Teen Discovers 1.5 Million Unidentified Space Objects Based On NASA Data

Like many high schoolers, I spent my summer vacations goofing off, riding my bicycle everywhere, and reading books (yes, I was a nerdy high schooler). Matteo Paz put his time to much better use, devising a system that analyzed NASA data and discovered 1.5 million previously unidentified objects in the sky, reports Futura. Even my homemade ham radio antennas don't hold a candle to this.

It all started in 2022 when Paz joined the Caltech Planet Finder Academy, intended to give Pasadena high school students a taste of astronomy. He worked with mentor Davy Kirkpatrick on analyzing an enormous archive of data from NASA's NEOWISE satellite, originally intended to detect near-Earth asteroids. While doing that work, the infrared telescope also detected heat variations in more distant objects. Kirkpatrick's notion was to use some of this data to discover these objects. From Caltech:

"At that point, we were creeping up towards 200 billion rows in the table of every single detection that we had made over the course of over a decade," Kirkpatrick says. "So my idea for the summer was to take a little piece of the sky and see if we could find some variable stars. Then we could highlight those to the astronomic community, saying, 'Here's some new stuff we discovered by hand; just imagine what the potential is in the dataset.'"

However, Paz had other ideas. He wanted to put his knowledge of math, programming, and AI to use analyzing the complete dataset, covering the entire sky, to detect these objects automatically. Rather than tell him that's too much, Kirkpatrick encouraged him, and Paz got to work.

Unlike some modern students, Paz didn't just dump the data into ChatGPT and hope for the best. That's just as well, since we've already seen that Google AI can't even do basic math. Instead, with cooperation from other Caltech astronomers, Paz programmed his own algorithm, which broke down the 200 billion data entries into bite-size chunks, then analyzed them for the telltale infrared signatures that identify distant objects like binary stars, quasars, and black holes. 

The result is VARnet, which Paz describes in his paper as "a capable signal-processing model for rapid astronomical time series analysis." Yes, the high schooler even published a paper in The Astronomical Journal about his findings. Caltech is already putting his research to use to study binary star systems. Paz told Smithsonian Magazine that his AI model isn't limited to astronomy, but could be used for "anything else that comes in a temporal format," like analyzing stock market data or environmental effects like pollution. Paz and his family had to evacuate for the Eaton Fire last year, so it's only natural that down-to-earth applications like this come to mind.

In recognition of his work, Paz won the prestigious Regeneron Science Talent Search. It includes a $250,000 prize that he intends to put toward college, he told FOX 11 Los Angeles. While I spent my high school summers reading science fiction, this guy was doing actual science, something we never imagined would even be possible for a student back then.