice. Vectors act as virus hosts and adapt to climate conditions, such as temperature and precipitation, to survive and reproduce, which could affect disease transmission.
ProjectX has paired the team with Dr. Swaroop Vedula, an assistant research professor at the Malone Center for Engineering in Healthcare. Vedula is an epidemiologist and a medical doctor by training. As a ProjectX advisor and mentor, he facilitates communication with program organizers and guides the team’s research. The team has also reached out to infectious disease experts with the help of the competition’s organizers. Figdore explained that the team has felt supported throughout the entire process.
“ProjectX is very well run, and I appreciate the resources the competition organizers have provided us. I'm really grateful to get to work with my teammates on this project,” Figdore wrote.
The team recently submitted a written project proposal to the competition. Figdore, Loggia and Weber completed extensive research on infectious disease and formulated ways to approach solutions in the field, and while the exact proposal is still confidential, the team revealed that they combined their individual specialties to form a creative model.
Specifically, Fidgore is using his background in artificial intelligence to help set up the computing infrastructure and contribute to model development. Other members of the team have strong backgrounds in machine learning and prior experience conducting research in the field.
The team is working hard to complete the final draft of the research paper by the Nov. 23 deadline. Final research papers will be judged by researchers in academia and experts in the industry. ProjectX will select the winning teams by December.
The Hopkins team has enjoyed the competition process.
“We're all taking part in writing our manuscript. It's been exciting to meet and work with the other team members so far,” Fidgore wrote.
ProjectX gives undergraduate students a rare opportunity to apply machine learning to solve real-world problems. The winners of the competition will present their research at the UofT AI Group’s annual conference in February and receive the $70,000 prize. Figdore, however, is more excited about how his team’s work might be shared with the larger scientific community rather than the prize money.
“I think it would be exciting to publish our work if we're able to get strong results with our machine learning models,” he wrote.
After the competition ends, the team hopes that their research will have a real impact on both climate change and infectious disease.
Correction: This article originally did not include team members Sally Cao, Daniel Borders and Gary Yang and included Spencer Loggia. This roster was not correct at the time of publication. The News-Letter regrets this error.
