Artificial intelligence
Preamble The
whole civilizational history of the development of mankind people sought to become stronger and less vulnerable, which was caused by the need to eternal struggle for vital resources and ensure their own security. Homo-Sapiens always wanted to develop their physical and mental abilities and capabilities. A person wanted to see much more and more clearly than he usually can, wanted to fly high and swim deep, hear from afar and feel sharper, remember more and think faster ...
In almost all of these aspirations, these sought-after goals were usually always achieved by creating any technical amplifying products based on known physical laws. And here people, practically, in all cases achieved success, except, perhaps, one old dream is the creation of “ Artificially Reasonable Intelligence ”, that is, such a device that would be able to think with inhuman speed and operate with Knowledge not worse, what the most advanced thinking beings can do.
When a computer was created in the middle of the 20th century, which magnified man’s counting capabilities and became the most powerful operational accelerator of data processing , the best minds of the academic circles of the leading countries of the planet directed their efforts to turn the computer into the required “ thinking apparatus ”, trying to make operate it with knowledge. But years went by and more than half a century had passed since the beginning of the first developments in this direction, and all the titanic efforts to create a “ reasonable machine ” have so far been in vain.
Here arises a reasonable question: "What is this insurmountable problem does not allow a huge army of ambitious scientists, researchers and enthusiasts all over the world come to the long-sought goal, forcing the computer to think of Knowledge , surely the human conscious activity is not subject to physical devices?"
In the first approximation, the answer to this The question is pretty simple. The computer is able to “recognize and read the Data”, but it completely “ does not understand Knowledge” and, moreover, cannot operate with them, at least up to the present. This thesis requires an expanded explanation of what exactly Knowledge represents in relation to what we call “Artificial Intelligence” (AI).
About Intellect
TermIntellectus is derived from the Latin notion of “knowledge,” “reason,” “understanding,” or the ability to rational thinking.
From the position of AI developers, “ Intellect ” is understood as the System of Technological Behavior of an Object in the surrounding world. Stone, water, wood, animals and human flesh are all Essences, as well as computers or other artificial devices.
As for the term “ Object ”, it is understood as a group of structurally hierarchically organized elements ,independently identified and existing in the Spatiotemporal continuum, registered by the human sense organs, united as a common Formwithin common boundaries, and with a single Entity based on physical laws.
In principle, all the physical Objects on Earth known since ancient times that satisfy this requirement are divided into the following four categories:
1. Physio-minerals and alloys . (Inanimate inanimate unreasonable substances);
2. Phyto-flora . (Living inanimate unreasonable plants in the period of their growth or development);
3. Bio-fauna . (Live animate unreasonable animals during their biological life);
4. Mento-Ratio Sphere . (Living animated rational people in the period of their “intelligent life”).
Within the framework of this classification, there are, respectively, the inherent four forms of behavior in the Natural Intelligence (EI) format :
1. Unreasonable Intelligence of inanimate inanimate Mineral Nature Objects;
2. Unintelligent Intellect of living inanimate Flora Objects;
3. The unintelligent Intellect of the living animate Objects of the fauna;
4. Reasonable Intellect of living animate Objects together with their conscious Subjects.
In general, the consciousness of people identifies the following eight behavioral functions for all four categories of Objects with an
Unified Unity : Fn1 => Existence => Be,
Exist Fn2 =>Transformation => Become, Change
Fn3 => Join => Feel, Feel
Fn4 => Reflection => React,
Excite Fn5 => Action => Act, Produce, Make
Fn6 => Transport => Move, Move, Move
Fn7 => Intention => Want, Want,
Wonder Fn8 => Thinking => Think, Think, Realize, Speak
Note that any processes in which the Objects of the Universe participate are described by verbs in natural languages with exactly these eight behavioral functions. At the same time, different categories of Objects of the physical world have, respectively, a different set of behavioral functions.
1. A simple stone has only 4 first functions => Fn1 - Fn4.
That is, Objects of the first category can Exist , Change, Register and React. So, if you repeat these 4 natural behavioral functions on an artificial level, you will get 4 types of “artificiality”:
- Artificial Existence ;
- Artificial Transformation ;
- Artificial Registration ;
-Artificial Reflection .
