W&M IT plays integral role in undergraduate research
Students detail their appreciation for the impact IT has had on their work
At William & Mary, 80% of undergraduates engage in research. While research interests run the gamut, every project relies on the work of W&M Information Technology in some way. Access to dependable WiFi, communication and collaboration tools like email and Teams, or video-conferencing services are helpful when working on any project, but the students who rely most heavily on IT have been those involved with the Geospatial Evaluation and Observation Lab (geoLab).
Elias Wolman ’22, John Hennin ’22 and Jason Lin ’22 are geoLab undergraduates who have been directly impacted by the work of IT. Their research projects are funded through a Commonwealth Cyber Initiative grant, and look at computational modeling and satellite imagery, as well as creating convolutional neural networks that can classify satellite imagery. To work on each of their respective projects, they needed access to the university’s high performance computing cluster, as well as the expertise of Eric Walter, W&M IT’s director of research computing.
Each student is leading a team of undergraduate researchers, some of which are at other universities. They have had to navigate the complexities of working with the HPC both on campus and off, and Eric has been instrumental in getting them up and running, as well as creating tutorials to help remote students.
For Wolman, a computational and applied mathematics and statistics major, his team is collecting educational data and trying to predict test scores based on satellite imagery in multiple countries. Wolman and his team worked with Phil Fenstermacher, W&M IT’s manager of systems design & architecture, to develop a database and website where they are putting together their data, which is the first of its kind. “Phil was instrumental in helping us get our website up and running, as well as integrating it with the Central Authentication System (CAS),” said Wolman, who has recently spoken with a professor in Japan who is interested in using his team’s data.
Hennin, a data science major, is using convolutional neural networks and importing satellite imagery to predict COVID cases in various countries. “This is especially helpful because at the onset of the pandemic everyone was caught off guard and there was not a lot of infrastructure in place to collect this data and standardize it,” said Hennin. “If you know where the sensitive areas are you can get ahead of the curve and calculate speed of the disease, which would be helpful if we ever experience another pandemic.”
Lin, a computer science and data science major, is focusing on testing a range of attack and defense strategies for data poisoning in the context of satellite imagery and deep learning. He and his team of undergraduate researchers are exploring the susceptibility of satellite imagery to data poisoning attacks.
Hennin says that the opportunity to work with external partners is unique to the William & Mary undergraduate research experience. All three students have had the opportunity to present their research to the Department of Homeland Security and other government entities.
“Student research at William & Mary is unique because we are kind of thrown in and given the opportunity to do some pretty important things and work with advanced technology,” said Wolman. “When we first started this project, I had access to two GPUs (graphics processing unit) and now have access to eight which is kind of unheard of.”
Walter laughs and adds, “We do not baby the undergraduates here. Many other institutions would either require a sponsor to cordon off an area for undergraduates to work, but we let them dive right in and they follow the same rules as everyone else.”
“Getting to use the HPC as a resource has been tremendous and Eric has been a huge help getting everyone set up,” says Lin. “We have 60 students who need access to this. A big part of this grant is also that it’s an educational opportunity to bring in more diversity and students who would not typically be exposed to these machine learning opportunities. So, for people who may be unfamiliar with an HPC environment, Eric has helped by setting us up with a JupyterHub environment where people can log in and view instructional notebooks and helpful graphics.”
“There is no way we could have done these projects without the help of IT,” said Hennin.
As the three students wrap up their projects and prepare for Commencement, they are all getting ready to embark on the next chapter of their lives: their careers. Lin will be starting a role at Capital One after graduation, Hennin is going to be a consultant at West Monroe, a boutique consulting firm in Chicago and Wolman will be traveling to Germany as a Fulbright scholar to teach English.
