According to foreign media reports, neural network is an interconnected computing system similar to the human brain in structure. The artificial neural network created by computer scientists consists of various nodes that process and transmit signals. These nodes are similar to biological neurons. The network can change as it is used – for example, by increasing the weight given to certain nodes and connections, so that the entire network can “learn” as it runs. For example, let the neural network look at a group of pictures of cats, and then the neural network can learn the characteristics of cats by themselves, and then find the cats from a group of animals. < / P > < p > but, fundamentally, everything we know may be one of these systems. This concept is proposed to coordinate the so-called “classical” physics and quantum mechanics, which has been a problem for a long time. < p > < p > we’re not just saying that artificial neural networks can be used to analyze physical systems or discover laws of Physics – we’re talking about how the world around us works, which is a very bold proposition that can be seen as a suggestion for all theories, and because of this, it should also be easy to falsify. All that needs to be done is to find a physical phenomenon that cannot be explained by neural networks. However, it is easier said than done. < / P > < p > when considering the operation of the whole universe on a large scale, physicists usually use a specific set of theories as tools. These tools are “classical mechanics” based on Newton’s law of motion and Einstein’s theory of relativity, which can be used to explain the relationship between time and space and how mass distorts the structure of space-time to produce gravitational effects. But for physical phenomena at the atomic and subatomic levels, physicists have found that so-called “quantum mechanics” can better explain the universe. < / P > < p > in this theory, physical quantities such as energy and momentum are limited to discrete values rather than continuous values, i.e. “quantum”. All objects are particle and wavy, and finally, measuring this behavior changes the object itself. Finally, this feature is called Heisenberg’s “uncertainty principle”. In short, some related attributes, such as the position and speed of an object, cannot be accurately measured at the same time, so there is a theory of probability. < / P > < p > although these theories can explain the universe well in their respective scales, physicists have always wanted to find a way to coordinate these theories into a universal theory, which is sometimes called the “quantum gravity puzzle”. If the two theories are to be combined, gravity (described by general relativity as the curvature of matter / energy in space-time) may need to be made up of quanta, and thus have its own basic particle, graviton. Unfortunately, the influence of a single graviton on matter will be very weak, which makes the quantum gravity theory almost impossible to verify, so it is impossible to judge whether the theory is correct or not. However, the concept of neural network has a new way. Instead of trying to reconcile general relativity and quantum mechanics into a basic universal theory, we believe that there are deeper reasons behind the behaviors observed in this theory. < / P > < p > in the new study, scientists set out to establish a model to observe how neural networks work in a system with a large number of single nodes. Under certain conditions, the learning behavior of neural networks can be approximately explained by quantum mechanical equations, but other times it needs to be explained by classical physical laws. This theory may also be used to explain the so-called “hidden variables”, i.e., the unknown properties of objects proposed by some physicists, so as to explain the inherent uncertainty in most quantum mechanics theories. In the emerging quantum mechanics, hidden variables are the states of each neuron, and they are trainable variables – such as deviation vectors and weight matrices, which are quantum variables Quantity. < / P > < p > in such a neural network, everything from particles and atoms to cells to everything else will slowly emerge in a process similar to evolution / natural selection. The structure of micro neural network is more stable, while other structures are less stable. The more stable the structure, the more likely it will survive evolution, while the unstable structure will be eliminated. < / P > < p > at the smallest scale, it is natural to select structures with very low complexity, such as neuron chains. However, the larger the scale, the more complex the structure is. However, this process should be limited to a specific length range. So the concept means that everything we see around us – particles, atoms, cells, observers, etc. – is the result of natural selection. Most people in physics don’t seem to buy in on whether the “universe is a neural network theory” will be accepted or not. 99% of physicists will tell you that quantum mechanics is the main theory, and everything should come from it in some way. However, this principle contradicts the viewpoint that quantum mechanics is not the basic law. Experts in physics and machine learning are also skeptical about this new theory. (Henglin) < A= https://ibmwl.com/category/global-tech/ target=_ blank>Global Tech