2. In a typical plant, the behavioral functions are 2 more units => Fn1 - Fn6.
Here, Objects of the second category can still be Produced and even Transported. Therefore, their technical counterparts may additionally have two more types of “artificiality”:
- Artificial Productivity ;
- Artificial Transportation .
3. In different animals, up to primates, their number reaches 7 functions => Fn1 - Fn7
Therefore, to copy the behavior of animals, add another:
- Artificial Intention .
4. But people have all eight functions at maximum => Fn1 - Fn8
It follows that Artificially Reasonable Intelligence should be able to perform the function:
- Artificial Thinking .
“Homo Sapiens” (Homo Sapiens) differs from Objects of the first three categories in that, in addition to his material body, as “ physical objectivity ”, he also has “ mental subjectivity ” in the form of Consciousness . That is, people, or rather, their bodies are not only Objects from the standpoint of physical (material) nature, but also act as a mental (conscious) Subject .
It is easy to see from here that only “ Artificial Thinking ” can actually add to the machine of subjectivity, making it truly “An Intelligent Object . ”
Thus, we note that all AI systems created in the world at the moment are INEFINITE , since they do not have mental abilities.
Analysis and Synthesis
At the basis of the functioning of any Substance, whether it is a physical Object, for example, the human body or the same mental Subject in the form of its Consciousness, are, as you know, two technologies, conventionally referred to as:
1. Analysis of the registered Situation ;
2. Synthesis Reactions to the situation;
Objects of Fauna and Flora, along with all inanimate nature, including human bodies, have parametric registrationphysical effects, which is a physical Analysis of the Situation, and are also capable of reflex reactions , which, respectively, is a physical Synthesis of Reaction.
In general, the physical Analysis used by certain Objects of the physical world is based on the Fn3 function (“ Registration ”), assigned to various sensory organs and receptors of the Objects of the physical world, in which there are 16 types of Receptors:
- Vizio (visual, light) )
- Audio (acoustic, sound)
- Olfacio (olfactory)
- Gevzio(taste)
- THERMO (heat)
- Proprio (vestibular)
- Taktilio (kinesthetic, tactile, contact)
- dermis (skin, itch)
- RE (electrical)
- Magnetite (magnetic)
- Radio (RF)
- Roentgen (X-ray radiation )
- Emotion (neuro-emotional)
- Eksio (sensual)
- Nestesio (pain, neuro-pain)
- Nevropsio(neuro-neural-impulse)
Note that, practically, all of the above-mentioned types of receptors have linear parametric dimensions that fully comply with the data format , therefore, physical Analysis is Data Analysis . Accordingly, physical synthesis is based solely on the results of physical data analysis. From this it follows that all these physio-functions and the set reaction rules are relatively easy to model in computer programs in the form of physical Analysis and physical Synthesis .
Thus, all modern AI devices are implemented exclusively by “ Physical REACTION TECHNOLOGY (TFR) based on the result of a physical Analysis", And represent a simple physical response of the Object to physical stimuli.
But the Analysis, which is used by the mental Substance in the form of human Consciousness, uses the function Fn8 (“ Thinking ”). That is, a person, as an Object of the 4th category, owning the subjective function of the Mind, becomes thinking only if he learns to CONSCIOUSly speak , think , understand , comprehend , create , compose ...
Hence we have that “Artificially Thinking Machine” should have “ The technology of mental reaction (TSR) according to the result of mental analysis "that is, the TSR must replicate the Object's mental response to mental situations.
Summarizing, we note that only people, in addition to physical Analysis (due to their physiological receptors), also possess mental ( mental ) Analysis (due to their consciousness apparatus), that is, a person is able to make a semantic assessment of the situation , as well as people own mental ( mental ) Synthesis, that is, a person is able to apply a reasonable logic of response , which, in general, corresponds to the format of Knowledge , and not Data.
