Path to Intellect. Step one. Overall picture
The article makes an attempt to understand what Intellect is and explain in simple terms the principles of its construction and implementation. The concept of "Intelligence" generates fog even in the minds of smart people. As a rule, everyone understands Intelligence so that he "thinks like a person." Bearing in mind the formal side of thinking. But a person can think not only formally (logically), but also emotionally, as well as abstractly (not to be confused with logical), but also sensory sensations (specifically).
The topic of Intelligence is interesting for its inconsistency and confusion. Here a large number of different concepts are mixed in one pile, somehow related to each other. What is intelligence? And how does it relate to Reason and Consciousness? Is it the same or not? What is the subconscious? What is meant by thinking? Does the cell have consciousness? Do monkeys have a mind? How is frog intelligence different from bee intelligence? Or do they have no intelligence? It is clear that a million such questions may arise. Therefore, I consider it extremely important to define a framework - what is Intelligence - and therefore determine where it is and where it is not.
Intelligence(from lat. intellectus - cognition, understanding, reason), the ability of thinking, rational knowledge ... etc. On the Web you can find many similar definitions, from where it becomes clear that they give us little. For example, from everyday ideas it can be considered that dogs possess intelligence. But whether they possess rational knowledge is more difficult to say. And what is rational knowledge? Looking at the moon through a telescope? Then they don’t possess ... well, etc. Based on the practice on the thematic forums, I have come across many different opinions, sometimes extremely opposite.
For example, there were allegations that there is an “animal” and a “civilizational” intelligence. Or “figuratively-sensual” and “logical-abstract” types of intelligence. It turns out, for example, that figurative thinking cannot be abstract. But what about Salvador Dali? Some also say that only a person possesses Intelligence, and in all animals only "adaptive behavior". Many people prefer to equate: Intelligence = Consciousness = Reason. Some people think that Consciousness begins with a cell. Or Intellect begins with a cell, but they put their understanding into this term. There is also a theory [Simonov] that there is the Subconscious, Conscious and Superconscious. The latter is such a creative Consciousness that only units can have. Or, for example, there is an opinion that mechanisms (including electronic ones) have Intellect, but "very, very little," the dog has "more," and the person has "very much." It turns out that Intellect is just some elusive substance that exists everywhere and nowhere ...
Before moving on, it is necessary to briefly dwell on the concept of Model, which is crucial in the science of Intelligence.
A model (lat. Modulus - “measure, analog, sample”) is a system, the study of which serves as a means to obtain information about another system , this is a simplified representation of a real device and / or processes occurring in it, phenomena ... There are other similar definitions, however, there is a need to clarify them. In general, a Model can be understood as anything, even a paper boat. It can be the same theory and formula, it’s just that they will be different types of models: physical, mathematical, formal, etc.
So, in our reasoning, there is a place to be a Model and “another system” - the Original. The main statements will be as follows:
1) The model is much simpler than the Original;
2) The model contains the “essential” properties of the Original;
3) their “materiality” is determined by the purpose of building the Model.
Without the Purpose it is impossible to distinguish the “essential” properties of the Original, which means that without the Purpose there is no Model. In other words, the Model is a look at the Original from the point of view of the Purpose. The trick is that due to the complexity of real objects, there can be many such points of view and all of them will be correct. Thus, a triple of interconnected concepts arises: Original, Model and Purpose. Remember this.
A simple example shows that a frog that lives in a pond and a heron that preys on it have different goals and different resources to achieve their goals. It is also natural to assume that they will have different models of the same pond. Does this mean that these models will not be correct? No, it does not, because the goals will also be different. And then what is the criterion for the adequacy of these models? The answer is obvious - the degree to which the goal is achieved: the more efficiently the goal is achieved, the more adequate the model.
Now back to the definitions of Intelligence, of which many are interesting in the following. Intelligence - the ability of a system to display patterns of the outside worldand use them to achieve your goals [Samarin]. In my opinion, here the phrase “reflect the laws of the outside world” needs to be clarified and replaced with “dynamically form sensory models of the environment”.
