The Book of Genesis?). Reflections on the nature of mind. Part I

• What is the mind, consciousness.
• How is knowledge different from awareness?
• Consciousness, self-consciousness - the same thing?
• Thought - what is thought?
• Creativity, imagination - something mysterious, inherent in man, or ...
• How the mind works.
• Motivation, goal setting - why do something at all.

Artificial Intelligence is the Holy Grail of any person who has connected his life with IT. The crown of development of any automation, programming, designing mechanisms is the pinnacle of everything. However, the question “What is consciousness, intellect?” Is still open. I do not understand how a huge number of people can deal with a subject for which there is no definition, but I really did not find a concept that would satisfy me. And I had to invent her myself.

Disclaimer: This opus does not pretend to revolutionize the AI ​​paradigm, or revelation from above, it is simply the result of thinking on this topic and to some extent introspection. Also, I do not have any serious practical results, so the text is more philosophical than technical.

UPD:While preparing the article, I stumbled upon several very close concepts ( for example , and even on a habr ). On the one hand, it is a little disappointing that I “rediscovered the bicycle” again. On the other hand, it’s not so scary to bring your thoughts to the public when they are not only mine!

Basic theory

I will not beat around the bush, and give long lyrical digressions of the form "how I came to this" (although it might be worth it). I will begin immediately with the main thing: the wording.

Here she is:

Mind is the ability of a creature to build a complete, adequate and consistent model of reality.

Of course, in its pure form, such a definition gives more questions than answers: how to build, where, what, in fact, means “complete” and “non-contradictory”? And the very " reality given to us in sensation " (c) Lenin is the subject of many philosophical disputes. However, a start has been made - we have a definition of reason. We will expand, complement and expand the concept.

It was not for nothing that I gave the famous quote about reality: in order to build a model for anything, it is necessary to “feel” something. One must be a being , i.e. existand to have ways of perception, data entry channels, sensors - that’s all. Those. our hypothetical AI exists in a certain world and interacts with this world. The main idea of ​​this paragraph - it is foolish to expect a meaningful conversation about football with AI, if all that it interacts with is an indexed knowledge base, like Wikipedia! However, this idea is not new: even the first experiments with a deterministic and understandable world were very impressive . And this is 50 years ago, by the way!

Let's start with the model. Which is complete, adequate and consistent. The definition from Wikipedia at this stage is quite suitable for us: Model - is a system, the study of which serves as a means to obtain information about another system. Its principal structure is not so important, although I have some thoughts on this subject. It is important that, on the basis of the available input data (of that very “feeling of reality”), the mind forms a kind of abstract idea of ​​“how it all is in reality”.

The completeness of this model is critical . It is important to understand that this is all : any knowledge is in a certain way inscribed in the global universal model of reality, or uncognizable! .. Rather, we can say that this is the subtle difference between just knowledge (information) and awareness (unique placement inside the model). You can learn the text in Chinese., you can according to the patterns given to you, find the corresponding piece ... Why, there you can, if you wish, you can teach more and not to such tricks - the Chinese will be in shock! But all this is not related to the intellectual activity of the first type.

Completeness does not necessarily imply maximum detail. The mistake of people trying to go in this direction (creating a comprehensive knowledge base, at the cost of incredible resource expenditures) in an attempt to describe everything at once. The simplest model of all: <All> . In itself, one word implies an indivisible single description of the world. The next possible level of description of reality: ( <something> , <! Something> ) = <All> . Those. there is something and everything else besides this. And together they are all.

A newborn baby initially sees almost nothing. Light and shadow. Gradually, he begins to distinguish certain dark spots on a light background and <something> appears . Almost immediately with the advent of this first element of the model, three more appear: <space> , <time> and the idea of <movement> - changing the position (size?) In space over time. The idea of ​​the length of <existence> is realized soon enough - there was nothing, then something appeared, was and eventually disappeared ( <birth> and <death>?) We still have a very simple model, but there is already a lot of things in it: being and non-being, beginning and end, movement, and so on ... And, most importantly: it still includes all of the perceptions available to the mind. This is a complete description of the surrounding world.

By the way, the question is: how fully can you describe the world around you, having these concepts here (objects, space, time, movement, beginning and end) and only them? ;)

With the advent of concepts of color and form - the number of objects of the model grows. Other senses give a field for the formation of associative links. A built-in unconditioned reflexes form an evaluation function: some prerequisites form a model that in the future has a reality that is evaluated positively (tasty, warm, pleasant), others scare (the last time was bad). Again, unconditional mechanisms make us respond positively to the “good” reality (smile, rejoice) and negatively to the bad (yell!).

