Complex iteration and nesting, recursive operations, can traverse all possibilities at the same time, but not actually search, but like the calculation of SVM's kernel function, it can build low-dimensional simple operations equivalent to high-dimensional complex operations, so that the final optimal solution can be obtained with lower computing resources, just like quantum computing considers all possibilities at the same time and then anneals. Like the construction of a function, it is the convergence of the parameters of the neural network algorithm.

There is only one world line, and while there are theoretically multiple possibilities, such as parallel spaces, which are objects to be considered in calculations, in practice, the result of the final convergence can be seen as a selective expression (parallel worlds are constantly creating, but at the same time shrinking, with the emphasis on specific nodes having a greater impact).

The convergence of the world line, this is a dynamic process, just like the decision tree decision, different decision sequences may lead to changes in the decision tree structure of the limbs, such as the adjustment of parameters, which may be a Bayesian process, it can be said that the algorithm is not fixed, but can adjust the parameters with the input of the data, so that the same input produces different outputs at different times.

In the realm of computing, there is no separation of time and space, everything is equivalent, so what happens in the future may affect the past, continuity is actually an illusion, and we have done a lot of discrete iterative nesting.

Therefore, the convergence of the possibilities brought about by different decisions makes the final logical chain relatively fixed, that is, the causal cycle that we can understand.

This paradoxical existence is actually a structure of cyclic computing, which ultimately requires the convergence of the powerful computing power provided by quantum computers to map to reality.

Or at this level, what we call the universe is a Turing machine with super-computing power. We can understand that everything is an expression of probability, an application of Gödel's law of incompleteness in reality: in a formal system of axioms, there are always propositions that can neither be verified nor falsified.

These unreasonable existences make possible the existence of axiomatic systems that we can deduce, like the tip of the iceberg, the sea water allows our knowledge icebergs to float at sea level, and constantly increase the total amount of icebergs above sea level, that is, the scope of human knowledge of the universe is constantly expanding.

On this basis, there are logical, common sense, emotion, and other objects that we are accustomed to. So the question is, how do we use this idea to write and implement this algorithm?

I don't know