Now we turn to Fig. 1, which shows a graph of the levels of the effectiveness of the behavior of living organisms, depending on the presence of certain thinking mechanisms. The graph has an abrupt threshold nature, like most natural processes. Obviously, a jump in efficiency appears after the emergence of a new mechanism of thinking.
1. The level of automaton behavior. It is characterized by “hard-wired” (static) behavior, which changes little or does not change at all when external conditions change. It is characterized by rigidity and low flexibility. The ability to modify (learn) is weak and only in a narrow range of possibilities. This level of behavior is characteristic of insects.
2. The level of adaptive behavior arises as a result of the formation of sensory models of the environment. These models are formed dynamically, i.e. in the process of life. The mechanism for the formation of such models is training. Adaptive behavior will also be called the behavior of "simple" intelligence. This behavior is characteristic of amphibians and reptiles.
3. The level of subconscious behavior occurs after the appearance of a mechanism that can dynamically form sensorypredictive models of the environment. In accordance with the Theory of Functional Systems (TFS) [Anokhin], the prognostic properties of intelligence are provided by the Acceptor of the result of an action (ARD). These will be systems with “true” Intelligence, which we will also call systems with “complex” intelligence. An example would be higher mammals. For the reasons stated above, it can be assumed that cells and insects do not have any “intelligence”. It never was and never will be.
4. The level of conscious behavior arises as a result of the appearance of iconic models of sensory modelsWednesday. Those. a label — a label, or a label, or a sign is “glued” to the sensor model, after which it is possible to operate with a reference to the model, and not the sensor model itself. Sign systems can be any: graphic, sound, receptive (Braille), sign language, etc. Then we have the following chain: The
Real World -> Sensory Models -> Sign Models
The model of the model leads to even more simplification of complex objects of the real world, which makes it possible to learn more effectively. Pavlov’s well-known idea that a word is a signal of a signal, as a result of which a person has a second signal system, confirms this logic.
Outwardly, it looks like a speech, in one form or another. It also leads to the appearance of rational behavior and the phenomenon of Consciousness, when one's own thoughts and actions are realized. They can be voiced, explained to another person, etc. Experiments with higher primates show that they have protoconsciousness, because successfully master a simple language. We can say that they are at the foot of the last "step" on the chart. For this reason, cats, amphibians and cages have no consciousness and cannot be. The aforesaid can be briefly expressed as follows:
1. (Automatic behavior) <- Rigid static models of the environment
2. (Adaptation = Simple intelligence) <- Touch dynamic models of the environment
3. (Intelligence = Complex intelligence) <- Touch models + ARD (Forecast)
4. (Reason = Consciousness) <- Sensory models + ARD + Sign models
Hence, difficulties in solving the problem of understanding the natural language and human thinking become clear. This is due to ignoring the stages of sensory modeling and sensory prediction. Researchers immediately grab onto higher functions, which are based on primitives of a lower level, forming, ultimately, a hierarchical system of models. This means that phased work on the implementation of these levels is necessary, of which levels No. 2 and No. 3 are of significant interest. The implementation of level 1 seems trivial. Therefore, it makes sense to consider the known data regarding the "constructive" understanding of the mechanisms of thinking in the field of neurobiology, neuroanatomy and psychophysiology. Here, “thinking mechanisms” refers to architectural concepts of the most general form.
Fig. 2 shows a model of behavior, also known as the architecture of a functional system within the framework of TFS [Anokhin]. Fig. 3 shows a model of a behavioral act taking into account needs, motivation and emotions [Simonov]. Some analysis reveals that both well-known scientists depicted the behavior in fact in the form of time diagrams , and not in the form of static block diagrams, as is customary, for example, in digital circuitry. There are other similar schemes, for example, in Fig. 4 [Redozubov], in Fig. 5 [Zhdanov] and many others, which also do not bring much clarity.
