A high school student collaborating with researchers at Caltech has created an exciting AI model that discovered 1.5 million new objects in space, tapping into the vast archives of NASA’s old NEOWISE telescope. This breakthrough highlights a whole new way for astronomers to study the cosmos and was even featured as a single-author paper in The Astronomical Journal.
A Journey from Stargazing to Discovery
Meet Matteo Paz, a high school student who has always loved astronomy. His passion sparked at local stargazing lectures and grew when he joined the Planet Finder Academy in 2022. Under the mentorship of Davy Kirkpatrick, a senior scientist at IPAC, Paz dove into an ambitious project aimed at sifting through NEOWISE’s extensive data archive.
“I remember telling Davy I wanted to write a paper, which felt like a big goal. He didn’t discourage me,” said Paz, recalling his early days in the academy.
Kirkpatrick recognized Paz’s potential and encouraged him to explore data from NEOWISE, which had collected information for over a decade. Surprisingly, the telescope’s dataset, filled with nearly 200 billion rows of observations, contained signals from various celestial objects, including quasars and eclipsing binaries. However, until now, most of this data had gone unexplored for dynamic changes.
Digging Into the Unused Data
While Kirkpatrick suggested focusing on a selected section of the sky, Paz opted for a more innovative approach. With a solid background in AI and math from his studies in the Pasadena Unified School District’s Math Academy, he figured out how to automate the search for new variable stars.
“His schoolwork helped him see the challenge in a fresh way. He even studied AP Calculus BC in eighth grade,” the article noted. Paz’s advanced understanding of machine learning allowed him to develop an algorithm that flagged objects by detecting tiny changes in brightness in the infrared data over time.
Innovative Applications for AI
Over two years, Paz and the team refined this AI model, ultimately enabling it to process the entire NEOWISE dataset. As a result, they discovered 1.5 million never-before-seen celestial objects. These find could provide scientists with valuable insights into cosmic events like exploding stars and binary systems, broadening the scientific applications of NEOWISE well beyond its initial goals.
“The model I created can extend beyond astronomy. It could help analyze stock market data or study pollution trends, where time series analysis is critical,” Paz explained. This adaptability shows that the techniques developed could apply to a range of fields.
The model, dubbed the “Submillisecond Fourier and Wavelet-based Model,” is remarkable, specializing in detecting fleeting or gradually changing events that traditional methods often miss.
Passing on Knowledge to Future Generations
After his groundbreaking work in 2022, Paz returned to the Planet Finder Academy in 2024, this time mentoring fellow high school students. Today, he is even working as a paid employee at Caltech under Kirkpatrick’s guidance.
“Every meeting with Davy feels like a mix of work and casual chats about science,” Paz shared, emphasizing the supportive atmosphere.
Kirkpatrick, reflecting on his own journey as a mentor, stated, “If I see their potential, I want to help them reach it.”
The complete catalog of Paz’s findings will be published in 2025, paving the way for new astronomical research. This story showcases the amazing power of AI and the profound impact of mentoring, proving that even young students can make significant discoveries that push the boundaries of our understanding of the universe.
With astronomy changing quickly, this innovative research could inspire similar projects elsewhere. The journey of Matteo Paz reminds us that curiosity combined with technology can lead to incredible breakthroughs. As we continue to explore and learn more about our universe, who knows what else we might find?
Check out this related article: Discover the Fascinating History of Earth’s Oceans: From Green Waters to Potentially Purple in the Future!
Source link