The research, led by University of Toronto undergraduate student Peter Ma, who began working with the Breakthrough Listen team while still in high school, identified around 3 million signals in scans of 820 stars observed with GBT.
"The key issue with any technosignature search is looking through this huge haystack of signals to find the needle that might be a transmission from an alien world," explained Dr. Steve Croft, an astrophysicist with the Breakthrough Listen team at the University of California, Berkeley (and one of Ma's research advisors).
"The vast majority of the signals detected by our telescopes originate from our own technology - GPS satellites, mobile phones, and the like. Peter's algorithm gives us a more effective way to filter the haystack and find signals that have the characteristics we expect from technosignatures."
Classical technosignature algorithms compare scans where the telescope is pointed at a target point on the sky with scans where the telescope moves to a nearby position, in order to identify signals that may be coming from only that specific point. These techniques are highly effective - for example, they can successfully identify the Voyager 1 space probe, at a distance of 20 billion kilometers, in observations with the GBT. But these algorithms struggle in crowded regions of the radio spectrum, where the challenge is akin to listening for a whisper in a crowded room.
The process developed by Ma inserts simulated signals into real data, and trains an artificial intelligence algorithm known as an autoencoder to learn their fundamental properties. The output from this process is fed into a second algorithm known as a random forest classifier, which learns to distinguish the candidate signals from the noisy background.
"In 2021 our classical algorithms uncovered a signal of interest, denoted BLC1, in data from the Parkes telescope," said Dr. Andrew Siemion, Breakthrough Listen's Principal Investigator. "Peter's algorithm is even more effective in finding signals like this. Any technosignature candidate needs to be confirmed, however, and when we looked at these targets again with the GBT, the signals did not reappear. But by applying this new technique to even larger datasets, we can more effectively identify technosignature candidates, and hopefully eventually even a confirmed technosignature."
"It's exciting to see new approaches like this being developed by imaginative young people like Peter at the beginning of their scientific careers," said Breakthrough Initiatives Executive Director Dr. S. Pete Worden. "We'll continue to monitor the stars Peter observed, and we'll continue to develop our use of artificial intelligence to help us try to answer humanity's most profound question: are we alone?"
Cherry Ng, another of Ma's research advisors at the University of Toronto and now an astronomer at the French National Center for Scientific Research said, "These results dramatically illustrate the power of applying modern machine learning and computer vision methods to data challenges in astronomy, resulting in both new detections and higher performance. Application of these techniques at scale will be transformational for radio technosignature science."
Larger datasets are imminent. The team recently announced the start of observations using the MeerKAT telescope in South Africa, where efficient techniques for finding signals of interest in 24/7 observations with this powerful array are urgently needed.
"There is such a large volume of data," explains Ma. "The more traditional ways of searching for extraterrestrial life are just not sufficient."
A preprint of Ma's paper, links to the datasets examined, and an artist's illustration of the search, are available here
Research Report:Overview - Deep Learning Search for Technosignatures
Related Links
Breakthrough Listen
Lands Beyond Beyond - extra solar planets - news and science
Life Beyond Earth
| Subscribe Free To Our Daily Newsletters |
| Subscribe Free To Our Daily Newsletters |