For example, if in Fig. 2 there is a block “Decision Making”, then in Fig. 3 and Fig. 4 it is absent, in Fig. 3 there is a block “Emotion”, and in Fig. 2 it is absent. The same can be said about “Acceptor of the result of an action”, “Motivation”, “Formation of memory”, “Knowledge base”, “Formation of action”, “Pattern recognition”, etc. Thus, it is impossible to understand from known sources what the architecture of a technical system should be, capable of carrying out intellectual functions in accordance, for example, with the level of efficiency of behavior No. 2. From this it is clear that for the formation of such an architecture an engineering approach is necessary taking into account the general picture considered.
To be continued.
The used literature
Aleksandrov Yu.I. Psychophysiology. St. Petersburg: Peter, 2007.
Anokhin P.K. Fundamental questions of the general theory of functional systems // Principles of systemic organization of functions. M .: Nauka, 1973.
Velichkovsky B.M. Cognitive science: the foundations of the psychology of cognition. M .: Sense, 2006.
Danilova N.N., Krylova A.L. Physiology of higher nervous activity. M .: Educational literature, 1997.
Zhdanov A.A. Autonomous artificial intelligence. M .: Binom, 2009.
Luria A.R. Higher cortical functions of man. M .: Moscow State University, 1962.
Redozubov A.D. The logic of emotions. St. Petersburg: Amphora, 2012.
Samarin A.I. Mobile robots and self-organization in neural structures - a retrospective view // From behavior models to artificial intelligence. Ed. V.G. Redko. M .: URSS, 2006.
Simonov P.V. The emotional brain. M .: Nauka, 1981.
Sudakov K.V. Physiology. Basics and functional systems. M .: Medicine, 2000.
Turchin V.F. The phenomenon of science: a cybernetic approach to evolution. M .: ETS, 2000.
Hubel D. Eye, brain, vision. M .: Mir, 1990.
Shvyrkov VB Introduction to objective psychology: the neuronal basis of the psyche. M .: Institute of Psychology RAS, 2006.
Ashby U.R. Brain construction. M .: Foreign literature, 1962.
The topic of Intelligence is interesting for its inconsistency and confusion. Here a large number of different concepts are mixed in one pile, somehow related to each other. What is intelligence? And how does it relate to Reason and Consciousness? Is it the same or not? What is the subconscious? What is meant by thinking? Does the cell have consciousness? Do monkeys have a mind? How is frog intelligence different from bee intelligence? Or do they have no intelligence? It is clear that a million such questions may arise. Therefore, I consider it extremely important to define a framework - what is Intelligence - and therefore determine where it is and where it is not.
Intelligence(from lat. intellectus - cognition, understanding, reason), the ability of thinking, rational knowledge ... etc. On the Web you can find many similar definitions, from where it becomes clear that they give us little. For example, from everyday ideas it can be considered that dogs possess intelligence. But whether they possess rational knowledge is more difficult to say. And what is rational knowledge? Looking at the moon through a telescope? Then they don’t possess ... well, etc. Based on the practice on the thematic forums, I have come across many different opinions, sometimes extremely opposite.
For example, there were allegations that there is an “animal” and a “civilizational” intelligence. Or “figuratively-sensual” and “logical-abstract” types of intelligence. It turns out, for example, that figurative thinking cannot be abstract. But what about Salvador Dali? Some also say that only a person possesses Intelligence, and in all animals only "adaptive behavior". Many people prefer to equate: Intelligence = Consciousness = Reason. Some people think that Consciousness begins with a cell. Or Intellect begins with a cell, but they put their understanding into this term. There is also a theory [Simonov] that there is the Subconscious, Conscious and Superconscious. The latter is such a creative Consciousness that only units can have. Or, for example, there is an opinion that mechanisms (including electronic ones) have Intellect, but "very, very little," the dog has "more," and the person has "very much." It turns out that Intellect is just some elusive substance that exists everywhere and nowhere ...
