Blugold research team studies use of artificial intelligence during COVID-19 pandemic

His hope, Langlois says, was to find a way to integrate his prior health care experience with his interest in technology, something he thought the University of Wisconsin-Eau Claire could help him do.

“UWEC was the complete package deal that offered a great location, the programs that I specifically wanted and the timeline for graduation that would set me up for success in the future,” says Langlois, an Altoona native.

A year later, Langlois is among the student authors of a published paper, “A Comprehensive Review of Machine Learning Used to Combat COVID-19,” a study reviewing how machine learning was used in the diagnosis and treatment of COVID-19 in computed tomography (CT) and X-ray imaging.

“I’ve always been technology oriented and so the field of computer science was a natural choice for me,” Langlois says. “I knew I wanted to integrate my health care experiences into some of the computer science work that I would be doing. Luckily, I managed to do that with the research that I work on.”

The research project involved examining what other researchers were doing and how successful they were in creating models to solve problems related to COVID-19, along with examining the limitations and difficulties they experienced, Langlois says.

This summer, the research team’s paper, co-authored by Dr. Rahul Gomes, assistant professor of computer science, Dr. Papia Rozario, assistant professor in geography and anthropology, and nine undergraduate students — including Langlois — was published in Diagnostics, a professional journal on medical diagnosis.

A comprehensive review

Artificial intelligence (AI) — which allows machines to efficiently solve problems — was used extensively during the pandemic, Gomes says. For example, it helped clinicians use CT scans and radiographs to diagnose patients with COVID-19, and AI models predicted disease progression over time, he says.

Many researchers documented how AI was used to combat COVID-19, Gomes says, noting that a 2020 paper showed that more than 20,000 articles were published about it in just six months.

The UW-Eau Claire research team’s comprehensive review focused on published papers that looked at machine learning used to combat COVID-19, where imaging data played a significant role, Gomes says.

“We explored how the research progressed within a specific time frame of 2021 and 2022 and investigated its distribution both geographically and by publishers,” Gomes says.

Langlois says it was interesting to see how COVID-19 research evolved throughout the pandemic.

“Papers were being published every day, so we always had new content to sort through,” Langlois says. “The availability of information also increased as public data sets became more prevalent for researchers to utilize. With the pandemic looming in the background, it just made the research more personal and the effects of this research more tangible.”

Since the use of geographic information systems (GIS) to determine geographical distribution patterns of diseases has increased significantly in recent years, GIS played a significant part in the project, Rozario says. The project “showed us the positive impact GIS can have on health care,” she says.

Gomes agrees, saying projects like this one are “possible only if students and faculty from different disciplines share their knowledge and expertise.”

One of the takeaways from the project is the need for a centralized data repository and a means to efficiently distribute knowledge across a larger population when necessary, Rozario says.

The project also documents the enormous impact AI, machine and deep learning can have during a massive crisis like the COVID-19 pandemic, Gomes says.

“I am not looking forward to another pandemic in my lifetime, but if a crisis arises, we do have the AI means and capability to handle this situation,” Gomes says.

Real-world learning

The project helped student researchers gain new skills by contributing to the body of research that focuses on the role AI and machine learning can play in understanding and addressing COVID-19.

“The students did a great job,” Rozario says. “They reviewed more than 150 papers to curate the data and learned a great deal about the spatial distribution characteristics of the disease.”

The comprehensive review “really helped our students know how machine learning is vital and applicable with direct outcomes,” Gomes says.

Gomes says this was the first time his research team did such an extensive review paper and on such a vital topic.

“We were nervous all along, but our students did a fantastic job curating data and presenting them in an orderly manner,” Gomes says. “Review papers can be very challenging to write. The skills that our students demonstrated have been phenomenal.”

Langlois knew within weeks of being a Blugold that he wanted to be part of UW-Eau Claire’s nationally known undergraduate research program. He told Gomes that he was interested in the work his undergraduate research team was doing in collaboration with Mayo Clinic Health System. Gomes invited him to join his research team and assigned him to the then-new COVID-19 project.

The research project helped him learn valuable new skills, but it also has him thinking differently about his future, says Langlois, who will graduate in May 2023.

“The amount of knowledge that I learned from this research project has been immense,” Langlois says. “Machine learning is an extremely important but difficult field to navigate, and Dr. Gomes helped us through every step of the way.”

Research gives students opportunities to see the bigger picture and explore specializations in their career as a computer science major, Gomes says. It also leads to other opportunities, he says, noting that students in his research lab have had internships at Mayo Clinic that specialize in deep learning.

“This is a very dynamic field with a lot of potential and direct patient outcomes, which makes students very excited and gives them confidence in their ability to solve complex problems,” Gomes says.

That’s certainly true for Langlois, who says the research he’s doing is giving him a new appreciation for what might be possible in his future career.

Having his work published as an undergraduate student will look great on a resume, but it goes beyond just that, Langlois says.

“Being exposed early in my computer science career to more advanced topics such as artificial intelligence helps open future potential interests,” Langlois says. “One of the fields that I’m most interested in is computer vision or how computers can understand images/videos, and this research helps to lay the foundation for that path.”

Langlois already is working with Gomes on another research project, this one involving using deep learning on spectral imaging of kidney tissue microarrays to detect certain kinds of cancer.

His current plan is to work in the field of software development, but “I’m open to wherever my computer science journey takes me,” Langlois says.