The human brain is a special structure. Humankind has been singled out from the animal kingdom for being most effective in recognizing the ...
The human brain is a special structure. Humankind has been singled out from the animal kingdom for being most effective in recognizing the relationships between the events that take place around and in being able to classify those relationships into causal relationships.
Man-made Artificial Intelligence (AI) also attempts to model and try to overtake the human brain in the ability to recognize correlation. Significant results have already been achieved in the development of Artificial Intelligence.
The brain is a physically limited structure, it can work with a potentially limited amount of data. AI is physically expandable, hence it can handle a potentially unlimited amount of information. Advanced mathematical and computational methods are available on large datasets to recognize relationships between data elements. AI is already more effective than the human brain - especially in the case of large data sets - in recognizing the relationships between data elements. AI is more effective than the human brain, for example, in categorizing images, in rule-based games, and even in processing data and knowledge based on information in the form of natural language (spoken or written).
In data-processing activities in which AI is able to defeat humans, it is characteristic that the relationship between information is also a causal relationship. AI has already defeated humans in these areas thanks to its potentially unlimited computational power and by utilizing applications of advanced mathematical methods recognizing the relationships between data elements.
However, AI performs very poorly compared to the human brain in areas of data processing where the coincidence between the data is not clearly the same as a causal relationship. In these areas, AI often draws silly, obviously erroneous conclusions compared to humans.
The reason for this operational limitation is that the current AI is unable to handle the meaning of the data, unable to work with the purport of the information.
The meaning of things and events is currently a philosophical concept, we cannot assign a well-defined definition to this property. However, based on the consideration above, we can formulate a definition, we can connect the concepts of meaning with a well-defined property.
The meaning of things and events is the causal relationship, the cause and effect connection of information related to that specific thing or event.
The human brain is able to understand the meaning of things and events, able to handle the purport of information, so using the previous definition, the human brain is able to recognize causal relationships between correlated information.
Why is the human brain suitable for this? The human brain is a product of biological evolution. Natural evolution has limited resources at its disposal, so it strives for optimization. Evolution seeks to make the most from the limited resources as possible. If the brain is able to select the true causal relationship from the coincidences in the related information, so it can recognize the meaning of things and events, it can function more efficiently with fewer resources.
However, even the brain can only perceive information and can detect coincidences. The physical construction of the brain, its operation is mechanical, it does not contain a mysterious knowledge or mechanism that fundamentally recognizes the essence and purport of things and events. There is no process in the brain that knows definitively and unambiguously which correlation is relevant, i.e. which is the causal relationship.
We live in the world, but we are only observers of the world. At birth, we know nothing about the world, we have only a few reflexes needed to sustain life at the beginning of our lives. After we are born, we begin to integrate into the environment - whatever it may be - through active observation (examining the environment) and gathering experiences. Our brain is able to select the coincidences, correlations, and even causal relationships, the meaning of things and events from the observed, collected information. With this property of the human brain, it is particularly suitable for adapting to the environment. We are already able to survive even in space.
AI currently has no methods to select cause-and-effect relationships from correlations between related information. The human brain is more suitable for recognizing causal relationships, for recognizing meaning, and essence beyond coincidences.
What are the mechanisms the brain uses to select between coincidences, to excrete cause-and-effect relationships? Obviously, the brain does not know that the selected coincidence is a causal relationship. However, the brain has mechanisms by which it can grade correlations, assign reliability property to a relationship. This ranking ultimately creates how we interpret the world, what is what we consider to be a cause-and-effect relationship.
The strong correlation of relation is supposed to be causal relation, the meaning and essence of things and events.
It is important to emphasize that the brain’s ability to classify correlations by reliability does not mean that the brain is able to objectively recognize cause-and-effect relationships, it does not mean that the brain is able to recognize the objective reality of the world with certainty. A strong correlation is not necessarily a causal relationship. However, mankind is curious about objective reality, therefore it consciously and intentionally enhances the credibility and reliability of the ranking between correlations. In this process of recognizing the world, we call this activity, this procedure as the scientific method.
What are the neural processes that support the ranking of correlations between information by the brain?
Weighting in multiple sensory correlations
The brain is connected to different senses, allowing us to learn about the world from different perspectives. The reality of the world, the cause-and-effect relations are objective, so the coincidences between the information recognized and sent to the brain by the different senses in the case of causal relation cannot contradict each other. The brain uses several senses at the same time and takes them into account when ranking correlations between perceived information, weighting multi-sensory coincidences as classifying them according to reliability.
Active observation
Not only does the brain passively observe the environment, but during the observation, it tests and clarifies the recognized relationships with effector activity - intervening by changing the circumstances. Possible occasional, non-causal coincidences typically do not change proportionally with changing the correlated parameters. By modifying the influencing parameters of the correlation (active observation), coincidences can be weighted, and causal relationships can be presumed.
Creating hierarchical, non-contradictory correlation structure
The world operates in a causal structure. However, an outside observer only perceives correlations. Causal coincidences are distinguished from occasional coincidences by the fact that causal coincidences operate in a relational structure that is non-contradictory. Due to its structure and principle of operation, the brain is particularly suitable for recognizing hierarchical and non-contradictory relationships.
These are the procedures used by the brain, the nervous system, to rank the coincidences between information and, ultimately, to explore causality.
The development of Artificial General Intelligence (AGI) means the development of general causal machines. Cause-and-effect correlations can be identified by weighting the correlating information. Only recognizing correlations during AI development is not enough for efficient function. The weighting of correlations provides the method of recognizing causal relationships and, ultimately, leads to meaning-based IT operation.
No comments