A team of researchers at the University of Georgia has confirmed evidence of a previously unknown extrasolar planet and used machine learning tools to detect it.
A recent study by the team showed that machine learning can examine protoplanetary disks, the gas around newly formed stars, to correctly determine whether exoplanets are present.
The newly published findings represent a first step toward using machine learning to identify previously overlooked exoplanets.
“We used traditional methods to identify the planets, but our model told us to run those simulations and showed us exactly where the planets were,” said UGA Franklin College. said Jason Terry, a PhD student in the Department of Physics and Astronomy at the College of Arts and Sciences. First author of the study.
“Applying the model to a series of older observations identified a disk that had already been analyzed but was not known to contain planets. As with previous discoveries, we performed simulations of the disk. and discovered that planets could re-form, making observations.”
According to Terry, the model suggests the presence of a planet, indicated by several images that strongly highlight certain regions of the disk that were found to have characteristic signatures of a planet. Anomalous deviations in the velocity of the gas near the planet.
“This is an incredibly exciting proof-of-concept. From previous work, we knew that machine learning could be used to find known forming exoplanets.” “Now we are confident that we can use it to make completely new discoveries.”
This finding highlights the power of machine learning to enhance the work of scientists. Utilizing artificial intelligence as an additional tool to increase researchers’ precision and save time more efficiently when engaging in enormous endeavors such as deep space exploration.
The model was able to detect signals in data that people had already analyzed. They found something previously undetected.
“This demonstrates the ability of our models, and machine learning in general, to quickly and accurately identify important information that people might miss. It can dramatically speed up theoretical insight,” said Terry. “It only took about an hour to analyze that entire catalog and find strong evidence of a new planet at a particular location, so as the dataset grows ever larger, more important locations for these types of techniques become apparent. I think there is.”
