An American schoolboy discovered 1.5 million previously unknown objects in space with the help of AI
An 18-year-old high school student from the United States developed an algorithm based on artificial intelligence, which made it possible to detect 1.5 million hitherto unknown objects in space. He presented the results of his work in a study published in The Astronomical Journal, informs Phys.org.
Matteo Paz’s interest in astronomy arose in elementary school, after his mother took him to an open lecture on star observations at the California Institute of Technology. The teacher identified the boy’s scientific potential and recommended that he take additional courses to prepare for college.
In 2022, Matteo returned to school to study astronomy and related computer science for the summer. During his studies, he approached Senior Research Fellow Devi Kirepatrick to be his research supervisor. Kirepatrick invited the student to explore a small area of space for variable objects — phenomena that are difficult to record. These include, in particular, exploding stars, those that eclipse each other, and quasars — superluminous nuclei of galaxies.
However, Matteo was reluctant to manually sift through the existing data to discover new objects. Since, in addition to astronomy, he was fascinated by artificial intelligence, he decided to use its capabilities in his scientific work.
Using his knowledge of mathematics and programming, he created a machine learning model that allowed him to analyze large amounts of data to discover new objects in space. The mentor introduced the young man to other astronomers who had experience in using similar technologies.
In 2024, Matteo refined his algorithm by adapting it to analyze raw data from the WISE infrared orbiting telescope.
The algorithm was trained to recognize subtle changes in radiation, allowing it to detect and classify 1.5 million potentially new objects. A complete catalog of these bodies, Matteo, together with Davy Kirepatrick, is planned to be published in 2025.
“My model can be used for other space-time studies in astronomy, and perhaps anything else that comes in a time format”, – explained the schoolboy.
Algorithms also have promise for analyzing stock market charts and atmospheric phenomena, including air pollution, he said. For his achievement, he won the top prize of the prestigious Regeneron Science Talent Search competition for high school students, receiving a prize of $250,000.
Matteo is currently finishing school and working at the California Institute of Technology in the space mission data archiving and analysis department.