If we investigate more deeply the reason for the difference between TFR and TSR, then we can come to the unequivocal conclusion that all FM Objects operate exclusively with Dataand use for this (together with different iterative formulas) only one single implicative (conditional) Operand of the format:
“If A = X, then do Y, otherwise do Z” .
At the same time, any physical dimensional value always acts as a Data, which is compared with a given Reference Data. For example, if you gently tap a finger along an ordinary stone, then it will certainly not collapse. And if it is powerfully hit with a sledgehammer, then , most likely, the stone should collapse or at least be damaged as a result of overcoming the strength of the stone by the force of an impact. That is, it can be said that the stone " feels”The force of the strike, comparing it with some of its internal standard, and“ behaves ”strictly in accordance with the measured parameters of the impact: the force of the impact, the area of the impact, the specific power of the impact and the other according to the behavior program for the stone set by the physical world.
From the standpoint of a computer simulating the natural behavior of a stone, these parameters are input signals in any computer-digital format of the Data . Therefore, today, the developers of AI were able to quite easily achieve imitation of Natural Intellect with varying degrees of completeness and similarity only for the first seven functions inherent in the EI. And this is called “ Seek AI ” or SAI - (“Weak AI”).
In principle, the same ordinary computer executing a program is the most typical example of SAI.
But the eighth function - Thinking , which can turn an unreasonable animal or even Mowgli into Homo Sapiens - has become a real stumbling block in the way of the creators of “ Power AI ” - PAI (“Strong AI”) or as it is also called Artificial Super Intelligent - ASI (“ Artificial Super Intellect - ICI ”). And here the matter is that “a rational person”, as mentioned earlier, unlike the other three categories of Nature Objects, is able to operate not only with Data, but also with Knowledge !
The global scientific and technical community involved in the creation of Super Intellect understands this well, and is confident that if the desired computer format of the reference Knowledge is created , it will be possible to create, by analogy with the current Data processing technology, the Knowledge Processing technology using relevant knowledge base(BR) and believing that if we replace the arguments in the implicative Operand in the form of Data with the arguments in the form of Knowledge, then the problem will be solved immediately. However, the failure of all the numerous attempts to find a solution for modeling the machine operation of Knowledge for such a long period of time has led today only to the fact that the overwhelming majority of the developers of the FIC either refused to continue working in this direction, or chose a different way - to create technologies based on so-called “ Artificial Neural Networks ” (INS), relying on the fact that sooner or later INSs can become “reasonable”.
Currently, no one can tell whether the INSs will lead to the desired result or not. Nevertheless, now developed countries are investing heavily in this direction. For example, China announced that it intends to invest 2 billion US dollars in the construction of the “City of Artificial Intelligence”, where more than 400 IT companies will be concentrated, the task of which is to make China by 2030 the world leader in Strong AI. The company Ilona Mask is now competing with this, declaring that he is going to invest 1 billion US dollars in the creation of ICI.
Problem
Now let us consider the difficulty of solving the creation of machine format of Knowledge and why the leading Super Intelligence developers in the world have put a cross on this direction. The consolidated response of many experts on this subject states that, unlike the term " This ", for the concept of " Knowledge ", unfortunately, there is not yet a clear, generally accepted and unambiguous definition, on the basis of which one could create for the Knowledge machine format so that this “digitized Knowledge” could already operate the computer.
Knowledge is always more than this. Knowledge and Data correlate with each other as, for example, System and Elementwhere the latter is part of the System, but the Element does not define the System itself. Only the sum of the properties and abilities of all the Elements included in the System can provide new integral functionality, but in a different format than the essence of a single Element, like, for example, individual frames of a film (Elements) cannot give an idea of the plot of the film (System), until the Play e den film itself and do not realize the audience.
In cybernetics, there already exists a data recognition technology , without which, in fact, they could not be operated on. For this purpose, a database is created in the computer memory.(DB), as, for example, the dictionary of some natural language in case of need for recognition, for example, text graphemes. Using such a database, the input Data (grapheme) is compared with the similar Data (graphemes) available in the database in order to identify and uniquely identify the input Data (grapheme). Therefore, it would be natural to solve the problem by going by analogy and creating a BZ to compare all input Knowledge with it.
