Our world is a causal system, our world is based on cause and effect. To understand our world, to live in it effectively, to behave in the ...
Our world is a causal system, our world is based on cause and effect. To understand our world, to live in it effectively, to behave in the world in a truly meaningful way, it is necessary to recognize the cause-and-effect relationships in the functioning of our world.
There are innumerable correlations in the world, only a part of which is the causal relationship. A causal relationship is a necessarily connected sequential correlation. Causality is the essence of how our world works.
A causal relationship is also the medium of meaning. Any meaning can be described as a causal relationship. Consequently, anything that can recognize the cause-and-effect relationships of the world is, and only is, capable of understanding the world.
Non-causal correlations also play an important role in our world. These correlations represent the properties of things that exist in the world.
Artificial intelligence is able to recognize correlations more effectively than the human brain modeling the world due to its suitable mode of operation and lack of performance limitations, but it does not have a procedure for selecting cause-and-effect relationships between correlations, i.e. for recognizing meaning. Therefore, artificial intelligence can be used effectively, even more efficiently than the human brain, in areas where it is necessary to recognize connections in large data sets and where correlations are essentially cause-and-effect relationships.
Because of its operating mechanism, the human brain is effectively capable of recognizing and selecting cause-and-effect relationships among correlations. The human brain still retains its superiority in recognizing the essence and meaning of the world.
Basically, even the human brain can only recognize correlations, because only correlations can be recognized by observation. However, as a result of its operating procedures, the brain is capable of functions that can rank relations according to order, and as a result of ranking, causal relations can be established probabilistically. The brain is capable of the emergent function of prediction error minimization, which is functionally equivalent to the implementation of cause-and-effect recognition, because of its mode of operation, which is based on neural resonances that are constantly formed by the continuous external and internal conditioning of the firing connected neurons.
The brain works in a fundamentally different way than artificial intelligence, which also recognizes correlations. To enable the identification of correlations between data using computational tools, to achieve the current level of artificial intelligence, it was necessary to find the appropriate architecture and suitable mathematical procedures, the implementation of which allowed the simulation of the functioning of the brain on computing devices to recognize correlations that led to the creation of the intelligent function.
The appropriate operation of artificial intelligence that recognizes correlations has been achieved by modeling the architecture of the brain on computing devices, supplementing the intelligent function of recognizing correlations by applying mathematical procedures that can be used natively in the computing environment, but which is completely different from the operating principle of cooperating brain neurons based on resonance. However, recognizing correlations is not a satisfactory way for rationality, for intelligent operation according to causality.
Similar to the development of operations based on the recognition of correlations, the artificial realization of rationality and the recognition of causality must first determine the characteristics of causality and then formulate them in a mathematical language so that they can be implemented on computational devices.
Causality is a correlation where two basic characteristics apply simultaneously, sequentiality and exclusivity, i.e. strict correlation.
The mathematical techniques used to implement artificial intelligence to detect correlations work well. Since cause-and-effect relationships are a subset of correlations, a procedure suitable for selecting causality must select correlations based on the properties of cause-and-effect relationships using mathematical methods suitable for search.
Causal relationships are sequential in nature: a mathematical procedure implemented using computational tools must be able to identify correlations that are sequential. Considering time in the computational environment when selecting from the set of correlations does not present a theoretical difficulty, only an appropriate mathematical method must be used to detect the sequential correlations on the data set.
The causal relationship is a strict correlation: the mathematical procedure implemented by computer technology must be able to select, among the correlations that follow each other in time, those that strictly occur every time. The mathematical procedure for processing the correlations must also be able to identify the exact initial conditions associated with the strictly occurring correlations that occur sequentially, i.e. the closed set of initial conditions that consequently generate the occurring state.
When the initial state that creates the causal relationship is a complex structure, the causal relationship is typically a characteristic correlation, and the final state depends critically on the nature of the initial state. The mathematical procedure implemented on the computational tools that search for causality must be able to recognize the optimal arrangement of initial conditions that create the final state among the strictly sequential correlations.
This function can be implemented either in a passive way or in an active way by observation.
In the case of passive observation, in the event of the appearance of the selected and repeatedly occurring condition among the data characteristic of the monitored system, it is necessary to examine in what form the initial conditions are present, and by examining them, the precise characteristics of the relationship must be determined. Passive monitoring, obviously, presupposes the availability of a large amount of data and requires the activity of the monitored system in order to provide sufficient information for successful recognition.
When sufficient data is not available, active observation is the appropriate method to determine causal relationships. Human intelligence also uses active observation, which is much more effective at identifying cause-and-effect relationships in insufficiently active systems. Active monitoring requires direct interaction with the monitored system.
Active observation is based on actively modifying the initial conditions as causes and observing the effect of the changes on the causal state that occurs. Active observation is directed toward finding the optimal configuration of causes, with the goal of identifying the characteristics that are most effective in producing the state that occurs. A possible way of mathematical formulation of the task could be, for example, the modification of the variables creating the local extreme value (correlation) in such a way that the function reaches the given extreme value with the steepest possible course of the function.
The causal relations always appear in correlations. Artificial intelligence capable of selecting causal relations, i.e. capable of recognizing meaning, can generally behave intelligently with less available data, just like human intelligence.
Artificial intelligence using generative technology, which is currently the most advanced form of intelligence, uses human language to recognize correlations. However, unlike natural systems, human language is not a causal system by nature. In fact, natural language is a universal modeling tool. This is why artificial intelligence that uses only generative methods to recognize correlations in language as a data set can lead to false and untrue conclusions.
However, artificial intelligence optimized for recognizing cause-and-effect relationships would also be efficient and effective at filtering out false and untrue statements, and would be effective in language-based practical application of meaning as an implementation of the philosophical concept.
On the other hand, the universal modeling function of language also means that a set of linguistic data that carries only real relationships and is constructed according to strict internal rules, such as a linguistic model created with the help of computer languages that strictly serve to describe causal systems and apply strict logic, can be a suitable and ideal medium for recognizing cause-and-effect relationships and fundamental meaning. Such a language used to describe the systems of the world and constructed according to strict internal rules can be the ideal and universal form of a data set that can be used by artificial intelligence, which without any theoretical performance limit, could implement the recognition of causal relationships in the world using this language as a data source to build a knowledge system that recognizes the meanings of the world.
No comments