HS Student Builds AI Pipeline to Discover 1.5M Lost Space Objects
4 days ago
- #Space Discovery
- #Astronomy
- #Artificial Intelligence
- A high school student from California, Matteo Paz, used AI to identify over 1.5 million previously unknown space objects from NASA's NEOWISE mission data.
- Paz developed an AI model in six weeks that detected faint, variable light sources, revealing objects like quasars, eclipsing binary stars, and supernovae.
- The AI utilized Fourier transforms and wavelet analysis to identify subtle infrared variations, uncovering slow transients and cataclysmic variables.
- Paz's findings, peer-reviewed and published in The Astronomical Journal, will be fully cataloged in 2025, aiding future observations by telescopes like Vera Rubin and JWST.
- Still in high school, Paz is now a paid research assistant at Caltech's IPAC, mentoring others and expanding his AI pipeline for interdisciplinary applications.