Avi Devy Mohan has had multiple research experiences as a Blugold after transferring to UW-Eau Claire from a university in Malaysia. A computer science and mathematics major, Mohan has worked alongside faculty research mentors from three academic areas on campus as well as with physicians at Mayo Clinic Health System. (Photo by Bill Hoepner)
The opportunity to do undergraduate research is among the reasons Avi Devy Mohan decided to transfer to the University of Wisconsin-Eau Claire from a university in her native Malaysia.
Still, the Blugold — now a senior computer science and applied mathematics major — never expected that her research would include collaborations with faculty in three academic departments as well as physicians at one of the most respected health care systems in the world.
Even more surprising is that her research has ranged from biomedical technology to recreational mathematics, helping her understand the many ways researchers can influence everything from health care to wallpaper.
“Taking part in research has been a great learning experience,” Mohan says. “I have learned so much about deep learning and how it can be used in different fields.”
Embracing research opportunities
Mohan is part of a UW-Eau Claire computer science research team that is collaborating with physicians at Mayo Clinic Health System, a project that involves creating a deep learning model that can detect inferior vena cava (IVC) filters in scans.
“Deep learning” is a subset of machine learning that derives its working logic from the functioning of the brain. Due to its high accuracy, deep learning finds application in things such as self-driving cars that process high-resolution images on the fly as well as in biomedical technology.
For the deep learning research project with Mayo Clinic Health System, Mohan helps develop ideas that may improve on the base architecture. She also assists with literature reviews for future papers and presentations.
In addition, Mohan is part of a research team involving geography faculty and students, again using deep learning techniques but, in this project, focusing on techniques that are applicable for remote sensing. She helps to develop different deep learning models and to process data for the models, as well as assisting with tasks such as literature reviews.
“In both research projects, we are trying to apply deep learning techniques on problems in a way that has not been done before,” Mohan says of why and how the projects interest and challenge her.
Mohan also was involved in a much different research project, one that she describes as “very fun to work on” that involves collaborating with mathematics faculty. That project, she says, fell within the field of “recreational mathematics” and involved creating knitted versions of different wallpaper group patterns.
“In a broader sense, these research projects together opened my eyes to the many possibilities that can be achieved through technology in different fields,” Mohan says.
An outstanding student, researcher
Mohan’s intelligence and curiosity make her a valuable student to have working in his research lab, says Dr. Rahul Gomes, an assistant professor of computer science and one of Mohan’s research mentors for the past year.
“She is very motivated and eager to learn about machine learning, especially deep learning and how it can be used in image classification and segmentation,” Gomes says of his research assistant.
Recently, Mohan began exploring unsupervised deep learning algorithms to curb the need for using labeled images while training. This branch of research will reduce the dependence of deep learning algorithms on labeled training images and masks as creating them can be a tedious process, Gomes says.
“Her research dives into modifying the underlying framework of deep learning by using experimental layers of TensorFlow instead of using the models which are readily available on the internet,” Gomes says. “For example, in our Mayo Clinic research project on IVC filter detection, Avi is exploring custom image filters in the deep learning model that could have the potential to enhance IVC filters from CT scans based on its shape and reflectance.”
The primary task of this filter is to break blood clots and prevent them from traveling to the heart and lungs. The research team’s goal is to automate the detection of these filters when a patient gets a CT scan, Gomes says, adding that they are using deep learning to accomplish their goal.
Mohan’s contributions to that project and others is invaluable, Gomes says.
“She is an expert in running parallel jobs on the Blugold Center for High-Performance Computing,” Gomes says. “It is great to have her as a research assistant.”
Finding a potential career path
Mohan came to UW-Eau Claire knowing she would pursue her passions for computer science and mathematics, two academic areas that have long interested her.
“I have been interested in computers since I was in middle school, specifically the problem-solving aspects involved in programming,” Mohan says. “I decided to take on an additional math major because I really enjoyed the math classes that I was in, and it had great applications in certain computer science areas.”
It is through her research, however, that Mohan is learning about biomedical technology and the increasingly large and critical role it plays in health care.
“I previously had very little knowledge of the medical field and the technology used,” says Mohan, who will graduate in May 2022. “Working with Mayo has allowed me to gain that knowledge through our collaboration. Actually, doing research has shown me that this is something that I would like to do as a career.”
Mohan’s plans now include earning a Ph.D., which would set her on a path toward a career in research on machine learning.
The research she is doing as a Blugold will help her achieve those goals, Mohan says, noting that her research experience will strengthen her application when she’s applying to graduate programs.
“More importantly, it has taught me how to actually conduct research and how to be an independent learner,” Mohan says. “That will help me throughout grad school and beyond.
“Since technology is an ever-evolving field, I will be able to cope with the changes easily and stay relevant in my field — especially because research teaches you to think outside of the box. Other little things I have picked up like working with others, working outside of my comfort zone and dealing with ‘failures’ will help me in the long run as well.”
While her research experience has strengthened her understanding of deep learning and related concepts, it’s also helping her build other skills that will help her thrive in her life as well as her career.
“On a more personal level, research has taught me that some things take time,” Mohan says. “As a perfectionist, I like being able to finish my task as soon as possible, and as perfectly as possible. That’s not always the case in research. Sometimes it takes you weeks or days to get where you need to be. Sometimes you chase rabbit holes that end up nowhere or methods that do not work as well as you thought they would. This is still something that I struggle with, but I am improving with time.
“And, of course, research is always a collaborative effort, so it has improved my communication skills and my ability to work with others.”