Or another example - these are Phrase combinations and their Meaning . Here are just Baz Smyslov , as BZ, yet no one has created.
And who, in general, could help here with the solution of the problem of the definition of the term “Knowledge”? If we apply for academic assistance, then the closest science to this topic is linguistics, which, at first glance, would have to cope with such a difficulty, but it, unfortunately, again, deals exclusively with Data, that is, words, graphemes, semes and all that stuff, but completely incompetent in the field of human thinking and knowledge theory. Despite the section “Semantics”, linguistic theory does not develop a Base of Meanings, which should contain various phrases in natural languages. This task goes into the field of Big Data, but experts in the field of Big Data are also far from solving this problem.
Other sciences and theories - from cognitive science and artificial neural networks to philology with cryptography - are even further removed from the desired solution. Therefore, the main reason for the fact that the Machine Knowledge Format (MFZ) was not created until now is the dominant principle:
“It is impossible to program something that cannot be described in known truths”.
Thus, it turns out that if you develop the desired MDF and create a BRZ on its basis, and also if you teach a computer to extract Knowledge from information and operate on it, then the problem of creating a “Strong AI” is resolved, as it were, by itself.
However, unfortunately, the desired result cannot be achieved even at this stage, even if someone could present the Knowledge in machine format. MFZ and BZ - this is only half the battle. And the whole essence lies in the fact that the BR is needed, in turn, to extract Sense from information, because people in reality exchange Sense , not Knowledge and Data, with the help of which, in fact, people only encode and decode Sense, using language tools. And here a new problem arises - how can you teach a computer to understand the meaning of a text or even a context, like how people can do it? After all, the human consciousness operates, as is known, with Thoughts !
Meaning
Homo sapiens, as is known, does not own telepathy, unlike some animal species, since once humanity lost 9 pairs of chromosomes responsible for telepathy. Therefore, in order to exchange information, people have to encode their Thoughts by various language means. The idea of any author, described by him, for example, with the help of the Russian language, becomes information for all of his addressees who speak the given language, and this idea is later required by the recipients to be decrypted with the help of Knowledge. That's what, ultimately, humanity needs Knowledge!
So, Knowledge is necessary to restore the author's Thoughts from textual information, which, in fact, is nothing more than extracting Sense from natural language messages, that is, without Knowledge, it is impossible to understand what is being said through speech, texts, gestures ... Therefore, the thinking computer in addition to the BZ need to have more and technology for extracting meaning !
Thus, in order to create an artificially-thinking machine, it is necessary to develop such software that could successfully decode the author's Idea, encrypted, for example, in a natural-language text. However, to achieve success, it is necessary to understand what is the procedurality of the Thinking of the human consciousness, from what other Operands, besides the usual implicative, does the “thinking process” consist, etc.? In other words, in order to develop the technology for decrypting author's Thought, you need to know the coding technology itself!
Approaches
Scientists and developers of FIC to obtain the desired offspring on the basis of a computer have already tried many approaches from various sciences and theories, attracting linguistics, statistics, and stochastics and more. In recent years, unsuccessfully having tried everything previously known, the majority of enthusiasts, as mentioned above, in creating computer thinking have directed their main efforts to developing ANNs and trying to create FIS using the new direction based on the so-called “ machine learning ” technology , thanks to which achieved - it should be noted - very significant results in the recognition of audio-visual objects.
However, this approach also usesthe technology of identifying any patterns by statistical data processing and other mathematical tricks , independently, without using the technology of semantic processing of natural language information (texts), is completely incapable of processing Knowledge (at least today and in the near future) and, all the more , is impotent in terms of extracting Sense from an information code, which is a natural-language text.
In addition, at the moment, based on information from open world sources, there is not a single development known to the author even at the level of an idea that could teach a computer to operate Knowledge and extract Meaning from the context, restoring copyright Thoughts.
