‘Are we alone?’ is perhaps one of the most profound questions, asked repeatedly by humanity since the dawn of space exploration. Now, with the rapid development of artificial intelligence, we may be one step closer to knowing the answer.
Artificial intelligence (AI), or machine learning, has long been used in the fields of astronomy and astrophysics due to its ability to spot patterns and sift through huge data sets. Now it’s being put to the test in the search for extraterrestrial life. In a paper published at the end of last month in Nature Astronomy, a group of scientists lead by Peter Ma from the University of Toronto outlined the result of their AI-based search for intelligent life beyond our planet. The AI algorithm they built was designed to sieve through data from telescopes that detect signals from space at radio frequency and identify anything that could not have been produced by natural astrophysical processes. This data was then sorted through and signals that could be due to man-made interference (common in the radio frequency band due to GPS, Wifi, satellites and phones) was disregarded. They were left with eight signals that were not natural or man-made, leaving the potential that these could have been produced by some technologically advanced society light years away.
The basis behind this search is that any society advanced and intelligent enough to have built communication devices will be emitting what’s known as technosignatures, which our telescopes would be able to detect. Conversely, another intelligent society may be able to detect technosignatures coming from planet Earth and be able to identify us as intelligent life. Radio frequency was chosen out of the many frequency bands that signals could be emitted at due to its ability to easily propagate through space and the relative ease of constructing radio receivers and transmitters. There is however always the chance that some far-off society is communicating at an entirely different frequency or at signals too weak to be picked up by our devices.
Before any signals from extraterrestrial intelligent life could be detected, the algorithm had to be trained to spot them. AI algorithms are not able to understand or think for themselves. They just do what humans programme or teach them to do. Since we don’t possess any genuine alien radio signals to teach them with, a training set of simulated technosignatures was created in order to show the algorithm what to look for. After it had been trained, the algorithm was fed more than 480 hours of telescope observation data from 30 to 90 light years away, initially identifying 20’515 signals of interest. These signals were then manually inspected and whittled down to eight signals from five different stars. The primary reason for the rejection of signals was due to human radio interference, the main challenge associated with the search for extraterrestrial life using this method. Despite locating Earth’s radio telescopes in radio quiet zones where they are far from major sources of radio signals, the issue is still prominent. The hope is that future algorithms will be able to filter out these man-made radio signals.
The eight signals identified as potentially being from extraterrestrial life forms had never previously been detected and were of great interest. Unfortunately, when re-observation took place to try and verify these signals, researchers were unable to detect them. It’s likely that these signals were also due to man-made radio interference and are not a sign of intelligent life far away, but the findings were not conclusive. All the scientists were able to say was that while we do not know the origins of the signals, they are not constant throughout time.
While the search for extraterrestrial life has not yet been successful, the use of artificial intelligence is definitely a big step towards data-driven astronomy. Detecting nothing does not mean we are alone: with more sensitive future telescopes and pioneering algorithms, hopefully we will soon be able to identify a definitive signal of intelligent life beyond our planet.
Image credit: Uninteneded Concept on Unsplash
