The first-ever AI simulation of the universe seems to work like the real factor — and is almost as mysterious.
Researchers reported the brand new simulation on June 24 in the journal Proceedings of the National Academy of Sciences. The objective was to create a virtual model of the cosmos to simulate different conditions for the universe’s beginning. However, scientists additionally hope to study their simulation to understand why it works so well.
“It is like teaching picture-recognition software with numerous footage of cats and dogs. However, then it can recognize elephants,” research co-author Shirley Ho, a theoretical astrophysicist at the Center for Computational Astrophysics in New York City, mentioned in a press release. “No one knows how it does this, and it is a terrific mystery to be solved.” [Far-Out Discoveries About the Universe’s Beginnings]
The model may also be a time-saver for researchers interested in universal origins. The brand new neural network might full simulations in 30 milliseconds, compared to several minutes for the fastest non-artificial intelligence simulation technique. The network additionally had an error rate of 2.8%, compared with 9.3% for the existing fastest model. (These error rates are in comparison with a gold standard of accuracy, a model that takes hundreds of hours for every simulation.)
The researchers now plan to vary other parameters within the new neural network, examining how components like hydrodynamics, or the movement of fluids and gases, may have shaped the universe’s formation.