A California high school student has identified over 1.5 million previously unknown space objects by developing an artificial intelligence model to analyze a decade's worth of archived NASA data. The discovery, made during a summer program, showcases how modern AI can unlock new insights from existing astronomical surveys.
Matteo Paz, a student from Pasadena, created a machine learning pipeline to sift through nearly 200 billion observations from the NEOWISE telescope. His work has since been recognized in a scientific journal, highlighting the significant contributions young researchers can make to the field of astronomy.
Key Takeaways
- Matteo Paz, a high school student, discovered more than 1.5 million unidentified space objects.
- He developed a custom artificial intelligence (AI) model to analyze archived data from NASA's NEOWISE telescope.
- The project was part of the Planet Finder Academy, a summer program at the California Institute of Technology (Caltech).
- The AI successfully identified faint and variable objects, such as quasars and supernovae, that were previously missed in the massive dataset.
A Summer Program with Cosmic Implications
In the summer of 2022, Matteo Paz participated in the Planet Finder Academy at Caltech. This program connects high school students with real-world scientific challenges. Paz was assigned to work with a vast and underutilized dataset from NASA's Near-Earth Object Wide-field Infrared Survey Explorer, better known as NEOWISE.
Under the guidance of Davy Kirkpatrick, an astronomer at Caltech’s Infrared Processing and Analysis Center (IPAC), Paz was tasked with exploring the NEOWISE archive. The telescope, originally launched in 2009, had accumulated over a decade of infrared observations of the entire sky, resulting in a dataset containing almost 200 billion lines of data.
What is NEOWISE?
The NEOWISE telescope was designed to track near-Earth objects like asteroids and comets. However, in its mission to scan the sky, it also captured data on a wide range of distant cosmic phenomena. This created an enormous public archive rich with potential discoveries, but its sheer size made manual analysis impractical for scientists.
The immense scale of the data presented a significant obstacle. Analyzing such a volume of information with traditional methods would be incredibly time-consuming. This challenge prompted Paz to seek a more innovative and efficient solution.
Developing an AI for Astronomical Discovery
Drawing on his skills in programming and mathematics, Paz decided to build a custom artificial intelligence model. Over a period of just six weeks, he developed a complete machine learning pipeline designed to automate the search for specific types of celestial objects within the NEOWISE data.
The AI was specifically trained to detect variable light sources. These are objects in space whose brightness changes over time. Such fluctuations can indicate interesting astronomical events, including distant quasars, eclipsing binary star systems, or the explosive deaths of stars known as supernovae.
The Power of AI in Data Analysis
Paz's AI model processed the massive dataset far faster than a human team could. It was designed to recognize subtle patterns of dimming, flickering, or pulsing light that are often too faint or infrequent for conventional software tools to flag, effectively revealing objects that had been hidden in plain sight within the data.
His mentor, Davy Kirkpatrick, noted the immediate effectiveness of the tool. In a statement to Phys.org, Kirkpatrick said,
“The model showed promise almost immediately.”As Paz continued to refine the algorithm, its ability to pinpoint these previously unseen objects grew, leading to the identification of the 1.5 million new candidates.
Unlocking Secrets in Archived Data
The success of Paz's project demonstrates a crucial principle in modern science: old data can yield new discoveries when viewed through a modern lens. The NEOWISE archive, while years old, held secrets that were inaccessible until a new method was applied to it.
The AI used advanced mathematical techniques, including Fourier transforms and wavelet analysis, to study signals that change over time. This allowed it to overcome a key limitation of the NEOWISE survey's sampling rate, which sometimes missed objects that changed in brightness too slowly or too briefly.
What Kind of Objects Were Found?
The variable objects identified by the AI are of great interest to astronomers. They fall into several important categories:
- Quasars: Extremely luminous and distant galactic nuclei powered by supermassive black holes.
- Eclipsing Binaries: Systems of two stars orbiting each other, where one periodically passes in front of the other from our perspective, causing a dip in brightness.
- Supernovae: Powerful and brilliant explosions of stars, which are crucial for understanding stellar evolution.
- Slow Transients: Events that change in brightness over long periods and are difficult to detect without sustained observation.
By identifying these objects, Paz's work provides a new catalog for astronomers to study, potentially leading to a better understanding of the life cycles of stars and the behavior of galaxies.
Inspiring a New Generation of Scientists
Matteo Paz's achievement is more than just a scientific finding; it serves as an example of how accessible tools and public data can empower young researchers. His project highlights the growing importance of skills in data science and machine learning for the future of space exploration.
The project underscores the value of mentorship programs like the Planet Finder Academy, which provide students with access to real scientific resources and expert guidance. By tackling a genuine research problem, Paz was able to make a tangible contribution to the field of astronomy before even starting college.
The AI pipeline he created may continue to be used to analyze other large astronomical datasets, potentially uncovering even more cosmic secrets hidden in archives around the world. This work proves that with curiosity, creativity, and the right tools, groundbreaking discoveries are within reach for anyone, regardless of age or experience.
