UMD astronomy Ph.D. student Shaniya Jarrett uses artificial intelligence to decode gravitational waves, some of the universe’s most elusive signals.
When gravitational waves were first detected in 2015, they opened an entirely new window into the universe—allowing scientists to “hear” cosmic events like black hole collisions that were previously undetectable.
University of Maryland astronomy Ph.D. student Shaniya Jarrett works to explore how artificial intelligence could help process gravitational-wave signals more efficiently—making gravitational-wave science faster, more accurate and more accessible to researchers.
Working with Astronomy Professor Coleman Miller, Jarrett examined one of astrophysics’ most persistent challenges: the time-consuming process of generating gravitational wave templates. These templates are theoretical models that predict what specific cosmic events, such as the collision of two black holes, should look like when detected as gravitational waves.
“Scientists need thousands of these templates to compare against real space observations. We use them to identify what actually created each detected signal,” Jarrett explained. “The problem we have now is that if we want to be highly accurate, it’ll take an extremely long time. But on the other hand, if we do want to be faster, we end up with less precise information.”
Jarrett’s goal was to find a happy medium—to churn through vast quantities of data faster without losing accuracy and precision. She explored the use of transformer neural networks, the same artificial intelligence technology behind large language models like ChatGPT, to develop these synthetic gravitational waves and templates that are more reliable than existing alternatives.
“It made sense to me to turn to artificial intelligence as a tool, seeing as its strength is working with huge amounts of data to find patterns,” Jarrett said. “I wanted to see if we could find a way to generate synthetic gravitational wave signals and make something that’s easier for us to work with, so we can study them in theory and create better templates for when we encounter real gravitational waves in practice.”
From sci-fi fan to space detective
Just a few years ago, Jarrett would never have imagined herself studying astrophysics or listening for the universe’s most secret sounds.
“Growing up, I was a big fan of science fiction, but I never considered any of that as a career,” she said. “After graduating from high school, I started college with the hopes of being involved with a job in international affairs, which is a very different type of profession. But I eventually fell in love with space and ended up doing the work I do today, even if it wasn’t exactly the most straightforward path.”
Jarrett earned her bachelor’s degree in astrophysics from the University of Colorado Boulder in 2020, and then she became a satellite test engineer for General Atomics. There, she got her first taste of working on complex software and hardware systems, conducting functional testing for million-dollar satellites like the company’s GAzelle spacecraft—a satellite that carries instruments that help monitor wildlife, ocean buoys and environmental conditions on behalf of the U.S. National Oceanic and Atmospheric Administration and French National Centre for Space Studies (CNES).
To take her interest in space to the next level, Jarrett earned her master’s degrees in physics at Fisk University through the Fisk-Vanderbilt Master's-to-Ph.D. Bridge Program in 2024.
During the program, Jarrett established a unique education program called AstroBeats. Middle school students learned how to convert real NASA data into music using sonification software, culminating with a professional ensemble performing the students’ compositions live. The students and their families listened to the science happening through sounds, like the dips in brightness when an exoplanet transits across a star. For Jarrett, it was an opportunity to combine her love of science with music and mentorship, something that continues to drive her passion for scientific outreach at UMD.
“I specifically target the middle school age range because it’s typically when most girls lose their science identity,” Jarrett explained. “Preventing that loss and making sure that they get the resources they need to build confidence and experience in STEM is critical.”
Building a better data pipeline today for tomorrow’s discoveries
Jarrett’s innovative work with space sounds and transformer neural networks naturally transitioned into her current research, which focuses on improving the data pipelines needed for future space missions, including NASA’s Laser Interferometer Space Antenna (LISA) mission set to launch in 2035.
At NASA’s Goddard Space Flight Center, Jarrett works with NASA research astrophysicist Scott Noble to design systems that can analyze light patterns from distant galaxies to distinguish between light curves with supermassive black holes versus those with two black holes in the process of merging.
One of the challenges LISA will face in detecting gravitational waves from binary supermassive black holes is accurately identifying which galaxy hosts them. Miller compared the gravitational waves as “a cacophony rather than a finely tuned instrument, making it extremely difficult to identify the subtle signatures of binary black holes.”
That's where Jarrett's machine learning expertise comes in. She's applying AI to simulated active galactic nuclei (galaxies whose centers are powered by matter falling into supermassive black holes)—some with single supermassive black holes and some with binary pairs—and training her systems to detect subtle differences that human observers might miss.
The project represents the cutting edge of multi-messenger astronomy—combining gravitational wave detection with traditional light-based observations to create a more complete picture of cosmic phenomena. If Jarrett's research succeeds, it could one day allow X-ray or optical detection telescopes to focus on more specific regions of sky where black hole mergers may occur, dramatically improving the efficiency of gravitational wave detection and follow-up observations.
As NASA’s LISA mission prepares to detect gravitational waves billions of light-years away, Jarrett hopes that when those cosmic whispers reach Earth, we’ll be ready to listen—and understand what they're telling us about the universe's most violent and energetic events.
“Once we can pinpoint which galaxies are most likely to host upcoming mergers, we'll be able to really open more doors,” Jarrett said. “Scientists will not only deepen their understanding of these systems, but also better coordinate traditional observations with upcoming gravitational-wave detections.”
