Astronomy: Machine learning combs radio signals from space

A machine learning method that could be used to efficiently identify unusual radio signals from space for further investigation, while filtering out interference, is reported in a paper published in Nature Astronomy. The research uses data from the Search for Extraterrestrial Intelligence (SETI) Breakthrough Listen Initiative and identified eight previously undetected signals of interest, although they have not been re-detected in follow-up observations.

It has been suggested that the detection of certain types of radio signals could be an indication of potential technological life, given that artificial radio signals can be distinguished from natural ones. SETI programmes have been scanning the sky with radio telescopes for decades to detect unambiguous artificial signals coming from the stars. However, this search is complicated by interference from human technology, which can generate false positive identifications that are time-consuming to filter out from large data sets.

Peter Ma and colleagues present a machine learning-based selection method that they apply to more than 480 hours of data from the Robert C. Byrd Green Bank Telescope, observing 820 stars. The method analysed 115 million snippets of data, from which it identified around 3 million signals of interest. The method was then able to further reduce this to 20,515 signals, which is more than 100 times less than previous analyses of the same dataset. The authors inspected the 20,515 signals and they identified 8 previously undetected signals of interest, although follow-up observations of these targets have not re-detected them.

The authors suggest their method could be applied to other big datasets to accelerate SETI and similar data-driven surveys.