The economics of property rights, transaction costs, which are the tendency of individuals to form a dissipative structure to resist the trend of entropy increase, and the existence of costs makes the network converge to a certain path, that is, the finite possibilities that we can observe, which is consistent with the idea of allocation function and path integral.

Marginal analysis is a mathematical method of finding the limit, at which possible trends can be understood, i.e., high-dimensional structures can be extracted.

The logical structure formed by the key information can express more information and thus get closer to the real situation. For example, the conserved sequence of organisms accounts for a larger proportion of the importance distribution function, showing a certain power-law distribution.

The observation of macroscopic effects, such as externalities, is a theoretical matching of information that can precisely define the various properties of specific objects. But this requires a certain amount of resources and costs to do, and we can still roughly divide it into a limited level in the end, which can be regarded as an equilibrium formed by a multi-level game, that is, it may change under the premise of changing the actual situation, and of course we also have a path dependence problem. This measurement of an uncertain object can be equal to the selective expressive relationship between different levels.

The interaction between levels is the basis for the formation of network structure, which is the formulation of various network protocols in computers and changes in the expression patterns of various genes and proteins in biology. This is the spirit of contracts and contracts in economics.

The coupling of multi-level information is an approximation of the real situation, which is a Bayesian computational process. But no matter how hard we try, there are still many unobservable variables that affect the final result, and we can only abstract into a certain x variable and participate in this calculation process, after all, the differentiation of the network is multipath. At the same time, these paths are engaged in a certain game competition, which can form a certain equilibrium and correspond to the actual situation.

Find a mathematical way to describe a specific object and use a model to explain the phenomenon.

Everything has a cost, which is a manifestation of the trend of increasing entropy. This could explain what we think is an unreasonable existence. All of them are network-based probabilistic expressions, but only when the expression probability is high enough can there be the possibility of continuation, thus forming path dependence. Therefore, the nature of emergence is always the most adaptable to the environment, and there is greater efficiency to integrate all aspects of resources. However, this is not a rational theoretical basis for existence, and the situation in the past is different from the current situation, so there needs to be a certain change in order to achieve the optimal adaptability. Since this is a selective expression of the network, it is probabilistic, so there is no good or bad decision, everything depends on the specific fitness of the specific environment. Of course, path dependence makes this equilibrium change very difficult, which is the introduction of other levels of game competition, such as some people's nostalgia and so on.

Selective representation at a particular level can approximate the real situation with arbitrary precision, like the Fourier transform of linear algebra, with a linear combination of a set of bases. This is why we want to explain specific objects using equilibrium formed at different levels of game theory, and the complex results of this multi-level coupling can form a high match. is an important module of our network thinking. It can be abstracted into a locally optimal approach, i.e., the benefits outweigh the costs, so that a specific model is chosen. The problem is that both the benefits and costs are multi-layered, and there are different weights in different levels of competition. Theoretically, a more certain weight distribution ratio should be determined, but in reality, decisions are made more based on empirical judgments, which is also a Bayesian calculation process, which is actually equivalent to the former. In any case, it is a model formed by information gathering ~ logical judgment ~ decision. This is the path collapse of the network. In addition to these multi-level considerations, there are also these levels of considerations, that is, in most cases convergence, that is, they account for a very low proportion in the process of decision formation, after all, this kind of computing resources are not available to everyone. The final result is also multi-layered (the stability of the network is considered, diversity equals stability, just as the genetic diversity of organisms can make a population always have some individuals able to survive in a drastically changing environment), but the observable results are limited. The former is the proportion of states like the hidden Markov sequence, and the latter is the specific sequence expression.

The construction of equivalence, like the law of conservation of energy, is the result of a multi-level game, and the price is based on the selective expression of value. If the two do not match, the equivalence transformation of the network automatically causes it to reach a new equilibrium. For example, the gray income of doctors is to a certain extent to make up for the current unreasonable medical pricing system. This is the regulating mechanism of the market.

Everything has a cost, and that's something we have to consider when modeling. Especially for indicators that are difficult to quantify, we can't use simple cases and non-cases to do simple addition and subtraction, which is a multi-level game that needs to be considered and analyzed at the level of big data.

Grasping the key problems from complex phenomena and grasping the key points is a kind of path collapse of experience-based Bayesian networks.