High school student stuns astronomers by revealing 1.5 million previously unknown objects in space using AI

The research by the high school student has opened doors to unexpected space explorations.
PUBLISHED APR 17, 2025
(L) STS 2025 First Place Winner Matteo Paz talks about his work (Photo by Society for Science / Youtube); (R) Humans are using laptops and computers to interact with AI (Photo by Wanan Yossingkum / Getty Images)
(L) STS 2025 First Place Winner Matteo Paz talks about his work (Photo by Society for Science / Youtube); (R) Humans are using laptops and computers to interact with AI (Photo by Wanan Yossingkum / Getty Images)

There is no set standard for intelligence and innovation, nor is there an age limit for wonderful discoveries. This is an apt fit for a local high school student, Matteo (Matthew) Paz, who revealed 1.5 million unknown secrets of the universe, as per Caltech. The Pasadena High School senior developed an AI algorithm that led to these discoveries. This was presented at the 2025 Regeneron Science Talent Search, run by the Society for Science. This algorithm can be useful in space exploration and for other astronomers and astrophysicists to adapt for their own research.

General view of California Institute of Technology on September 22, 2020 in Pasadena, California. (Photo by AaronP/Bauer-Griffin/GC Images)
General view of California Institute of Technology on September 22, 2020 in Pasadena, California. (Photo by AaronP/Bauer-Griffin/GC Images)

Paz was awarded first place, along with $250,000, for his innovation that was published in a single-author paper. The study described the AI algorithm that was capable of sifting through multiple spatial datasets. The paper was published in The Astronomical Journal. This discovery began with the idea of the variable objects present in NASA’s observations. The NASA telescope detected heat signatures with varied intensities of distant cosmic objects. However, their data was not cataloged due to their bulk, as they were unnecessary inputs during important observations.

IC 63, AKA The Ghost Nebula, is about 550 light-years from Earth. The nebula is classified as both a reflection nebula and an emission nebula. (Photo by Javier Zayas Photography / Getty Images)
IC 63, AKA The Ghost Nebula, is about 550 light-years from Earth. The nebula is classified as both a reflection nebula and an emission nebula. (Photo by Javier Zayas Photography / Getty Images)

To record this data, astronomers were dependent on NEOWISE (Near-Earth Object Wide-field Infrared Survey Explorer). This retired infrared telescope had scanned the entire sky to search for asteroids and other objects for more than 10 years. If the NEOWISE team could go through the records and catalog the required objects, it would be easier for astronomers to observe them for changes. This was of interest to Astronomer and IPAC senior scientist Davy Kirkpatrick, who was Paz’s mentor. He decided to form a team to scan portions of the sky and report anomalies.

 An artificial space object is on its way, in orbit, around the blue planet. (Photo by OsakaWayne Studios / Getty Images)
An artificial space object is on its way, in orbit, around the blue planet. (Photo by OsakaWayne Studios / Getty Images)

"At that point, we were creeping up towards 200 billion rows in the table of every single detection that we had made over the course of over a decade," Kirkpatrick stated. "So my idea for the summer was to take a little piece of the sky and see if we could find some variable stars. Then we could highlight those to the astronomic community, saying, 'Here's some new stuff we discovered by hand; just imagine what the potential is in the dataset,'" he added. However, Paz had a different idea in mind and wanted to subtract the task’s manual labor from the equation.

A programmer working on coding from multiple system setups (Photo by Emilija Manevska / Getty Images)
A programmer working on coding from multiple system setups (Photo by Emilija Manevska / Getty Images)

Paz’s innovation involved his interest in AI from a school elective that combined coding, theoretical computer science, and formal mathematics, as per Phys Org. As AI is best trained on large datasets, he created an algorithm that would analyze the entire dataset and catch any potential variable objects. He drafted the model for the machine learning technique by consulting his mentor about relevant astronomy and astrophysics. As NEOWISE data did not consider minute detections, Paz refined his model to process all of the observed raw data and analyze the results.



 

This detection of minute changes in the telescope's infrared measurements allowed the algorithm to identify 1.5 million new objects. Paz and Kirkpatrick plan on publishing the entire dataset this year. "The model I implemented can be used for other time domain studies in astronomy, and potentially anything else that comes in a temporal format," Paz commented. With a prize-winning discovery in hand, the high school senior is already working for his mentor at Caltech in his first paid job. His innovation is believed to be significant to astronomers across the world.

MORE STORIES

A re-analysis of data from the defunct Cassini probe confirms the moon's subsurface ocean is more chemically rich than previously thought.
17 hours ago
According to preliminary analysis, the newly identified exoplanet TOI-4507 b is challenging fundamental theories governing how celestial bodies evolve.
17 hours ago
A new academic investigation suggests that previously unrecognized observations of the significant galactic supernovae SN 1006 and SN 1181 may be preserved within ancient Arabic writings.
19 hours ago
Dark matter is an invisible, enigmatic substance that makes up about 85% of the universe's total mass yet remains undetectable because it does not interact with light.
1 day ago
New satellite data confirms that the South Atlantic Anomaly, a region of weakened magnetic field, has expanded since 2014 by an area equivalent to nearly half of continental Europe.
1 day ago
The intricate magnetic lifecycle of Iota Horologii (ι Hor), a 600-million-year-old Sun-like star, was successfully mapped by scientists at the Leibniz Institute for Astrophysics Potsdam.
2 days ago
Using ALMA, an international team achieved the most comprehensive mapping of magnetic fields yet, using the ALMA telescope to survey 17 massive star-forming regions.
2 days ago
The study presents a novel technique that utilizes the dark, shadow-like regions in black hole images taken by 'Event Horizon Telescope' to hunt for the invisible material comprising about 85% of all matter.
2 days ago
Researchers argue that 3I/ATLAS, a body roughly seven miles in diameter, is a lithified clastic fragment ripped from a sedimentary basin on a distant exoplanet that may once have been capable of supporting life.
2 days ago
This model predicts that the universe will reach its maximum size in approximately 11 billion years, then rapidly start contracting.
5 days ago