Scientists devise machine learning method that identifies galaxies — DNOTICIAS.PT

It was announced today that a scientific team led by the Institute for Astrophysics and Space Sciences (IA), in collaboration with a company, has created an automatic learning method that recognizes “ultra-luminous” galaxies from the earliest days of the universe.

The method is described in an article published in the journal Astronomy and Astrophysics and, according to its authors, “will allow astronomers to be more effective in searching for so-called radio galaxies,” galaxies “with active nuclei that emit powerful jets.” The material, which fluoresces at radio frequencies,” says IA in a statement.

The international team led by IA, in cooperation with Closer, an artificial intelligence company, developed an algorithm that was trained using images of galaxies with different wavelengths of light.

When tested with other images, the algorithm has proven to be “capable of predicting four times as many radio galaxies as conventional methods with explicit instructions,” the IA statement says, adding that machine learning (a branch of artificial intelligence) can “help explain the phenomena.” “Physics.” That occurred in radio galaxies when the universe was a tenth of its current age.

One of the article’s authors, José Afonso, an IA researcher who specializes in studying galaxies, highlighted in the same statement the importance of “developing advanced techniques” for astronomy to process and analyze “enormous quantities.” data”.

“At IA we are developing and implementing these techniques to be able to decipher the origin of galaxies and the supermassive black holes that many host,” he stressed.