Khoury Ph.D. Candidate Wins Prestigious NSF Fellowship

Laura South, a Ph.D. candidate studying data visualization at Khoury, was recognized by the NSF Graduate Research Fellowship Program (GRFP)

The NSF Graduate Research Fellowship Program (GRFP) – the oldest graduate fellowship of its kind – recognizes and supports outstanding graduate students in NSF-supported science, technology, engineering, and mathematics disciplines pursuing research-based master’s and doctoral degrees. Notable recipients include numerous Nobel Prize winners as well as Google founder Sergey Brin and the U.S. Secretary of Energy, Steven Chu. Added to the list of recent fellows is Laura South, a Ph.D. candidate studying data visualization at the Khoury College of Computer Sciences.

Professor Michelle Borkin, South’s advisor, played a significant role in encouraging South to apply for the fellowship. “She also helped me revise my application and consider the logistics of my proposed research plan,” South explains. “It was very helpful to have a more experienced researcher on hand to answer my questions and give feedback on my work.”

South received her notification email on the fellowship early in the morning; it was the first thing she saw when she woke up. She went straight to a scheduled meeting with Borkin who “was thrilled” by the news. South describes the moment: “Everyone in my lab was so kind and congratulated me on receiving the fellowship.”

South works in data visualization, specifically studying computer science and data science’s effects on how humans share information, as well as artificial intelligence’s effects on society, especially in marginalized communities. She plans to use the fellowship to further her study of computer vision, helping to make computers view things the way humans view them.

One focus is identifying media that could pose harm to photo-sensitive conditions including epilepsy, such as a set of patterns in a video that could be a risk for a seizure. She came up with the idea when she heard about someone being sent a GIF by a Twitter direct message to deliberately trigger a photo-sensitive seizure. She likes the idea that her research “could directly help real people, something that can be hard to find in the deeply technical field.”

Originally from Arvada, Colorado, South graduated from Colorado State University (CS and Statistics) in 2018 and began her Ph.D. program at Northeastern the same year. “Coming from a small town,” she says of Boston, “there’s just so much going on – universities in the area with cool speakers, workshops and activities. I go and sit in on lectures in the music department and law school for example. It’s wild that I can walk down the street and be in a different world class university.”

As an undergrad at CSU, South began her work by identifying photosensitive risk factors for her undergraduate honors thesis. She wrote a Python program that took in GIFs frame by frame, and detected flashes, stripes, as well as reviewing a bunch of medical papers that touched on what factors cause a photo-sensitive harm. Her NSF-funded work will build on this initial work, specifically involving “automatically detecting photosensitive risk factors in online media and taking steps to protect users with photosensitivity from those hazardous elements so they don’t suffer from seizures or other symptoms.”

“Given how important the web has become in our daily lives, it is critical that we make it an accessible and safe experience for all,” says Professor Cody Dunne. “Laura’s work may one day protect people against a truly dangerous and personal attack.” Dunne co-leads the lab Visualization @ Khoury with South’s advisor.

Since she has just finished the first year of her Ph.D. program, South isn’t entirely sure what her dissertation topic will be, but she stresses that she “would love to incorporate my NSF fellowship work with photosensitive epilepsy in online environments into my dissertation in some way.”

She continues, “I will be starting the machine learning approach in the fall when I begin the fellowship.” In the past, she took a rule-based approach, but “now I will use machine learning which has the potential to be far more effective.”

She is currently working on a way to visualize political debates. Debates, she says, “affect everyone and the way power is achieved in debates, but the way we expose ourselves to them is watching them live. That’s not something everyone has the time and ability to watch. I use technology to make debates more accessible and highlight the issues interesting to people who are affected by the outcomes of political debates but might not have the time to really delve deep to understand debates.”

Since she loves interacting with students, South hopes to continue in academia after receiving her Ph.D., with her ideal position offering a balance between research and teaching responsibilities. “If anything,” she says, “the fellowship has strengthened my desire to continue with my research in an academic setting!

This article was written with additional reporting by Ben Hosking.