We welcome guests from another galaxy. Supercomputers have discovered a stream of stars falling into the Milky Way

Thanks to the latest data collected by the Gaia space observatory, scientists have identified in our direct star environment a vast stream of stars that did not form in our galaxy, and are only now joining it.

Nix, because this name was given to this peculiar stream, is most likely a remnant of a dwarf galaxy that collided with the Milky Way. Most often, these types of streams are only memories of former globular clusters or dwarf galaxies, which as a result of tidal interactions with the Milky Way were first stretched, and eventually burst and absorbed by our galaxy.

Nix was discovered in a rather unusual way, which perfectly shows how new scientific discoveries occur in today's astronomy and astrophysics.


Lina Necib deals with the kinematics of stars and dark matter in the Milky Way.

If we have groups, density of stars moving in the same characteristic way, then usually there is also a reason why they move in the same way.

Since 2014, as part of the FIRE (Feedback in Realistic Environments) project, researchers from several US universities have been working on creating highly detailed simulations of realistic galaxies. All of our current knowledge of how galaxies form and evolve is included in these simulations. Starting at the moment of the Big Bang, simulations lead to the formation of galaxies as we see them in the present universe.

Creating a Milky Way map

Even before the FIRE project began, the Gaia space observatory was launched into space in 2013. Its main task is to create a precise, three-dimensional map of about one billion stars in the Milky Way and in its surroundings.

The Gaia Observatory measures the location, distances, speeds and direction of movement of all observed stars. For seven million stars, we also have speed measurements in three dimensions, which means that we know exactly where these stars are and in which direction and how fast they are moving. Thanks to such data, it is now possible to carry out extensive analyzes that allow scientists to look more broadly at the structure of the entire Milky Way.

The Nix stream was discovered by combining data from these two astrophysical projects and using machine learning methods.

Both FIRE and Gaia are trying to answer the question of how the Milky Way has become the system we are observing today.

Galaxies often grow by absorbing other galaxies. We assumed that the Milky Way did not have too many collisions with other galaxies and for some time it surprised us, because our simulations indicate a lot of collisions. Now, thanks to the fact that we have access to data with a large number of smaller structures, it turns out that it was not as quiet here as it may seem. To reach this conclusion, however, we needed both tools and observational data and simulations. We are only at the beginning of the road to a true knowledge of the history of the formation of the Milky Way.

Gaia space observatory

Gaia requires deep learning algorithms

Creating a billion stars map is a real curse. On the one hand, every scientist dreams of such detailed data, and on the other hand, nobody can analyze such amount of data on their own. You can't look at the data on seven million stars and understand what they do, how they behave.

To this end, scientists created artificial Gaia catalogs. Based on simulations of galaxies in FIRE, scientists created artificial catalogs of stars, on which they tested subsequent models, whose task was based on the movement of stars in the simulation, to divide them into those that formed in the galaxy and which were taken from outside. Often, the differences between such stars are very small and subtle, but computer models should be able to extract them from the data.

After training the models, they were tested on data from the Gaia observatory.

We recommended that we identify stars outside the Milky Way based on what we learned in FIRE simulations, says Necib.

During analysis, the models assigned each non-galaxy star a degree of confidence from 0 to 1 that it was actually outside the galaxy. Scientists set the cut-off point at different levels of confidence and analyzed the results obtained in this way.

To check if the models were actually catching the stars that they should have been checked, they were checked in the Gaia data to catch those stars that we already know that they did not form in our galaxy. Such objects include Sausage - the remainder of a dwarf galaxy that merged with the Milky Way about 6-10 billion years ago and has a very distinctive, sausage-like shape.

If our neural network works as we wanted, it should immediately identify the stars belonging to the Sausage.

And in fact, Sausage was immediately noticed by the network. The same happened with the other objects that according to scientists, the network should see: the Helmi stream, which is a remnant of another dwarf galaxy that merged with our galaxy in the distant past, discovered only in 1999, or the galactic halo itself.

New facility: Nix

The model also identified another structure in the data: a 250 star cluster also orbiting the center of the Milky Way with a disk, but also clearly heading towards the center of the Milky Way.

At first reflex I found it a mistake. I thought "Oh no!" I didn't even mention it to other band members for three consecutive weeks. At that time, however, I realized that this is not a data error, it is actually a new structure that we have never seen before.

I began to analyze all the literature and check if it was sure that nobody would see it before. It turned out not to be. Thanks to this, I had the unique opportunity to name this new stream - and as we know it is the most exciting privilege in all astrophysics. I named her Nix, after the Greek goddess of the night. This is an extremely unique structure, because without machine teaching we would have little chance to see it at all.

The entire project required very advanced and high computing power at many stages. FIRE and FIRE-2 simulations are one of the largest computer models of galaxies ever created. Each of the nine main simulations required many months of calculations carried out on the largest and fastest supercomputers in the world. Blue Waters and Stempede2 computers were used to create them.

Other researchers used clusters at the University of Oregon to deeply learn models and apply them to huge databases from the Gaia Observatory. They are currently continuing their work on Fronter, the fastest university system in the world.

Each element of this project required enormous computing power and would not be possible without supercomputers - adds Nacib.

What's next?

Necib and his team plan to conduct Nix observations using ground telescopes. Thanks to this, it will be possible to determine the chemical composition of the entire stream and other details that will allow to determine when Nix joined the Milky Way, and thus where it comes from.

The next set of data from the Gaia observatory, which will be published in 2021, will contain data for another 100 million stars that may also provide new streams.

We welcome guests from another galaxy. Supercomputers have discovered a stream of stars falling into the Milky Way


Popular posts from this blog

What is VoLTE and how can you activate it on your Xiaomi

So you can check the battery status of your Xiaomi smartphone and how many cycles you have performed

How to exit the FASTBOOT mode of your Xiaomi if you have entered accidentally

Does your Xiaomi charge slowly or intermittently? So you can fix it

Problems with Android Auto and your Xiaomi? So you can fix it

If your Xiaomi disconnects only from the WiFi it may be because of that MIUI setting

How to change the font in MIUI and thus further customize your Xiaomi: so you can change the type, color and size of the letters of MIUI

What is the Safe Mode of your Xiaomi, what is it for and how can you activate it

Improve and amplify the volume of your Xiaomi and / or headphones with these simple adjustments

How to activate the second space if your Xiaomi does not have this option