A research team led by scientists from the University of Keele in the UK has developed an artificial intelligence (AI) technology that assists astronomers in estimating the ages of stars more accurately than current methods, which rely on inferring the chemical composition of stars.
Machine learning is a branch of artificial intelligence that aims to improve the quality of results based on learning from vast amounts of data. For example, machine learning algorithms are trained on tens or hundreds of thousands of dog and cat images to distinguish between a dog’s shape and a cat’s. The AI software then searches for statistical patterns in which cats resemble each other but differ significantly from dogs.
Afterwards, artificial intelligence converts these patterns into probabilistic rules to analyze new images it has not seen before. With each new photo added to its database, the AI improves its performance for future instances. This resembles how humans learn anything in our lives since childhood.
The researchers utilized data from the Gaia telescope on over 6,000 stars (European Space Agency).
According to a new study announced at the National Astronomy Meeting 2023 at Cardiff University in the UK, the researchers based their work on a scientific hypothesis called “exhaustion of lithium.” During a star’s lifetime, the increasing heat and pressure in its core change the chemical composition of its outer layers. Specifically, the amount of lithium decreases over time.
Consequently, this idea can be used to determine the age of a star by measuring the amount of lithium present. This is achieved by studying the light emitted from distant stars, where lithium leaves a distinctive signature in the spectrum of that light. This signature can be read using telescopes such as Gaia, a space observatory launched by the European Space Agency in 2013 to create a three-dimensional catalogue of nearly one billion stars.
In a press release from the Royal Astronomical Society, the researchers from this team utilized data from Gaia on over 6,000 stars to train the AI software. By examining the data, the software establishes a model of the relationship between the star’s temperature, the measured lithium abundance, and its age.
The researchers hope this new mechanism will enhance astronomers’ accuracy in estimating the ages of stars, mainly since this type of AI-based technology is still in its early stages, with developments progressing rapidly daily.
