Imagine looking up at the night sky and knowing you helped reveal over a million objects no one had seen before. For high school student Matteo (Matthew) Paz, this isn’t imagination, it’s reality. Through innovative work at Caltech, Paz developed an Artificial Intelligence (AI) algorithm that sifted through mountains of data, leading to a groundbreaking teen uses AI space discovery of 1.5 million previously unknown objects.
This remarkable achievement not only expanded the potential of a NASA mission but also resulted in a single-author paper published in the prestigious The Astronomical Journal. It’s a story of curiosity, mentorship, and the incredible power of combining young talent with cutting-edge technology.
Table of contents
- A Spark Ignited: From Stargazing to Caltech
- The Mentor’s Vision: Unlocking NEOWISE’s Potential
- Bridging Worlds: AI Meets Astronomical Data
- Collaboration and Refinement: Building the AI Model
- The AI Space Discovery Breakthrough: 1.5 Million Discoveries and Beyond
- Wider Implications: The AI’s Future Potential
- Looking Ahead: A Young Scientist’s Journey Continues
A Spark Ignited: From Stargazing to Caltech
Matteo Paz began his journey into the cosmos years ago. His mother introduced him to public Stargazing Lectures at Caltech during grade school, nurturing his early fascination with astronomy. That spark grew into a passion and led him to Caltech’s campus in the summer of 2022.
He joined the Caltech Planet Finder Academy, led by Professor of Astronomy Andrew Howard. Here, Paz delved deeper into astronomy and the related computer science that would prove crucial for his future discoveries. He became especially fascinated by the role of AI in space discovery, where artificial intelligence is increasingly used to analyze vast astronomical datasets and detect celestial phenomena beyond human capacity.
It was at the academy that he met his mentor, astronomer and IPAC senior scientist Davy Kirkpatrick.
“I’m so lucky to have met Davy,” Paz shared. He recalled expressing his ambitious goal of publishing a paper within just six weeks, a feat Kirkpatrick encouraged rather than dismissed. “He has allowed an unbridled learning experience. I think that’s why I’ve grown so much as a scientist.”
The Mentor’s Vision: Unlocking NEOWISE’s Potential
Kirkpatrick himself understood the power of mentorship. Inspired by his own ninth-grade teacher who set him on the path to becoming an astronomer, he aimed to pay it forward. “If I see their potential, I want to make sure that they are reaching it,” Kirkpatrick stated.
He saw potential not just in Paz, but also in the vast dataset from NEOWISE (Near-Earth Object Wide-field Infrared Survey Explorer). This now-retired NASA infrared telescope spent over a decade scanning the sky for asteroids. While doing so, it also captured the fluctuating heat signatures of countless distant cosmic objects.
These “variable objects” phenomena like distant quasars, exploding stars (supernovae), or stars orbiting and eclipsing each other, lay hidden within the data. The challenge was immense; the dataset contained nearly 200 billion individual detections. Kirkpatrick initially planned to manually analyze a small patch of sky to demonstrate the data’s potential.
Bridging Worlds: AI Meets Astronomical Data
Matteo Paz, however, had a different approach in mind. Manual sifting wasn’t his plan. His background, combining coding, theoretical computer science, and advanced mathematics (having completed AP Calculus BC in eighth grade via Pasadena Unified School District’s Math Academy), positioned him perfectly to tackle the challenge with AI.
He recognized that the massive, orderly NEOWISE dataset was ideal for training a machine-learning model. Paz set out to develop an AI algorithm capable of analyzing the entire dataset automatically. It flags potential variable objects far faster and more comprehensively than any human could. This represented a significant AI algorithm application in astronomy.
Collaboration and Refinement: Building the AI Model
During that initial six-week summer program, Paz began drafting the AI model, showing early promise in the realm of AI space discovery. His collaboration with Kirkpatrick was key, providing the necessary astronomical context. “Every meeting with Davy is 10% work and 90% us just chatting,” Paz noted, highlighting the supportive and engaging nature of their mentorship.
Kirkpatrick also connected Paz with other Caltech experts in machine learning and variable objects, like Shoubaneh Hemmati and Matthew Graham. They learned that NEOWISE’s specific observation pattern made it difficult to systematically detect objects that changed very quickly or very slowly. This understanding helped refine the AI’s approach. The work continued beyond the summer, with Paz even mentoring other students in 2024 while refining his model.
The AI Space Discovery Breakthrough: 1.5 Million Discoveries and Beyond
The result of this persistent effort is stunning. Paz’s refined AI model processed the entire raw NEOWISE dataset. Trained to spot subtle differences in infrared measurements over time, the algorithms successfully flagged and classified 1.5 million potential new variable objects.
This incredible catalog of cosmic variables, detailed in Paz’s Astronomical Journal article, is a treasure trove for astronomers worldwide. Paz and Kirkpatrick plan to publish the complete catalog in 2025, offering insights into how diverse celestial bodies evolve over years. This teen uses AI space discovery represents a massive contribution to the field.
Wider Implications: The AI’s Future Potential
The impact of Paz’s work extends beyond identifying these specific space objects. “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 explained.
He sees potential applications in analyzing stock market charts, where information also arrives as a time series with critical periodic patterns. It could even be adapted to study atmospheric phenomena like pollution, where daily and seasonal cycles are crucial factors. The core AI technique has broad relevance.
Looking Ahead: A Young Scientist’s Journey Continues
Matteo Paz’s story is far from over. While still finishing high school, he is now officially a Caltech employee, working for Kirkpatrick at IPAC. IPAC is the data hub for NEOWISE and other major space missions, a fitting place for someone who has already unlocked so much from their archives.
His journey underscores the incredible potential that arises when young talent is nurtured through mentorship and given access to challenging, real-world problems. The teen uses AI space discovery led by Paz is a powerful testament to how fresh perspectives and modern tools like AI can revolutionize scientific exploration, revealing secrets hidden in plain sight within vast datasets. The universe just got a little bit bigger, thanks to the vision of a dedicated high school student.
| Latest From Us
- FantasyTalking: Generating Amazingly Realistic Talking Avatars with AI
- Huawei Ascend 910D Could Crush Nvidia’s H100 – Is This the End of U.S. Chip Dominance?
- Introducing Qwen 3: Alibaba’s Answer to Competition
- Google DeepMind AI Learns New Skills Without Forgetting Old Ones
- Duolingo Embraces AI: Replacing Contractors to Scale Language Learning
