Matteo Paz maps 1.5 million unknown space objects using AI
Updated
Updated · ECOticias · May 4
Matteo Paz maps 1.5 million unknown space objects using AI
10 articles · Updated · ECOticias · May 4
The Pasadena High School student used NASA's NEOWISE infrared archive, with Caltech mentor Davy Kirkpatrick, to identify previously unrecognised variable objects in a peer-reviewed 2024 study.
His VARnet system analysed nearly 200 billion detections over 10.5 years, classifying about 1.9 million infrared variable objects into 10 categories with reported high speed and accuracy.
The work won Paz the $250,000 Regeneron Science Talent Search and highlights how machine learning can unlock old scientific archives, with possible applications in tracking Earth patterns such as pollution.
Could VARnet's AI breakthroughs in astronomy transform how we discover hidden patterns in fields like climate science or financial markets?
As AI unlocks cosmic secrets, will its soaring energy demands force a reckoning between scientific progress and environmental sustainability?
High School Student Matteo Paz Discovers 1.5 Million New Celestial Objects Using AI Algorithm VARnet
Overview
In 2025, high school student Matteo Paz developed the advanced AI algorithm VARnet, which combined innovative techniques to analyze NASA's NEOWISE infrared data. This breakthrough enabled the discovery of 1.5 million previously unknown variable celestial objects, leading to the creation of the VarWISE catalog. Caltech researchers quickly used this catalog to identify new binary star systems and calculate exoplanet masses. Paz's work earned him the top Regeneron Science Talent Search prize and recognition from NASA. Mentored by Dr. Kirkpatrick, Paz transitioned into a research role and began mentoring others. His success highlights how AI and open data empower young scientists to make major contributions while raising important ethical discussions about AI in research.