Artificial Intelligence Helps Find New Fast Radio Bursts

Artificial Intelligence Helps Find New Fast Radio Bursts

Machine learning algorithms applied to data from the Green Bank Telescope find new pulses from the mysterious repeating source FRB 121102. 

robot
Credit: Breakthrough Listen / Danielle Futselaar

 

New detections mark the first time that machine learning techniques have been used to directly detect a fast radio transient.  Successful application of machine learning for signal detection promises to open up new avenues for identifying signals from extraterrestrial intelligence.  

 

September 10, 2018 – Mountain View, CA – Researchers at Breakthrough Listen – the initiative to find signs of intelligent life in the universe – have applied machine learning techniques to detect 72 new fast radio bursts emanating from the "repeater" FRB 121102. Fast radio bursts, or FRBs, are bright pulses of radio emission, just milliseconds in duration, thought to originate from distant galaxies. Most FRBs have been witnessed during just a single outburst. In contrast, FRB 121102 is the only one to date known to emit repeated bursts, including 21 seen during Breakthrough Listen observations made in 2017 with the Green Bank Telescope (GBT) in West Virginia[1].

 

The results of this research have been accepted for publication in the Astrophysical Journal and will be available on the arXiv service on Monday 10 September, 2018.  UC Berkeley doctoral student Gerry Zhang is the lead author of the paper; Dr. Andrew Siemion, Bernard M. Oliver Chair for SETI at the SETI Institute, Berkeley SETI Research Center Director and Breakthrough Listen Principal Investigator, co-authored the work.

FRBs from 121102 originate in a dwarf galaxy 3 billion light years from Earth, but the nature of the object emitting them is unknown. There are many theories, including that they could be the signatures of technology developed by extraterrestrial intelligent life.

 

In August of 2017, the Listen science team at the University of California, Berkeley SETI Research Center[2] observed FRB 121102 for five hours, using digital instrumentation at the GBT. Combing through 400 TB of data, they reported (in a paper led by Berkeley SETI postdoctoral researcher Vishal Gajjar, recently accepted for publication in the Astrophysical Journal[3]) a total of 21 bursts. All were seen within one hour, suggesting that the source alternates between periods of quiescence and frenzied activity.

 

Now, Zhang and collaborators have developed a new machine learning algorithm, and reanalyzed the 2017 GBT dataset, finding an additional 72 bursts that were not detected originally. They trained an algorithm known as a convolutional neural network to recognize bursts found by the classical search method used by Gajjar and collaborators, and then set it loose on the 400 TB dataset to find bursts that the classical approach missed.

 

"Gerry's work is exciting not just because it helps us understand the dynamic behavior of FRBs in more detail," remarked SETI Institute Bernard M. Oliver Chair for SETI Dr. Andrew Siemion, "but also because of the promise it shows for using machine learning to detect signals missed by classical algorithms."  He went on to say that “these new techniques are already improving our sensitivity to signals from extraterrestrial technologies.”

 

The SETI Institute has been using IBM Cloud and AI algorithms to analyze over 20 million signals captured by the ATA radio telescopes, using the power of machine learning to greatly improve how anomalous signals of interest can be identified and flagged for further examination. In addition, scientists at the SETI Institute have used use the IBM Cloud to search for wide-band signals from exoplanets of interest, such as the TRAPPIST-1 system, allowing the ATA to be used for entirely new types of observation programs. With a very wide field-of-view and unprecedented instantaneous frequency coverage, the ATA is well positioned to conduct searches for broadband fast transients, including FRBs.

 

“These results hint that there could be vast numbers of additional signals that our current algorithms are missing and clearly demonstrate the power of applying modern data analytics and AI tools to astronomical research,” said SETI Institute President and CEO Bill Diamond.  “Applying these techniques in the search for evidence of extraterrestrial technologies, or technosignatures, is incredibly compelling, together with addressing the tantalizing phenomena of FRBs” he continued.

 

Additional FRB research may provide clues about whether or not they are signatures of extraterrestrial technology. In the meantime, the SETI Institute’s search to understand the origin and nature of life in the universe, and the evolution of intelligence will continue.

 

Additional details, including raw data, code and a preprint publication describing the work are available at http://seti.berkeley.edu/frb-machine

 

About the SETI Institute

Founded in 1984, the SETI Institute is a non-profit, multi-disciplinary research and education organization whose mission is to explore, understand, and explain the origin and nature of life in the universe and the evolution of intelligence. Our research encompasses the physical and biological sciences and leverages expertise in data analytics, machine learning and advanced signal detection technologies. The SETI Institute is a distinguished research partner for industry, academia and government agencies, including NASA and NSF.

Contact information

Rebecca McDonald

Director of Communications
SETI Institute
189 Bernardo Ave., Suite 200
Mountain View, CA 94043
rmcdonald@seti.org
www.seti.org