How a high school student used artificial intelligence to uncover hidden patterns in one of astronomy’s greatest mysteries
On most nights, the universe appears quiet.
The stars shine steadily. Galaxies drift through the darkness. Planets follow predictable paths around their suns.
But hidden within that apparent calm are cosmic flashes so powerful that, for a few thousandths of a second, they can release as much energy as our Sun emits over several days.
Astronomers call them Fast Radio Bursts, or FRBs.
They arrive without warning, blaze across the cosmos, and disappear almost instantly.
No one knows exactly what creates them.
Now, a 17-year-old student has used artificial intelligence to help decode these mysterious signals, revealing patterns that human researchers might have missed and demonstrating how AI is rapidly becoming one of astronomy’s most valuable tools. His work achieved an impressive 98% classification accuracy and even identified potential new repeating sources hidden within existing data.
The story is remarkable not only because of the science involved, but because it highlights a larger shift taking place across research. Increasingly, discoveries are emerging from collaborations between human curiosity and machine intelligence.
And sometimes, those collaborations begin in high school classrooms.
The Mystery of Fast Radio Bursts
To understand why this achievement matters, it helps to understand the problem.
Fast Radio Bursts are among the strangest phenomena ever detected in space.
First discovered in 2007, FRBs are extremely brief bursts of radio energy that can travel billions of light-years before reaching Earth. Despite lasting only milliseconds, they are extraordinarily powerful. Scientists still do not fully understand their origins, although theories range from highly magnetized neutron stars known as magnetars to other extreme astrophysical events.
What makes FRBs particularly challenging to study is their unpredictability.
Imagine trying to investigate lightning if each lightning strike lasted less than a blink and could happen anywhere in the sky.
That’s the challenge astronomers face.
Over the past decade, observatories around the world have detected thousands of these bursts. Yet every new discovery seems to generate as many questions as answers.
Some FRBs appear only once.
Others repeat.
Why?
That simple question has become one of the central puzzles in modern radio astronomy.
A Teenager Takes on a Cosmic Puzzle
Rohan Arni, a student at High Technology High School in New Jersey, decided to investigate that mystery using artificial intelligence. Instead of focusing solely on traditional astrophysics methods, he combined machine learning with observational astronomy to analyze large datasets collected by the Canadian Hydrogen Intensity Mapping Experiment, better known as CHIME.
CHIME is one of the world’s most productive radio observatories for detecting FRBs.
Every day, it collects enormous amounts of information from the sky.
The volume of data is so vast that manually analyzing every signal would be impractical.
This is where AI becomes valuable.
Machine-learning systems excel at finding subtle patterns hidden within large datasets. They can examine millions of relationships simultaneously and identify distinctions that may not be obvious to human observers.
Arni’s goal was straightforward in theory but challenging in practice: train an AI system to distinguish between repeating and non-repeating Fast Radio Bursts.
The results surprised even experienced researchers.
Training AI to Listen to the Universe
Using deep learning techniques, Arni developed a model capable of classifying FRBs with approximately 98% accuracy.
That number is significant.
In machine learning, accuracy alone does not guarantee meaningful scientific results. A model must also reveal useful information about the underlying phenomenon being studied.
In this case, the AI did more than simply categorize signals.
It identified characteristics that appear to separate repeating bursts from one-time events.
According to the analysis, repeating FRBs generally seem to be located closer to Earth and often occupy narrower frequency ranges than non-repeating bursts.
Those observations may offer clues about the environments where these signals originate.
More importantly, the findings provide researchers with new directions for future investigations.
Rather than searching blindly, astronomers can focus on patterns identified through the machine-learning analysis.
The Discovery Hidden Inside the Data
Perhaps the most intriguing aspect of the project wasn’t the classification accuracy.
It was what the AI found afterward.
While examining the data, the model identified four sources that may actually be repeaters even though they had previously been categorized differently. Two of those candidates had independently attracted attention in earlier studies, adding credibility to the AI’s conclusions.
