Big Bang: Symmetry is at the heart of physics

2022-06-11 0 By

Bo Vivian from concave the temple qubit QbitAI | public, the big bang happened is that what kind of a few seconds?It’s one of the most complex problems in physics. Take a particular form of existence in the universe just a few millionths of a second after the Big Bang.It’s a super-hot “perfect liquid” that has enormous implications for the exploration of the structure and environment of the universe’s original matter.In the lab, it would have to be at temperatures 150, 000 times hotter than the center of the sun to successfully simulate it.To analyze or process this highly complex physical form, a supercomputer would need such a long time to approximate its form that it would be difficult for a classical AI or CNN to make a meaningful interpretation based on the physical concepts involved.But now a paper in the physics journal PRL proposes a new neural network structure called L-CNN that addresses the above question: How to deal with canonical invariants Before we get into the structure of L-CNN, let’s clarify one thing: what is it that traditional AI and CNN can’t do?Take the “perfect liquid state” mentioned at the beginning, in which protons and neutrons are broken apart and recombined into a new form of matter called quark-gluon plasma (QGP) at extremely high energies and temperatures.When introducing AI to analyze and deconstruct the QGP form, Gauge Symmetry must be considered.Canonical symmetry means that the same event can be described in different ways. For example, we can describe an electron-photon system with a pair of phases and electromagnetic potential, or we can describe it with another pair, and the two descriptions should give the same physical substance.Since physical quantities are normal-invariant, particle fields and their interactions that appear to be mathematically different may actually be the same physical state.Traditional CNN is difficult to calculate or analyze these standard invariants, so it is naturally impossible to obtain meaningful computer simulation results.The l-CNN Lattice Gauge Equivariant neural network mentioned at the beginning is a brand new method that can calculate or analyze the standard invariants that cannot be processed by traditional CNN.The whole method is based on Lattice Gauge theory.At lattice points, gauge invariants are usually described as Wilson loops of different shapes.Specifically, adding a new convolution layer can form Wilson rings of arbitrary shape in the continuous bilinear layer while preserving the Gauge Equivariance.The set of all shrinkable Wilson rings can be generated by the above method, and together with the topological information from the non-shrinkable loop, it is possible in principle to reconstruct all gauge connections.With such a neural network, it is possible to make predictions for many complex systems in physics.Andreas Ipp, the paper’s author, also uses the quark-gluon plasma as an example: L-CNN, for example, can estimate what a quark-gluon plasma will look like at a later point in time without going through every intermediate step.It also ensures that the system produces only results that are not inconsistent with canonical symmetry, that is, results that make sense, at least in principle.This is difficult to achieve by all previous calculation methods, l-CNN undoubtedly provides a new idea for simulating complex physical phenomena.In the future, it will provide more help in exploring the environment in which life first existed, understanding the original state of matter in the universe, black holes and grand unified theory.There are four authors of the paper, all from the Institute of Theoretical Physics at TU Wien.In the lower right corner, David I. Muller, the corresponding author of this paper, is a postdoctoral fellow at the Institute of Theoretical Physics, Vienna University of Technology. His main research areas are high energy physics, lattice gauge fields and machine learning.Paper: reference links: – the – qubit QbitAI, focusing on our signing in the headline number, first learn cutting-edge technology and dynamic