Is there intelligence in modern “Artificial Intelligence”?
This is far from idle and not a rhetorical question. Over the creation of AI worked, and even today tens of thousands of scientific and research scientists, countless different research developers without a degree, coding in countless IT companies, and a lot of simply ambitious enthusiasts who rarely go out of their homes. Over the past half century, many dissertations, monographs, and articles have been written on this topic, and in this area a wide variety of robots and computer programs have been created, claiming, as their authors claim, to be "intellectual." However, to our disillusionment, that desired intelligence, which we have been waiting for so long from all these theories and applied approbations, is hardly visible.
How so - you can be puzzled - so much effort and that, all for nothing? To understand this and understand the root causes of the existing failures, let's try to use the methodology of systems analysis and technotronic terminology, according to which the "intelligence" of any system is understood as the technology of its behavior. So, if we consider human behavior as a “natural-intellectual” being, in this aspect, then it turns out that the basis of human behavior is determined not by one, but by exactly two technologies, which we conditionally call as: Technology “Mind” (TP) and Technology "Mind" (TR).
The essence of their differences lies in the fact that the basis of the TA, whether it is inherent in man or an animal, are reflex-program reactions that arise in response to signals and / or influences coming from the outside world. This technology is perfectly implemented in cybernetics and its machine tool - the computer.
')
But the TR is already based on an analysis of environmental perturbations (changes in the situation) around the individual and on his decision about the need for a reasoned response to the analytical results he himself obtained. And this is done on the basis of the knowledge and accumulated experience of the reasonable individual.
It is known that in TP only one conditional operator is used for analysis: “If A = X, then do Y, otherwise execute Z”. And there are eight of them at once in TR, the main of which are the so-called “urine” and “position” operators. But a computer or even an animal whose intelligence is based on the conditional operator alone is not at all interested in why this [A] should be equal to [X], and whether [A] can be equal to [X], and why it should be done [Y ] and even more so to do [Z], but the human “Mind”, on the contrary, is engaged in this very matter, asking various “difficult” questions and trying to find any ontological answers to them. That is, it is common for “intelligent technologies” to evaluate not a two-digit situation, but a multi-valued situation.
If you now look at the ideas discussed above at specialists working in the field of developing and creating AI, and also about what they are doing in real life, it will turn out that they are all working absolutely on creating something like this, perhaps even very interesting. , but only within the framework of "mental technology", and well, not in the area of "Mind".
Why does this happen? Because of what, in fact, doesn’t TR develop and what are the difficulties here? It turns out that the answer to this question can only be found by studying and knowing such a phantom as “the procedurality of human thinking”, and this thing is very incomprehensible for linguists and cybernetics, and even physiologists-brain therapists did not really succeed in this direction.
To answer for everything here should have been a brainnet, as a scientific discipline studying algorithms of mental activity. It was her responsibility to give software developers the necessary tool for understanding the text or for understanding the context for the purpose of cognitive knowledge, which, in fact, computational linguistics cannot do in its current form, since it is just outside its competence. But, unfortunately, for a long time, the branetics remained a lingual in linguistics in scientific circles and only now begins to gain the necessary strength and weight, gradually turning into a princess and occupying her status to solve the problem of creating the present, not a surrogate AI.
Why is it precisely Braynetics, and not computer linguistics at all, that can help in the creation of Artificial Intelligence? And exactly for the following reason. Natural intelligence, acting as a system of "rational behavior", necessarily uses knowledge about the surrounding world, building up a certain model of the universe in its natural consciousness in order to navigate it. “Psychic robots” that do not have such a consciousness cannot do this at all, because there is no place for knowledge in compiling, nor, especially, for the model of the universe. Cybernetics with linguists simply do not yet know how to create it in machine standards, and therefore operate on data alone. Operations with knowledge are unbearable aerobatics for them. Well, if there is no knowledge, then there is no "rational intelligence."
How so - you can be puzzled - so much effort and that, all for nothing? To understand this and understand the root causes of the existing failures, let's try to use the methodology of systems analysis and technotronic terminology, according to which the "intelligence" of any system is understood as the technology of its behavior. So, if we consider human behavior as a “natural-intellectual” being, in this aspect, then it turns out that the basis of human behavior is determined not by one, but by exactly two technologies, which we conditionally call as: Technology “Mind” (TP) and Technology "Mind" (TR).
The essence of their differences lies in the fact that the basis of the TA, whether it is inherent in man or an animal, are reflex-program reactions that arise in response to signals and / or influences coming from the outside world. This technology is perfectly implemented in cybernetics and its machine tool - the computer.
')
But the TR is already based on an analysis of environmental perturbations (changes in the situation) around the individual and on his decision about the need for a reasoned response to the analytical results he himself obtained. And this is done on the basis of the knowledge and accumulated experience of the reasonable individual.
It is known that in TP only one conditional operator is used for analysis: “If A = X, then do Y, otherwise execute Z”. And there are eight of them at once in TR, the main of which are the so-called “urine” and “position” operators. But a computer or even an animal whose intelligence is based on the conditional operator alone is not at all interested in why this [A] should be equal to [X], and whether [A] can be equal to [X], and why it should be done [Y ] and even more so to do [Z], but the human “Mind”, on the contrary, is engaged in this very matter, asking various “difficult” questions and trying to find any ontological answers to them. That is, it is common for “intelligent technologies” to evaluate not a two-digit situation, but a multi-valued situation.
If you now look at the ideas discussed above at specialists working in the field of developing and creating AI, and also about what they are doing in real life, it will turn out that they are all working absolutely on creating something like this, perhaps even very interesting. , but only within the framework of "mental technology", and well, not in the area of "Mind".
Why does this happen? Because of what, in fact, doesn’t TR develop and what are the difficulties here? It turns out that the answer to this question can only be found by studying and knowing such a phantom as “the procedurality of human thinking”, and this thing is very incomprehensible for linguists and cybernetics, and even physiologists-brain therapists did not really succeed in this direction.
To answer for everything here should have been a brainnet, as a scientific discipline studying algorithms of mental activity. It was her responsibility to give software developers the necessary tool for understanding the text or for understanding the context for the purpose of cognitive knowledge, which, in fact, computational linguistics cannot do in its current form, since it is just outside its competence. But, unfortunately, for a long time, the branetics remained a lingual in linguistics in scientific circles and only now begins to gain the necessary strength and weight, gradually turning into a princess and occupying her status to solve the problem of creating the present, not a surrogate AI.
Why is it precisely Braynetics, and not computer linguistics at all, that can help in the creation of Artificial Intelligence? And exactly for the following reason. Natural intelligence, acting as a system of "rational behavior", necessarily uses knowledge about the surrounding world, building up a certain model of the universe in its natural consciousness in order to navigate it. “Psychic robots” that do not have such a consciousness cannot do this at all, because there is no place for knowledge in compiling, nor, especially, for the model of the universe. Cybernetics with linguists simply do not yet know how to create it in machine standards, and therefore operate on data alone. Operations with knowledge are unbearable aerobatics for them. Well, if there is no knowledge, then there is no "rational intelligence."
Source: https://habr.com/ru/post/144371/
All Articles