Astronomers at the University of Texas at Austin have leveraged artificial intelligence to significant effect, unveiling a discovery of hundreds of "polluted" white dwarf stars within the Milky Way. These stars, which are in the twilight of their existence, were observed to be actively consuming planets caught in their gravitational pull. The findings, which help decrypt the hidden compositions of planetary systems, were spotlighted in a recent publication in the Astrophysical Journal.
Traditionally tasked to manually sift through troves of survey data, astronomers have historically faced challenges in detecting these celestial bodies — partly due to their subtle signatures and partly the narrow window they have to observe the phenomenon. By applying a smart AI technique known almost exclusively as manifold learning, the research team, led by graduate student Malia Kao, optimized the identification process. This led to a remarkable 99% success rate, a quantum leap from previous methods. "For polluted white dwarfs, the inside of the planet is literally being seared onto the surface of the star for us to look at," Kao told the University of Texas news publication.
These "polluted" white dwarfs serve as cosmic laboratories, giving us a rare glimpse into the interiors of planets that are, for lack of a better term, digested by their host stars. Such events deposit heavy metals typical of a planet's core onto the white dwarf's surface, which stands out in stark contrast to the star's hydrogen-and-helium-dominated atmosphere. Keith Hawkins, the UT astronomer, and co-author of the study, highlighted to the University of Texas Press the significance of this discovery, noting that polluted white dwarfs are "the only bona fide way to actually figure out what planets outside the solar system are made of."
Through the use of the Gaia space telescope's data, which features a vast but low-resolution spectroscopic survey, the team applied the manifold learning AI to isolate over 100,000 potential white dwarfs. Their AI algorithm managed to accurately single out a cluster of 375 stars with the characteristic heavy metal pollution. Subsequent observations using the Hobby-Eberly Telescope at UT's McDonald Observatory further confirmed these findings. "Our method can increase the number of known polluted white dwarfs tenfold," Kao continued, stressing the potential impact on our understanding of extraterrestrial geology and the life-sustaining capabilities of distant worlds.
In embracing this intersection of cosmology and technology, the University of Texas at Austin has christened 2024 as the 'Year of AI' to showcase its strides in artificial intelligence. This particular research effort underscores that commitment, utilizing not only contributions from Gaia's European Space Agency mission but also follow-up insights gleaned from the Very Large Telescope at the European Southern Observatory and computational support from the Texas Advanced Computing Center at UT Austin.