Output
And where is the light that will lead us out of this impasse? This readers will learn in the sequel to this publication.
whole civilizational history of the development of mankind people sought to become stronger and less vulnerable, which was caused by the need to eternal struggle for vital resources and ensure their own security. Homo-Sapiens always wanted to develop their physical and mental abilities and capabilities. A person wanted to see much more and more clearly than he usually can, wanted to fly high and swim deep, hear from afar and feel sharper, remember more and think faster ...
In almost all of these aspirations, these sought-after goals were usually always achieved by creating any technical amplifying products based on known physical laws. And here people, practically, in all cases achieved success, except, perhaps, one old dream is the creation of “ Artificially Reasonable Intelligence ”, that is, such a device that would be able to think with inhuman speed and operate with Knowledge not worse, what the most advanced thinking beings can do.
When a computer was created in the middle of the 20th century, which magnified man’s counting capabilities and became the most powerful operational accelerator of data processing , the best minds of the academic circles of the leading countries of the planet directed their efforts to turn the computer into the required “ thinking apparatus ”, trying to make operate it with knowledge. But years went by and more than half a century had passed since the beginning of the first developments in this direction, and all the titanic efforts to create a “ reasonable machine ” have so far been in vain.
Here arises a reasonable question: "What is this insurmountable problem does not allow a huge army of ambitious scientists, researchers and enthusiasts all over the world come to the long-sought goal, forcing the computer to think of Knowledge , surely the human conscious activity is not subject to physical devices?"
In the first approximation, the answer to this The question is pretty simple. The computer is able to “recognize and read the Data”, but it completely “ does not understand Knowledge” and, moreover, cannot operate with them, at least up to the present. This thesis requires an expanded explanation of what exactly Knowledge represents in relation to what we call “Artificial Intelligence” (AI).
About Intellect
TermIntellectus is derived from the Latin notion of “knowledge,” “reason,” “understanding,” or the ability to rational thinking.
From the position of AI developers, “ Intellect ” is understood as the System of Technological Behavior of an Object in the surrounding world. Stone, water, wood, animals and human flesh are all Essences, as well as computers or other artificial devices.
As for the term “ Object ”, it is understood as a group of structurally hierarchically organized elements ,independently identified and existing in the Spatiotemporal continuum, registered by the human sense organs, united as a common Formwithin common boundaries, and with a single Entity based on physical laws.
In principle, all the physical Objects on Earth known since ancient times that satisfy this requirement are divided into the following four categories:
1. Physio-minerals and alloys . (Inanimate inanimate unreasonable substances);
2. Phyto-flora . (Living inanimate unreasonable plants in the period of their growth or development);
3. Bio-fauna . (Live animate unreasonable animals during their biological life);
4. Mento-Ratio Sphere . (Living animated rational people in the period of their “intelligent life”).
Within the framework of this classification, there are, respectively, the inherent four forms of behavior in the Natural Intelligence (EI) format :
1. Unreasonable Intelligence of inanimate inanimate Mineral Nature Objects;
2. Unintelligent Intellect of living inanimate Flora Objects;
3. The unintelligent Intellect of the living animate Objects of the fauna;
4. Reasonable Intellect of living animate Objects together with their conscious Subjects.
In general, the consciousness of people identifies the following eight behavioral functions for all four categories of Objects with an
Unified Unity : Fn1 => Existence => Be,
Exist Fn2 =>Transformation => Become, Change
Fn3 => Join => Feel, Feel
Fn4 => Reflection => React,
Excite Fn5 => Action => Act, Produce, Make
Fn6 => Transport => Move, Move, Move
Fn7 => Intention => Want, Want,
Wonder Fn8 => Thinking => Think, Think, Realize, Speak
Note that any processes in which the Objects of the Universe participate are described by verbs in natural languages with exactly these eight behavioral functions. At the same time, different categories of Objects of the physical world have, respectively, a different set of behavioral functions.
1. A simple stone has only 4 first functions => Fn1 - Fn4.
That is, Objects of the first category can Exist , Change, Register and React. So, if you repeat these 4 natural behavioral functions on an artificial level, you will get 4 types of “artificiality”:
- Artificial Existence ;
- Artificial Transformation ;
- Artificial Registration ;
-Artificial Reflection .