Before moving on, it is necessary to briefly dwell on the concept of Model, which is crucial in the science of Intelligence.
A model (lat. Modulus - “measure, analog, sample”) is a system, the study of which serves as a means to obtain information about another system , this is a simplified representation of a real device and / or processes occurring in it, phenomena ... There are other similar definitions, however, there is a need to clarify them. In general, a Model can be understood as anything, even a paper boat. It can be the same theory and formula, it’s just that they will be different types of models: physical, mathematical, formal, etc.
So, in our reasoning, there is a place to be a Model and “another system” - the Original. The main statements will be as follows:
1) The model is much simpler than the Original;
2) The model contains the “essential” properties of the Original;
3) their “materiality” is determined by the purpose of building the Model.
Without the Purpose it is impossible to distinguish the “essential” properties of the Original, which means that without the Purpose there is no Model. In other words, the Model is a look at the Original from the point of view of the Purpose. The trick is that due to the complexity of real objects, there can be many such points of view and all of them will be correct. Thus, a triple of interconnected concepts arises: Original, Model and Purpose. Remember this.
A simple example shows that a frog that lives in a pond and a heron that preys on it have different goals and different resources to achieve their goals. It is also natural to assume that they will have different models of the same pond. Does this mean that these models will not be correct? No, it does not, because the goals will also be different. And then what is the criterion for the adequacy of these models? The answer is obvious - the degree to which the goal is achieved: the more efficiently the goal is achieved, the more adequate the model.
Now back to the definitions of Intelligence, of which many are interesting in the following. Intelligence - the ability of a system to display patterns of the outside worldand use them to achieve your goals [Samarin]. In my opinion, here the phrase “reflect the laws of the outside world” needs to be clarified and replaced with “dynamically form sensory models of the environment”.
Now we turn to Fig. 1, which shows a graph of the levels of the effectiveness of the behavior of living organisms, depending on the presence of certain thinking mechanisms. The graph has an abrupt threshold nature, like most natural processes. Obviously, a jump in efficiency appears after the emergence of a new mechanism of thinking.
1. The level of automaton behavior. It is characterized by “hard-wired” (static) behavior, which changes little or does not change at all when external conditions change. It is characterized by rigidity and low flexibility. The ability to modify (learn) is weak and only in a narrow range of possibilities. This level of behavior is characteristic of insects.
2. The level of adaptive behavior arises as a result of the formation of sensory models of the environment. These models are formed dynamically, i.e. in the process of life. The mechanism for the formation of such models is training. Adaptive behavior will also be called the behavior of "simple" intelligence. This behavior is characteristic of amphibians and reptiles.
3. The level of subconscious behavior occurs after the appearance of a mechanism that can dynamically form sensorypredictive models of the environment. In accordance with the Theory of Functional Systems (TFS) [Anokhin], the prognostic properties of intelligence are provided by the Acceptor of the result of an action (ARD). These will be systems with “true” Intelligence, which we will also call systems with “complex” intelligence. An example would be higher mammals. For the reasons stated above, it can be assumed that cells and insects do not have any “intelligence”. It never was and never will be.
4. The level of conscious behavior arises as a result of the appearance of iconic models of sensory modelsWednesday. Those. a label — a label, or a label, or a sign is “glued” to the sensor model, after which it is possible to operate with a reference to the model, and not the sensor model itself. Sign systems can be any: graphic, sound, receptive (Braille), sign language, etc. Then we have the following chain: The
Real World -> Sensory Models -> Sign Models
The model of the model leads to even more simplification of complex objects of the real world, which makes it possible to learn more effectively. Pavlov’s well-known idea that a word is a signal of a signal, as a result of which a person has a second signal system, confirms this logic.
