English
0

The largest collection of cosmic simulations for teaching artificial intelligence has been released

بزرگترین مجموعه شبیه‌سازی‌ کیهانی برای آموزش هوش مصنوعی منتشر شد

The CAMELS project has just released the largest set of simulations to simulate 4,233 worlds, millions of galaxies, and 350 terabytes of data. The goal of this project is to use cosmic simulations to teach artificial intelligence models so that we may be able to decipher some features of our world.

“Francisco Viascosa-Navarro” co-director of the project Says Scientists have already started using this information, which is available for free download, to further their studies. Viascosa Navarro is leading the project with scientists from the Flatiron Institute, along with Shay Genel and Daniel Angels-Alcazar.

“Machine learning has revolutionized many areas of science, but it takes a huge amount of data to use the full potential of this technology,” says Angeles Alcazar. “The general release of CAMELS data, which covers thousands of simulated worlds covering a wide range of physics, provides an exceptional opportunity for galaxy researchers and cosmologists to explore the capabilities of machine learning algorithms in solving a variety of problems.”

The CAMELS project team created these simulations using the IllustrisTNG and Simba project code. The team members are also a combination of the members of these two projects. About half of these simulations combine the physics of the universe with the physics of lower scales but influential in the formation of galaxies.

Cosmic simulations are slightly different

Each simulation is performed with relatively different assumptions about the world. These assumptions, for example, are about the amount of invisible dark matter in proportion to the dark energy that expands the universe. Researchers have designed these simulations to teach machine learning models so that they can later extract useful information from real-world observations.

The CAMELS project, based on the term “cosmology and astronomy with machine learning simulations”, is the largest collection of cosmological simulations built for machine learning algorithms. These data could help make new discoveries and link cosmology and astronomy.

One of the projects that used this data set is a researcher from the University of Valencia in Spain. With CAMELS simulations, he and his team developed an artificial intelligence model to measure the mass of the Milky Way galaxy and the dark matter ring around it, along with the Andromeda Galaxy and its rings. The result was that the mass of our galaxy was estimated to be about 1 to 2.6 trillion times the mass of the Sun. This estimate corresponds to the numbers obtained through other scientific methods and shows the high accuracy of artificial intelligence models.

More Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *

Fill out this field
Fill out this field
Please enter a valid email address.
You need to agree with the terms to proceed

Most Viewed Posts
Menu