2. In a typical plant, the behavioral functions are 2 more units => Fn1 - Fn6.
Here, Objects of the second category can still be Produced and even Transported. Therefore, their technical counterparts may additionally have two more types of “artificiality”:
- Artificial Productivity ;
- Artificial Transportation .
3. In different animals, up to primates, their number reaches 7 functions => Fn1 - Fn7
Therefore, to copy the behavior of animals, add another:
- Artificial Intention .
4. But people have all eight functions at maximum => Fn1 - Fn8
It follows that Artificially Reasonable Intelligence should be able to perform the function:
- Artificial Thinking .
“Homo Sapiens” (Homo Sapiens) differs from Objects of the first three categories in that, in addition to his material body, as “ physical objectivity ”, he also has “ mental subjectivity ” in the form of Consciousness . That is, people, or rather, their bodies are not only Objects from the standpoint of physical (material) nature, but also act as a mental (conscious) Subject .
It is easy to see from here that only “ Artificial Thinking ” can actually add to the machine of subjectivity, making it truly “An Intelligent Object . ”
Thus, we note that all AI systems created in the world at the moment are INEFINITE , since they do not have mental abilities.
Analysis and Synthesis
At the basis of the functioning of any Substance, whether it is a physical Object, for example, the human body or the same mental Subject in the form of its Consciousness, are, as you know, two technologies, conventionally referred to as:
1. Analysis of the registered Situation ;
2. Synthesis Reactions to the situation;
Objects of Fauna and Flora, along with all inanimate nature, including human bodies, have parametric registrationphysical effects, which is a physical Analysis of the Situation, and are also capable of reflex reactions , which, respectively, is a physical Synthesis of Reaction.
In general, the physical Analysis used by certain Objects of the physical world is based on the Fn3 function (“ Registration ”), assigned to various sensory organs and receptors of the Objects of the physical world, in which there are 16 types of Receptors:
- Vizio (visual, light) )
- Audio (acoustic, sound)
- Olfacio (olfactory)
- Gevzio(taste)
- THERMO (heat)
- Proprio (vestibular)
- Taktilio (kinesthetic, tactile, contact)
- dermis (skin, itch)
- RE (electrical)
- Magnetite (magnetic)
- Radio (RF)
- Roentgen (X-ray radiation )
- Emotion (neuro-emotional)
- Eksio (sensual)
- Nestesio (pain, neuro-pain)
- Nevropsio(neuro-neural-impulse)
Note that, practically, all of the above-mentioned types of receptors have linear parametric dimensions that fully comply with the data format , therefore, physical Analysis is Data Analysis . Accordingly, physical synthesis is based solely on the results of physical data analysis. From this it follows that all these physio-functions and the set reaction rules are relatively easy to model in computer programs in the form of physical Analysis and physical Synthesis .
Thus, all modern AI devices are implemented exclusively by “ Physical REACTION TECHNOLOGY (TFR) based on the result of a physical Analysis", And represent a simple physical response of the Object to physical stimuli.
But the Analysis, which is used by the mental Substance in the form of human Consciousness, uses the function Fn8 (“ Thinking ”). That is, a person, as an Object of the 4th category, owning the subjective function of the Mind, becomes thinking only if he learns to CONSCIOUSly speak , think , understand , comprehend , create , compose ...
Hence we have that “Artificially Thinking Machine” should have “ The technology of mental reaction (TSR) according to the result of mental analysis "that is, the TSR must replicate the Object's mental response to mental situations.
Summarizing, we note that only people, in addition to physical Analysis (due to their physiological receptors), also possess mental ( mental ) Analysis (due to their consciousness apparatus), that is, a person is able to make a semantic assessment of the situation , as well as people own mental ( mental ) Synthesis, that is, a person is able to apply a reasonable logic of response , which, in general, corresponds to the format of Knowledge , and not Data.
