The number of factors you have to consider when calculating how the big bang went down is tremendous. The key part of the story you have to model is quark-gluon plasma. This is a state where there were many tiny quantum entangled particles, which eventually reached a point of extremely high energy. When they collided, the aftermath was spewed out to create the physical universe we know today.
Extremely complex computer programming is needed to model this and their results are very difficult to evaluate. This is why a team from TU Wien, in Vienna, decided to design a special kind of AI in this modeling process.
“Simulating a quark-gluon plasma as realistically as possible requires an extremely large amount of computing time,” says Dr. Andreas Ipp, who worked on the project. “Even the largest supercomputers in the world are overwhelmed by this.” Making matters worse, the key forces between the particles being modeled can be described in a number of different ways, this is called gauge symmetry. Basically, the math is complicated!
Therefore, the team had to get creative. Using an approach called a “neural network”, more relevant information was able to be filtered out by the algorithm. Plus, this structure allowed for gauge symmetry to be taken into account.
“With such neural networks, it becomes possible to make predictions about the system—for example, to estimate what the quark-gluon plasma will look like at a later point in time without really having to calculate every single intermediate step in time in detail,” says Ipp. “And at the same time, it is ensured that the system only produces results that do not contradict gauge symmetry—in other words, results which make sense at least in principle.”
This creative new neural network tool, published in Physical Review Letters, provides a promising powerful approach to modeling physical phenomena.
Source study: Physical Review Letters – Lattice Gauge Equivariant Convolutional Neural Networks
