Artificial intelligence detects 72 extraterrestrial radio signals

For not the first decade, scientists have been trying to identify at least one signal sent by representatives of other civilizations. The difficulty lies in the high level of noise, which almost completely clogs up the signals from space. And then artificial intelligence came to the aid of scientists.

In a new study published in The Astrophysical Journal, the researchers describe how they used previously intercepted signals from a source billions of light-years away to train a neural network. After that, using artificial intelligence, they isolated several dozen more signals from the previously collected information.

The signals received the code name "fast radio bursts" - FRB (fast radio bursts). Another name for them is "cosmic whistle". They are powerful short pulses of energy reaching us from the depths of space. Scientists still do not know their nature. According to some hypotheses, the source of the signals is magnetized neutron stars and black holes. It is also possible that this was not without our distant space brothers in mind.

FRBs are so short-lived that it is almost impossible to fix them manually, so it made sense to connect AI to this process.

The first thing that researchers at the University of California at Berkeley (USA) did was create their own FRB detection algorithm. They then used this algorithm for analysis when analyzing previously received signals, as a result of which the AI ​​identified 72 new FRBs. In total, about 300 signals have already been detected.