If we investigate more deeply the reason for the difference between TFR and TSR, then we can come to the unequivocal conclusion that all FM Objects operate exclusively with Dataand use for this (together with different iterative formulas) only one single implicative (conditional) Operand of the format:
“If A = X, then do Y, otherwise do Z” .
At the same time, any physical dimensional value always acts as a Data, which is compared with a given Reference Data. For example, if you gently tap a finger along an ordinary stone, then it will certainly not collapse. And if it is powerfully hit with a sledgehammer, then , most likely, the stone should collapse or at least be damaged as a result of overcoming the strength of the stone by the force of an impact. That is, it can be said that the stone " feels”The force of the strike, comparing it with some of its internal standard, and“ behaves ”strictly in accordance with the measured parameters of the impact: the force of the impact, the area of the impact, the specific power of the impact and the other according to the behavior program for the stone set by the physical world.
From the standpoint of a computer simulating the natural behavior of a stone, these parameters are input signals in any computer-digital format of the Data . Therefore, today, the developers of AI were able to quite easily achieve imitation of Natural Intellect with varying degrees of completeness and similarity only for the first seven functions inherent in the EI. And this is called “ Seek AI ” or SAI - (“Weak AI”).
In principle, the same ordinary computer executing a program is the most typical example of SAI.
But the eighth function - Thinking , which can turn an unreasonable animal or even Mowgli into Homo Sapiens - has become a real stumbling block in the way of the creators of “ Power AI ” - PAI (“Strong AI”) or as it is also called Artificial Super Intelligent - ASI (“ Artificial Super Intellect - ICI ”). And here the matter is that “a rational person”, as mentioned earlier, unlike the other three categories of Nature Objects, is able to operate not only with Data, but also with Knowledge !
The global scientific and technical community involved in the creation of Super Intellect understands this well, and is confident that if the desired computer format of the reference Knowledge is created , it will be possible to create, by analogy with the current Data processing technology, the Knowledge Processing technology using relevant knowledge base(BR) and believing that if we replace the arguments in the implicative Operand in the form of Data with the arguments in the form of Knowledge, then the problem will be solved immediately. However, the failure of all the numerous attempts to find a solution for modeling the machine operation of Knowledge for such a long period of time has led today only to the fact that the overwhelming majority of the developers of the FIC either refused to continue working in this direction, or chose a different way - to create technologies based on so-called “ Artificial Neural Networks ” (INS), relying on the fact that sooner or later INSs can become “reasonable”.
Currently, no one can tell whether the INSs will lead to the desired result or not. Nevertheless, now developed countries are investing heavily in this direction. For example, China announced that it intends to invest 2 billion US dollars in the construction of the “City of Artificial Intelligence”, where more than 400 IT companies will be concentrated, the task of which is to make China by 2030 the world leader in Strong AI. The company Ilona Mask is now competing with this, declaring that he is going to invest 1 billion US dollars in the creation of ICI.
Problem
Now let us consider the difficulty of solving the creation of machine format of Knowledge and why the leading Super Intelligence developers in the world have put a cross on this direction. The consolidated response of many experts on this subject states that, unlike the term " This ", for the concept of " Knowledge ", unfortunately, there is not yet a clear, generally accepted and unambiguous definition, on the basis of which one could create for the Knowledge machine format so that this “digitized Knowledge” could already operate the computer.
Knowledge is always more than this. Knowledge and Data correlate with each other as, for example, System and Elementwhere the latter is part of the System, but the Element does not define the System itself. Only the sum of the properties and abilities of all the Elements included in the System can provide new integral functionality, but in a different format than the essence of a single Element, like, for example, individual frames of a film (Elements) cannot give an idea of the plot of the film (System), until the Play e den film itself and do not realize the audience.
In cybernetics, there already exists a data recognition technology , without which, in fact, they could not be operated on. For this purpose, a database is created in the computer memory.(DB), as, for example, the dictionary of some natural language in case of need for recognition, for example, text graphemes. Using such a database, the input Data (grapheme) is compared with the similar Data (graphemes) available in the database in order to identify and uniquely identify the input Data (grapheme). Therefore, it would be natural to solve the problem by going by analogy and creating a BZ to compare all input Knowledge with it.
