قالب وردپرس درنا توس
Home / Science / Astronomers trained AI to find collisions with ancient galaxies from the early Universe

Astronomers trained AI to find collisions with ancient galaxies from the early Universe



Everywhere in the universe there are bright lights where galaxies collide. And everywhere in the universe there are bright spots – especially in the distant universe – where galaxies produce stars at unusual rates. Now a team of astronomers has a new way to tell them apart.

Here's the problem: Telescopes can not detect distant, ancient parts of the Universe clearly enough to recognize the common signatures of galaxy fusion A good way to distinguish these two types of superbright galaxies.

In an article published in the July issue of the Monthly Notices journal of the Royal Astronomical Society, researchers have described a new system that allows distant galaxy fusions to be distinguished from distant galaxies with a super bright glow because of their unusually large number of new stars give birth.

Related: 9 Ideas About Black Holes That Will Blow You Away

In Relatively Close Parts of the Universe We can easily see galactic mergers based on tsunamis of stars on their peripheries. Long "tidal arms" of stars extend from the cores of galaxies that have recently joined together, like arrows that mark them as star archaeologists, who are visible through space. At the same time, the first few billion years of the history of our universe were shaped by galaxies that produced stars at high rates that would be unusual today. So it is unclear for a long time which distant superbright galaxies can see astronomers in the early Universe as a result of mergers and which are so bright by themselves.

The team argued that because we know which star-forming galaxies and galactic fusions look close up, it would be relatively easy to create fake images and then blur and distort them as if the light of those galaxies were from one of our galaxies Space telescopes would be captured remotely. That's what the team did by creating more than 1

million fake Hubble space telescopes and James Webb space telescope images. The astronomers knew which were blurry, distant images of galactic collisions versus blurry images of superbright, star-shaped galaxies, though they looked very similar at first glance. The researchers were thus able to find subtle signatures with which astronomers distinguish galaxy fusions from galactic star factories in the distant, ancient universe. And they trained their own machine learning algorithm to distinguish between images of the two types of galaxies.

That's a big deal, because the entire universe is full of merging galaxies, the researchers said in a statement – up to 5% of the galaxies are involved in mergers at some point, and even the Milky Way is expected to one day merge with his neighbor Andromeda.

The new method has its limitations, researchers wrote in the work. There is always the possibility of distortion in the simulation database, and some inaccurate attempts are primarily associated with the creation of the database. To further enhance the machine learning algorithm and distinguish older galaxy fusions, they need to create a much larger database.

Originally published on Live Science . [19659014]
Source link