Plugavel.
  • Home
  • Tech
  • Car
  • More
    • Privacy policy
    • About us
    • Contact us
No Result
View All Result
Plugavel.
  • Home
  • Tech
  • Car
  • More
    • Privacy policy
    • About us
    • Contact us
No Result
View All Result
Plugavel.
No Result
View All Result

Cosmological simulations with considerably reduced computing time

10 de May de 2021
in Tech
Une nouvelle technique de super-résolution basée sur l’IA permet de créer des simulations de l’Univers en haute résolution. © Marie-Lou Gendron-Marsolais, université de Montréal ; Julie Hlavacek-Larrondo, université de Montréal ; Maxime Pivin Lapointe ; CC by-sa 4.0
ADVERTISEMENT

Researchers have created a super-resolution technique for cosmological simulations using neural networks. It considerably reduces the necessary computation time and makes it possible to simulate large volumes of the Universe in high resolution on a simple graphics card.

To discover the secrets ofUniverse, as dark matter or thedark energy, researchers are working from cosmological simulations. The problem is that current methods force them to choose between a high resolution simulation of a small volume of space, or computing a large area of ​​the Universe with a much lower resolution.

In an article published in the journal Proceedings of the National Academy of Sciences, researchers at Carnegie-Mellon University in the United States describe how they discovered a technique to speed up simulations using artificial intelligence. They resorted to neural networks antagonists generative (GAN) to develop a form of super-resolution, similar to theupscaling used on televisions to improve the quality of low definition video. One of the networks starts from a low-resolution simulation and must invent its own techniques to multiply the resolution by 512. The second network guesses whether it was created using a new technique or the classic method. The loop continues until the two are indistinguishable.

The low-resolution simulation, on the left, is used to create the super-resolution version on the right, and compared to a classic method in the center. © Y. Li and al., Proceedings of the National Academy of Sciences 2021

From several months on a supercomputer to a few hours on a graphics card

This new technique considerably reduces the computation time required. For an area of ​​500 million years light, a simulation with 134 million particles requires only 36 minutes, compared to 560 hours with a conventional technique. Multiply by a thousand, or 134 billion particles, and it will take 16 hours with a graphics card, against several months on a supercomputer with current methods.

Researchers use this kind of simulations to test hypotheses, then use a telescope to check if the results are correct. This advance will make them much more accessible. However, this method has limitations, as it does not take into account the effects of the formation of stars, supernovas where the black holes. The researchers plan to include these phenomena in future versions.

—

Discover Fil de Science! Every Friday, from 6.30 p.m., follow the summary of the scientific news of the week, deciphered for you by the journalists of Futura.

—

Interested in what you just read?

ShareTweetPin

We would like to send you notifications with news, you can unsubscribe at any time.

Unsubscribe
  • Home
  • Privacy policy
  • About us
  • Contact us
© 2021 Plugavel - News about technology and cars on one site Plugavel.
No Result
View All Result
  • Home
  • Tech
  • Car
  • More
    • Privacy policy
    • About us
    • Contact us