With the help of AI (Artificial Intelligence), British researchers have found fifty new planets, marking a tech innovation in the field of astronomy.
Computer scientists and astronomers working at the University of Warwick developed a machine-learning algorithm to evaluate via former NASA data holding thousands of potential planet contenders.
“British researchers have identified 50 new planets using artificial intelligence, marking a technological breakthrough in astronomy.
….a machine learning algorithm to dig through old @NASA data containing thousands of potential planet candidates.”https://t.co/zg1tHOy5Bt
— Amit Paranjape (@aparanjape) August 26, 2020
However, it is not unambiguous which of these contenders are real and experts looking for exoplanets such as planets outside our solar system, they see dips in light that identify a planet crossing between their bar and the telescope. Besides this, those dips may be created by several other factors, such as errors in the camera or background interference.
Source: Web
A team of few specialists trained the machine learning algorithm by having it go with the data gathered by NASA’s Kepler Space Telescope (now-retired), which almost spent around 9 years in profound space on the world’s hunting mission. Furthermore, that algorithm, once, learned to perfectly separate real planets from wrong positives, it was applied to examine the old data sets that still not confirmed that’s why it recognized fifty exoplanets.
New exoplanets orbit around other stars
In a news published, the university said that those fifty exoplanets, which circulate around other stars, range in size froma s big as Neptune to smaller than our Earth. Several of their circulating paths are as extensive as two-hundred days, and few as a short single day. Now astronomers understand that the planets are real, and they can prioritize them for more evaluation.
The measures recognized by the researchers were aired previous week in the Monthly Notices of the Royal Astronomical Society.
In the news release, one of the study’s top authors, David Armstrong of the UOW (University of Warwick), said that in terms of planet authentication, no one had used that machine learning algorithm before. Moreover, machine learning was used to ranking planetary contenders but never in a probabilistic framework; that’s why you need to authenticate the planet truly.