This illustrates one of the most powerful applications of artificial intelligence in science.
AI is not simply automating existing work.
In some cases, it helps researchers notice things they weren’t actively searching for.
Scientific history contains many examples of accidental discoveries.
Penicillin.
Cosmic microwave background radiation.
X-rays.
Artificial intelligence may increasingly become a tool that creates new opportunities for those kinds of discoveries.
Not because machines are replacing scientists, but because they can help reveal hidden signals buried beneath overwhelming amounts of data.
Why Fast Radio Bursts Fascinate Scientists
FRBs remain one of astronomy’s most active areas of research because they touch on several fundamental questions about the universe.
Scientists still do not know exactly what powers them.
Some evidence points toward magnetars—highly magnetized neutron stars formed after massive stellar explosions.
Other theories involve exotic astrophysical environments that remain poorly understood.
What makes FRBs especially useful is that they also serve as cosmic probes.
As radio waves travel across intergalactic space, they interact with matter scattered throughout the universe.
By studying these interactions, astronomers can learn more about the otherwise invisible material that exists between galaxies.
In other words, FRBs may eventually help answer questions extending far beyond their own origins.
Understanding them could improve our understanding of the large-scale structure of the cosmos itself.
That’s why every new clue matters.
The Growing Role of AI in Astronomy
Arni’s project reflects a broader trend transforming scientific research.
Modern observatories generate data at a scale no human team can fully process alone.
The upcoming generation of telescopes will only accelerate that challenge.
Facilities such as the Vera C. Rubin Observatory and the Square Kilometre Array are expected to produce unprecedented amounts of astronomical information.
Traditional methods simply cannot keep pace.
Artificial intelligence is increasingly becoming essential for managing this flood of data.
Machine-learning systems are already being used to identify exoplanets, classify galaxies, detect gravitational lensing events, and search for unusual cosmic phenomena.
Fast Radio Burst research has also benefited from machine-learning approaches for years, as researchers have used AI systems to filter candidate signals and distinguish genuine astronomical events from interference.
What makes Arni’s work noteworthy is how effectively a young researcher applied these techniques to a major unsolved problem.
The project demonstrates that powerful scientific contributions are no longer limited to massive laboratories or decades-long careers.
Access to data, computing resources, and advanced AI tools is opening new pathways into research.
What This Means for Future Scientists
Stories about artificial intelligence often focus on replacement.
Will AI replace programmers?
Will AI replace writers?
Will AI replace scientists?
Projects like this suggest a different future.
The most exciting outcomes may emerge when humans and AI work together.
The teenager in this story did not simply ask an AI chatbot for an answer.
He developed a research framework, selected data, trained models, evaluated results, and interpreted findings.
The AI performed calculations and pattern recognition at scale.
The human provided direction, curiosity, and scientific judgment.
That partnership is increasingly becoming the model for modern research.
Future scientists may spend less time manually sorting through data and more time asking better questions.
A Glimpse of Astronomy’s Next Era
For centuries, astronomy advanced through increasingly powerful instruments.
Galileo had his telescope.
Edwin Hubble had Mount Wilson Observatory.
Modern astronomers have giant radio arrays and space telescopes.
Today, artificial intelligence is becoming another instrument—one that doesn’t collect light or radio waves, but helps us understand them.
Rohan Arni’s work offers a glimpse of what that future might look like.
A teenager with access to public scientific data and advanced machine-learning techniques was able to contribute meaningful insights to a mystery that continues to challenge professional astronomers.
The Fast Radio Burst puzzle remains unsolved.
Scientists still don’t know exactly what creates these powerful flashes from across the universe.
But the search has gained a new ally.
Somewhere within the torrent of cosmic data streaming toward Earth every day, more answers are waiting. And increasingly, artificial intelligence is helping researchers hear signals that would otherwise remain lost in the noise.
The universe is still speaking.
We’re just getting better at listening.