Outwardly, it looks like a speech, in one form or another. It also leads to the appearance of rational behavior and the phenomenon of Consciousness, when one's own thoughts and actions are realized. They can be voiced, explained to another person, etc. Experiments with higher primates show that they have protoconsciousness, because successfully master a simple language. We can say that they are at the foot of the last "step" on the chart. For this reason, cats, amphibians and cages have no consciousness and cannot be. The aforesaid can be briefly expressed as follows:
1. (Automatic behavior) <- Rigid static models of the environment
2. (Adaptation = Simple intelligence) <- Touch dynamic models of the environment
3. (Intelligence = Complex intelligence) <- Touch models + ARD (Forecast)
4. (Reason = Consciousness) <- Sensory models + ARD + Sign models
Hence, difficulties in solving the problem of understanding the natural language and human thinking become clear. This is due to ignoring the stages of sensory modeling and sensory prediction. Researchers immediately grab onto higher functions, which are based on primitives of a lower level, forming, ultimately, a hierarchical system of models. This means that phased work on the implementation of these levels is necessary, of which levels No. 2 and No. 3 are of significant interest. The implementation of level 1 seems trivial. Therefore, it makes sense to consider the known data regarding the "constructive" understanding of the mechanisms of thinking in the field of neurobiology, neuroanatomy and psychophysiology. Here, “thinking mechanisms” refers to architectural concepts of the most general form.
Fig. 2 shows a model of behavior, also known as the architecture of a functional system within the framework of TFS [Anokhin]. Fig. 3 shows a model of a behavioral act taking into account needs, motivation and emotions [Simonov]. Some analysis reveals that both well-known scientists depicted the behavior in fact in the form of time diagrams , and not in the form of static block diagrams, as is customary, for example, in digital circuitry. There are other similar schemes, for example, in Fig. 4 [Redozubov], in Fig. 5 [Zhdanov] and many others, which also do not bring much clarity.
For example, if in Fig. 2 there is a block “Decision Making”, then in Fig. 3 and Fig. 4 it is absent, in Fig. 3 there is a block “Emotion”, and in Fig. 2 it is absent. The same can be said about “Acceptor of the result of an action”, “Motivation”, “Formation of memory”, “Knowledge base”, “Formation of action”, “Pattern recognition”, etc. Thus, it is impossible to understand from known sources what the architecture of a technical system should be, capable of carrying out intellectual functions in accordance, for example, with the level of efficiency of behavior No. 2. From this it is clear that for the formation of such an architecture an engineering approach is necessary taking into account the general picture considered.
To be continued.
The used literature
Aleksandrov Yu.I. Psychophysiology. St. Petersburg: Peter, 2007.
Anokhin P.K. Fundamental questions of the general theory of functional systems // Principles of systemic organization of functions. M .: Nauka, 1973.
Velichkovsky B.M. Cognitive science: the foundations of the psychology of cognition. M .: Sense, 2006.
Danilova N.N., Krylova A.L. Physiology of higher nervous activity. M .: Educational literature, 1997.
Zhdanov A.A. Autonomous artificial intelligence. M .: Binom, 2009.
Luria A.R. Higher cortical functions of man. M .: Moscow State University, 1962.
Redozubov A.D. The logic of emotions. St. Petersburg: Amphora, 2012.
Samarin A.I. Mobile robots and self-organization in neural structures - a retrospective view // From behavior models to artificial intelligence. Ed. V.G. Redko. M .: URSS, 2006.
Simonov P.V. The emotional brain. M .: Nauka, 1981.
Sudakov K.V. Physiology. Basics and functional systems. M .: Medicine, 2000.
Turchin V.F. The phenomenon of science: a cybernetic approach to evolution. M .: ETS, 2000.
Hubel D. Eye, brain, vision. M .: Mir, 1990.
Shvyrkov VB Introduction to objective psychology: the neuronal basis of the psyche. M .: Institute of Psychology RAS, 2006.
Ashby U.R. Brain construction. M .: Foreign literature, 1962.