AI finds 'alien signals' hidden in radio data

The SETI (Search for Extraterrestrial Intelligence) Institute are renowned for scanning the skies for "technosignatures" that could tie to alien technology.

Its decade-long searches are yet to show any convincing leads. However, a new paper in the journal Nature Astronomy has expressed its hope for change by tackling the problem with a machine-learning method.

Taking data from 2016, the machine analysed over 480 hours of data from 820 stars and identified eight points of interest, that previous tech had not picked up on.

The first author of the paper said that while artificial intelligence has been applied to SETI's data in the past, this new approach takes the search out of human hands.

Peter Ma told Motherboard in an email: "Previously people have inserted ML [machine learning] components into various pipelines to help with the search,"

"This work relies entirely on just the neural network without any traditional algorithms supporting it and produced results that traditional algorithms did not pick up."

Sign up for our free Indy100 weekly newsletter

The authors followed up followed up eight signals (labelled MLc1-8) from seven stars – and hope for a repeat.

"We’re scaling this search effort to 1 million stars today with the MeerKAT telescope and beyond. We believe that work like this will help accelerate the rate we’re able to make discoveries in our grand effort to answer the question 'are we alone in the universe?'" Peter, an undergraduate at the University of Toronto, said.

In an email to IFLScience, Dr Franck Marchis of SETI said: "It is a pity that, despite attempts reported by the team, these signals could not be confirmed by other instruments,"

"The MLc1 and MLc7 signals are very interesting because they were recorded on two different dates, suggesting that they are not known interference if they are terrestrial in nature. Such a discovery requires confirmation by other instruments before we can be sure that we have detected extraterrestrial life.

"Nevertheless, this scientific result shows that it is now possible to announce this kind of detection quickly enough to do the necessary follow-up."

He added: "The arrival of large networks such as MeerKAT and the SKA, which will produce terabytes of data per week, make it imperative that SETI research adopt powerful algorithms such as deep learning,"

"We hope that this algorithm will be able to detect a signal more quickly than conventional methods because this will allow us to follow up with other antennas and therefore confirm whether a signal is extraterrestrial."

Have your say in our news democracy. Click the upvote icon at the top of the page to help raise this article through the indy100 rankings.