Or another example - these are Phrase combinations and their Meaning . Here are just Baz Smyslov , as BZ, yet no one has created.
And who, in general, could help here with the solution of the problem of the definition of the term “Knowledge”? If we apply for academic assistance, then the closest science to this topic is linguistics, which, at first glance, would have to cope with such a difficulty, but it, unfortunately, again, deals exclusively with Data, that is, words, graphemes, semes and all that stuff, but completely incompetent in the field of human thinking and knowledge theory. Despite the section “Semantics”, linguistic theory does not develop a Base of Meanings, which should contain various phrases in natural languages. This task goes into the field of Big Data, but experts in the field of Big Data are also far from solving this problem.
Other sciences and theories - from cognitive science and artificial neural networks to philology with cryptography - are even further removed from the desired solution. Therefore, the main reason for the fact that the Machine Knowledge Format (MFZ) was not created until now is the dominant principle:
“It is impossible to program something that cannot be described in known truths”.
Thus, it turns out that if you develop the desired MDF and create a BRZ on its basis, and also if you teach a computer to extract Knowledge from information and operate on it, then the problem of creating a “Strong AI” is resolved, as it were, by itself.
However, unfortunately, the desired result cannot be achieved even at this stage, even if someone could present the Knowledge in machine format. MFZ and BZ - this is only half the battle. And the whole essence lies in the fact that the BR is needed, in turn, to extract Sense from information, because people in reality exchange Sense , not Knowledge and Data, with the help of which, in fact, people only encode and decode Sense, using language tools. And here a new problem arises - how can you teach a computer to understand the meaning of a text or even a context, like how people can do it? After all, the human consciousness operates, as is known, with Thoughts !
Meaning
Homo sapiens, as is known, does not own telepathy, unlike some animal species, since once humanity lost 9 pairs of chromosomes responsible for telepathy. Therefore, in order to exchange information, people have to encode their Thoughts by various language means. The idea of any author, described by him, for example, with the help of the Russian language, becomes information for all of his addressees who speak the given language, and this idea is later required by the recipients to be decrypted with the help of Knowledge. That's what, ultimately, humanity needs Knowledge!
So, Knowledge is necessary to restore the author's Thoughts from textual information, which, in fact, is nothing more than extracting Sense from natural language messages, that is, without Knowledge, it is impossible to understand what is being said through speech, texts, gestures ... Therefore, the thinking computer in addition to the BZ need to have more and technology for extracting meaning !
Thus, in order to create an artificially-thinking machine, it is necessary to develop such software that could successfully decode the author's Idea, encrypted, for example, in a natural-language text. However, to achieve success, it is necessary to understand what is the procedurality of the Thinking of the human consciousness, from what other Operands, besides the usual implicative, does the “thinking process” consist, etc.? In other words, in order to develop the technology for decrypting author's Thought, you need to know the coding technology itself!
Approaches
Scientists and developers of FIC to obtain the desired offspring on the basis of a computer have already tried many approaches from various sciences and theories, attracting linguistics, statistics, and stochastics and more. In recent years, unsuccessfully having tried everything previously known, the majority of enthusiasts, as mentioned above, in creating computer thinking have directed their main efforts to developing ANNs and trying to create FIS using the new direction based on the so-called “ machine learning ” technology , thanks to which achieved - it should be noted - very significant results in the recognition of audio-visual objects.
However, this approach also usesthe technology of identifying any patterns by statistical data processing and other mathematical tricks , independently, without using the technology of semantic processing of natural language information (texts), is completely incapable of processing Knowledge (at least today and in the near future) and, all the more , is impotent in terms of extracting Sense from an information code, which is a natural-language text.
In addition, at the moment, based on information from open world sources, there is not a single development known to the author even at the level of an idea that could teach a computer to operate Knowledge and extract Meaning from the context, restoring copyright Thoughts.
Output
And where is the light that will lead us out of this impasse? This readers will learn in the sequel to this publication.