The emergence of AI has been hailed as a societal game-changer, offering a myriad of possibilities to enhance nearly every aspect of our lives. Now, astronomers are leveraging AI to measure the expansion of our universe.
Two recent studies led by Maria Dainotti, a visiting professor at UNLV’s Nevada Center for Astrophysics and assistant professor at the National Astronomical Observatory of Japan (NAOJ), have integrated multiple machine learning models to enhance the precision of distance measurements for gamma-ray bursts (GRBs)—the most luminous and violent explosions in the cosmos.
GRBs, which release an immense amount of energy equivalent to that emitted by our sun over its entire lifetime within a few seconds, can provide valuable insights into the oldest and most distant stars in the universe. However, due to technological limitations, only a small fraction of known GRBs possess the necessary observational characteristics to accurately determine their distances.
Dainotti and her teams combined data from NASA’s Neil Gehrels Swift Observatory with advanced machine learning techniques to overcome these limitations and estimate the proximity of GRBs with unknown distances more precisely. By discerning the locations of GRBs, scientists can gain deeper insights into the evolution of stars over time and the frequency of GRB occurrences in specific spatial and temporal contexts.
“This research represents a significant advancement in both gamma-ray astronomy and machine learning,” remarked Dainotti. “Continued research and innovation hold the promise of even more reliable results, offering solutions to some of the most fundamental cosmological inquiries regarding the origins and evolution of our universe.”
In one study, Dainotti and Aditya Narendra, a doctoral student at Poland’s Jagiellonian University, employed various machine learning methods to accurately gauge the distance of GRBs observed by the Swift UltraViolet/Optical Telescope (UVOT) and ground-based telescopes, such as the Subaru Telescope. These measurements were based solely on other properties of GRBs unrelated to distance.
In another study, Dainotti and international collaborators utilized machine learning to measure GRB distances using data from NASA’s Swift X-ray Telescope (XRT) afterglows, focusing on long GRBs. This approach, which combines multiple machine learning methods to enhance predictive accuracy, significantly expands the number of known distances for this type of burst.
Furthermore, a third study led by Vahé Petrosian from Stanford University and Dainotti utilized Swift X-ray data to address intriguing questions about GRB formation. Their findings suggest that the rate of long GRBs at small distances may not align with the rate of star formation, hinting at alternative mechanisms such as the fusion of dense objects like neutron stars.
With support from NASA’s Swift Observatory Guest Investigator program, Dainotti and her colleagues are working to make their machine learning tools publicly accessible through an interactive web application, further democratizing access to cutting-edge astronomical research.
