HOW DOES PHYSICS PLAY A ROLE IN DEEP LEARNING?
If deep learning can extend far beyond the plains, physics has something to do with it. The modern world sees artificial intelligence as the new physics that starts from the same principle of learning, reasoning and application. Physics can be useful in understanding gauge equivariance.
We have to go further back to Albert Einstein’s theory to be able to grasp the meaning. Certainly, the answer lies in learning and understanding particle physics. It turns out that this law has a gauge equivariance. This property wants to enlighten us on certain aspects of the universe and demonstrates that the quantities of particles and their relationships do not depend on gauges or arbitrary reference frames.
Let’s just take the example of a football field that is measured in yards and then in meters. The measurements taken do not give the same figures. This means that physicists always keep a certain consistency in the models of reality, regardless of the unit of measurement or the angle of view when taking a photo. The same is true for gauge NBCs. They adopt the same hypothesis as physicists.