Scientists Use AI to Find 72 Mysterious Radio Bursts From Far-Off Galaxy

Members of the Breakthrough Listen team released a new study on Monday, detailing their discovery of 72 new fast radio bursts from a mysterious source located some 3 billion light years from Earth.
Sputnik

Officials based at the University of California, Berkeley, Search for Extraterrestrial Intelligence (SETI) Research Center found the new signals after using an artificial intelligence algorithm to comb through 400 terabytes worth of radio data from a dwarf galaxy.

The data researchers recently re-examined was collected over a five-hour period by the Green Bank Telescope in West Virginia on August 26, 2017. A previous examination of the recording turned up only 21 bursts. Officials were able to find dozens of new bursts by using a new algorithm known as a convolutional neural network, according to officials.

Mysterious Signals From Space? Telescope Intercepts Strange Radio Bursts

"This work is exciting not just because it helps us understand the dynamic behavior of fast radio bursts in more detail, but also because of the promise it shows for using machine learning to detect signals missed by classical algorithms," Andrew Siemion, the principal investigator in the study, told Berkeley News.

The total number of fast radio bursts detected from the 2017 data currently stands at roughly 300 since the source, FRB 121102, was first discovered in 2012.

Although it's unclear what is causing the pulses, officials have come up with several theories that range from neutron stars being blasted by gas streams from supermassive black holes, to more bold suggestions that the bursts are being emitted by technology developed by extraterrestrial life.

In 2017, researchers at Harvard University theorized that the radio bursts were the result of energy leaks from transmitters built by alien civilizations attempting to send light sail ships on an interstellar expedition, Sky News reported.

"This work is only the beginning of using these powerful methods to find radio transients," Gerry Zhang, an astronomer who runs the Breakthrough Listen program, told Science Alert. "We hope our success may inspire other serious endeavors in applying machine learning to radio astronomy."

Meet 'Robat': Machine That Uses Sonar in Robotics to Generate Map of Space

In analyzing the frequency of the bursts, officials have already begun to speculate that they likely come from the vicinity of a black hole or a powerful nebula after detecting a strange warp in the pulses.

"Not all discoveries come from new observations," Pete Worden, executive director of Breakthrough Initiatives, said in a statement. "In this case, it was smart, original thinking applied to an existing dataset. It has advanced our knowledge of one of the most tantalizing mysteries in astronomy."

The results of the team's work have been accepted for publication in The Astrophysical Journal and are currently available on the Breakthrough Listen website.

Discuss