(RNN) – New technology has allowed researchers at the Berkeley SETI Research Center to identify far-off radio emissions that could be "signatures of technology developed by an advanced civilization."
The UC Berkeley "Breakthrough Listen" program used machine learning to identify 72 new recordings known as fast radio bursts that came from a strange repeating burst known as FRB 121102.
The bursts typically last milliseconds. According to the program, the burst can emit as much energy in 10 microseconds as the sun does in an entire year.
As a release from the program explained, "most FRBs have been witnessed during just a single outburst" while this signal "is the only one to date known to emit repeated bursts."
According to the group, that included 21 observations of the burst last year. The new set of 72 were found by examining old datasets.
Machine learning is a process by which, generally, computers are programmed to take huge sets of data and "learn" to classify and organize it into practical information.
"Not all discoveries come from new observations," Pete Worden, the executive director of the program, said in the release. "In this case, it was smart, original thinking applied to an existing dataset. It has advanced our knowledge of one of the most tantalizing mysteries in astronomy."
The source for the newly-found fast radio bursts is a galaxy about 3 billion light years from Earth.
FRB 121102 was first observed in 2012 and, in 2015, it was found to be repeating.
As a January report in The New York Times noted, "Among the more out-there explanations proffered was that they are lasers propelling alien interstellar spacecraft."
The researchers used an algorithm known as a "convolutional neural network" to comb through a massive set of data, 400 terabytes, and identify bursts missed during last year's research.
One terabyte of data is equivalent to about 310,000 pictures.
Andrew Siemion, the director of the Berkeley SETI Research Center, said in a UC Berkeley news item that the successful application of machine learning could lead to more breakthroughs.
"This work is exciting not just because it helps us understand the dynamic behavior of fast radio bursts in more detail, he said. "But also because of the promise it shows for using machine learning to detect signals missed by classical algorithms."