[ad_1]
Breakthrough Scientists have used artificial intelligence to scan radio signals recorded from a fast radio burst, capturing far more than humans could. (Breakthrough image listening)
Artificial intelligence is invading many areas, most recently astronomy and the search for a smart life in the universe, or SETI.
Discovery researchers Listen, a SETI project conducted by the University of California at Berkeley used machine learning to discover 72 new fast radio stations from a mysterious source located at $ 3 billion. Light years from Earth.
Fast radio bursts are light pulses of radio transmission only a few milliseconds, which are thought to be from distant galaxies. The source of these emissions is still unclear. Theories range from highly magnetized neutron stars blown by gas flows from a near supermassive black hole, to suggestions that bursting properties correspond to the signatures of technologies developed by an advanced civilization.
"This work is exciting not only because it helps us to understand the dynamic behavior of fast radio bursts in more detail, but also to promise to use machine learning to detect signals missed by conventional algorithms," explains Andrew Siemion, director of the Berkeley SETI research center and principal investigator for Breakthrough Listen, the initiative to find signs of intelligent life in the universe.
Breakthrough Listen also applies the successful machine learning algorithm to find new types of signals that can come from extraterrestrial civilizations.
Something emits repeated and powerful bursts of energy
While most fast radio bursts are unique, the source here, FRB 121102, is one of a kind with regard to issuing repeated bursts. This behavior has attracted the attention of many astronomers in the hope of pinpointing the cause and extreme physics involved in fast radio gusts.
The AI algorithms loaded with radio signals from the data were recorded over a five-hour period on August 26, 2017 by the Green Bank Telescope in West Virginia. An earlier analysis of the 400 terabytes of data employed standard computer algorithms to identify 21 bursts during this period. All were seen in one hour, suggesting that the source alternates between periods of rest and hectic activity, said Vishal Gajjar, postdoctoral researcher at Berkeley SETI.
UC Berkeley Ph.D. The student Gerry Zhang and his collaborators then developed a powerful new machine learning algorithm and reanalyzed the 2017 data, uncovering another 72 gusts that were not originally detected. This brings the total number of gusts detected from FRB 121102 to around 300 since its discovery in 2012.
"This work is only the beginning of the use of these powerful methods to find radio transients," Zhang said. "We hope our success could inspire further serious efforts in the application of machine learning to radio astronomy."
Zhang's team used some of the techniques used by Internet technology companies to optimize search results and rank images. They formed an algorithm known as a convolutional neural network to recognize the bursts found by the classical search method used by Gajjar and his collaborators, and then set up on the dataset to find bursts as the classical approach missing.
The results helped to establish new constraints on the frequency of FRB 121102 pulses, suggesting that the pulses are not received with a regular pattern, at least if the period of this pattern is greater than about 10 milliseconds. Just as pulse patterns of pulsars have helped astronomers compel computer models to extreme physical conditions of these objects, the new measurements of FRBs will help determine what powers these enigmatic sources, said Siemion.
"As the FRBs themselves turn out to be signatures of extraterrestrial technologies, Breakthrough Listen is helping to push the boundaries of a new and growing field of our understanding of the universe around us," he added.
The new results are described in an article accepted for publication in The astrophysical journal and is available for download from the Breakthrough Listen website.
RELATED INFORMATION
Source link
