Artificial intelligence has discovered new radio signals from outside the galaxy

SETI project researchers used machine learning algorithms to analyze data from the Green Bank radio telescope. This led to the discovery of new repetitive FRB signals from a mysterious source identified as FRB 121102. These pulses were missed when analyzed by conventional search methods.

Fast radio bursts – FRB) is a high-energy astrophysical phenomenon of unknown extragalactic origin, manifesting as krottotrial radio pulse, lasting an average of a few milliseconds. The first such signal was discovered in 2007. Since then, dozens of such signals have been recordedow. They lasted for fractions of a second, but carried a powerful amount of energy porown comparable to the energy emitted by the sun throughout the day.

Most FRBs occur without warning and only once. This means that they cannot be predicted; so astronomers are unable to schedule observations. These are detected only poSource in radio sky observation data.

But in 2012, a unique pulse was recorded – FRB 121102 – the only one of this type of signalow, whichory repeated. Four years poLater, six more FRBs were recorded coming from the same location, then 15 more. A year ago, it was determined that the FRB 121102 pulses originated from a region of theoin the stellarorch dwarf galaxy more than three billion light-years away from Earth. Researchers are convinced that the source of theothe source of FRB 121102 is a neutron star, although there is no shortage of ideas that an advanced extraterrestrial civilization is responsible for the signals.

We know that FRB 121102 is a repeating source of theoFRB bottom. That’s why researchers from the Breakthrough Listen project undertook prob to catch the signal, as it were „in the act”. And it was successful. Last August, they pointed the Green Bank radio telescope located in the US at the source of theoFRB 121102. The data was recorded for five hours. The observations produced 400 terabytesoin the data.

In analyzing the collected data, the researchers discovered 21 FRBs using standard algorithmsow of computer data, all within the first hour. They concluded that the source of theoThe source undergoes periods of stunning activity and relative calmness. But after using a powerful new algorithm to reanalyze the data, they found that the source of theoFRB 121102 bottom is much more active and probably more complex than originally thought. The new algorithm has discovered as many as 72 FRBs from this location, raising the total number of signalow from the same sourceodła to nearly 300.

A paper describing the results of the analysis was accepted for publication in the journal „Astrophysical Journal”. It appeared roalso on preprint serveroin the arXiv.

– This work is just the beginning of applying these powerful new methods. We hope our success can inspire other scientific teams to apply artificial intelligence to radio astronomy – said Breakthrough Listen project leader Gerry Zhang of the University of California Berkeley.

The researchers found that the impulses are not received with any regular pattern. Just as the signal patternsow with pulsarow have helped astronomers develop computer models of extreme conditionsow physical in such facilities, FRB measurements will allow you to know more detailsołow concerning the source ofoFRB 121102 and the FRB itself.

Breakthrough Listen is a scientific program focused on finding evidence ofow for intelligent life in the Universe. It aims to investigate a millionoin nearby stars and those of nearby galaxies in a wide range of radio and optical bands.