Machine Learning Discovers Previously Undiscovered Interest Signals in the Search for Extraterrestrial Life

Machine Learning Discovers Previously Undiscovered Interest Signals in the Search for Extraterrestri ...

For decades, the search for extraterrestrial life has been a topic of scientific inquiry and public fascination. To answer the question whether we are alone in the universe, scientists utilize various techniques, such as studying other planets and moons in our own solar system, analyzing distant stars' signals, and searching for biosignatures in the atmosphere of exoplanets.

The search for technologically advanced extraterrestrial life raises the question, "Where are they?" The answer often lies in the vastness of the universe and the limited scope of our search. Additionally, outdated computing techniques from the early days may not be efficient in processing today's massive petabyte data."

Deep Learning was used in a recent Nature Astronomy article led by University of Toronto undergraduate student Peter Ma and researchers from the SETI Institute, Breakthrough Listen, and other scientific institutions. This new technique discovered eight previously undiscovered clues.

"We've examined 150 TB of data of 820 nearby stars, on a dataset that had previously been searched through in 2017 but labeled as devoid of interesting signals," said Peter Ma, the lead author. "We're expanding 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 quest to discover 'are we alone in the universe?'

Waterfall plots of the eight indicated signals of interest. Each panel has a width of 2,800 Hz, and the x-axes are referenced to the source of the signal, as reported in column 3 of Table 1.

The most common technique used by SETI researchers is to look for radio signals. It can be used to intercept any radio signals alien civilizations might be transmitting at these distances.

The new data obtained by the Green Bank Telescope in West Virginia was used as part of a Breakthrough Listen campaign that initially did not have any goals of interest. The aim was to enhance the accuracy of the data by applying new deep learning algorithms to an established search algorithm that would yield faster results.

Cherry Ng, another of Ma's research advisors and an astronomer at both the SETI Institute and the French National Center for Scientific Research, stated, "These findings demonstrate the utility of modern machine learning and computer vision techniques in astronomy, resulting in both new detections and improved performance."

Although re-examinations of these new interests have yet to yield re-detections of these signals, this new method of analyzing data can help researchers better understand the data they collect and act swiftly to re-examine targets. Ma and his advisor Dr. Cherry Ng are looking forward to implementing extensions of this algorithm on the SETI Institute's COSMIC system.

Technological advances have enabled researchers to gather more data than ever before. This huge volume of data requires new computational tools to process and analyze it quickly to discover anomalies that may be evidence of extraterrestrial intelligence.

Peter Xiangyuan Ma, Cherry Ng, Leandro Rizk, Daniel Czech, Jamie Drew, Vishal Gajjar, John Hoang, Howard Isaacson, Matt Lebofsky, Imke de Pater, Danny C. Price, Sofia Z. Sheikh, and S. Pete Worden, 30 January 2023, Nature Astronomy. DOI: 10.1038/s41550-022-01872-z

You may also like: