AI uncovers hidden links between the Milky Way and Andromeda

Researchers at the University of Pennsylvania, using machine learning, have created a cosmic map that reveals hidden structures linking galaxies. Space reports.

The researchers say the map could yield new insights into the influence of dark matter on the evolution of the Universe, as well as the collision between the Milky Way and Andromeda in about 4.5 billion years. They discovered filaments of dark matter that connect the two galaxies and may affect their merger.

For the creation of the map, researchers trained a machine learning model on a large set of simulations of stellar systems called IllustrisTNG. The dataset included Milky Way-like galaxies to understand which of their properties best predict the distribution of dark matter.

Co-author Donghui Jeong said that over time, the large-scale structure and complexity of the Universe increase, so it becomes harder for scientists to study dark matter locally.

“Ironically, but the distribution of dark matter is easier to observe at a much greater distance from Earth, because it reflects a very distant past that is far less complex,” he added.

When the model was ready, researchers fed it real data from the Cosmicflows-3 catalog, which includes the motions and distribution of 17,000 galaxies within 200 megaparsecs of the Milky Way. As a result, AI revealed several new filaments that the team plans to study in the future, including connections between our galaxy and Andromeda.

According to the researchers, the map will become even more accurate after the launch of NASA’s James Webb Space Telescope in autumn 2021. With it, scientists will obtain data that will allow them to see even fainter and more distant galaxies.

Back in May 2021, American scientists unveiled an AI supercomputer that will help build the largest-ever 3D map of the visible Universe for studying dark energy.

In early May, researchers at Carnegie Mellon University developed an AI that models the Universe on a single GPU 1,000 times faster than existing methods.

Subscribe to ForkLog AI news on Telegram: ForkLog AI — all the news from the AI world!