And then there is a feedback. Or, perhaps, it appears earlier, when unconditioned reflexes work according to the program “tracking an object” and allow us not to let an object out of sight for as long as possible ... in her!

An important factor in refining the model is the ability to make hypotheses and the ability to test them. The basis of testing is the active perception of the world. In contrast to simple perception (contemplation), the verification of some assumptions requires the targeted acquisition of information. This is the process of learning . You ask the world a question - he answers ... One way or another.

It is important to understand that everything that the mind does is build a model. Consistent within themselves and adequate to reality.

Adequate- means corresponding to reality. If the input data does not fit into the model, then the model requires revision. But sometimes it requires too serious processing and temporarily some parts of the model may conflict with others, i.e. cause controversy. However, in most cases, after such inconsistencies, a new round of thinking will be provoked - this is the mechanism for resolving contradictions . Those. the desire for completeness, adequacy and consistency of the model - these are the basic functions on which the mind is built.

Changing the model, clarifying it - the essence of mental activity. Detailing the model if necessary, and vice versa - a generalization, if possible. Example: an apple and a ball of approximately the same shape / color and until a certain moment are recognized as one concept. However, an apple can be eaten, and the ball is not edible - it means these are different objects and you need to enter a parameter into the model that allows them to be distinguished when classifying (tactile differences, nuances of form, perhaps smell). On the other hand, an apple and a banana have very different external attributes, but obviously there must be ways to find a generalizing factor for them, since A number of common processes apply to them (eating).

If you have a thought , it does not matter - caused by association, external influence, internal trigger to eliminate contradictions, then this is:

• or an attempt to classify and place new information in the model,
• or the actual modeling of some part of the universal model (if from the past - the memory , if from the future, then the forecast or planning , it is possible to find the desired relationship, as an answer to the question ),
• or search and elimination of contradictions ( detailing / crushing , generalization , rebuilding , etc.).

I think in most cases this is all more or less one process, which is thinking .

But you can not only change the model. The mind is part of the world and is the active principle in the world. This means that it can initiate / participate processes that will bring the world into conformity with the model. Those. First there is a model of the world, where conventionally “everything is fine” and in this model, the mind takes certain steps to achieve the desired state of the system. Acting according to the model and having a sufficiently adequate model, the mind will receive a match. This action and motivation to action.

If we are talking about completemodels of the world - it must include the modeling itself. Awareness of one's own capabilities to know and change the world, plus the assessment of different versions of the model as positive or negative - motivation and motivation to action.

Inclusion of oneself in the final model is self-awareness, otherwise, self-awareness.

The model is not static . It necessarily exists in time, with a clear moment "now" and as a result - past and future. Causality, perception of processes rather than objects, is also an important criterion for the "completeness" of the model. On the subject of process perception, you need to do a separate article if the community is interested. ;) I must say at once that if this text seemed raw and heavy - everything is still worse there!

Thinking out loud

Reflections on the topic that came to mind later, or those that I could not fit into the main text ... Like the scene after the captions! ))

• Inclusion in the model gives recursion. However, we are IT specialists, we know what a link is! Yes, the fact that somewhere in the model of the universe is the model of the universe itself - it gives rise to the feeling of the OGVM, and its own exclusiveness! It’s true that each of us has a whole world.
• In fact, the implementation of all this in practice will be a very nontrivial task! “Model” is a too generalized concept, and this model should have a large number of properties that make it difficult to implement, if possible at all (sometimes it seems to me that everything I stated here is trivial, it all passed in the 80s and came to the conclusion that it is impossible to do so). For example, the model should have a rather large flexibility, multilevel, invariance, often with the properties of quantum physics (this is “being in several states at the same time”).
• It's funny that cognitive distortion occurs among people, when instead of taking specific steps that can be taken to bring the world and model into harmony, people simply plan circumstances that do not depend on them - what they will do best ... They say that they are dreamers and castles in the air are building ... Interestingly, within the framework of the theory, isn't it?
• Also, the models of the world in humans can often quite diverge from reality.
• Such purely human qualities (most often considered to be inaccessible to the machine) as creativity and imagination are easily explained within the framework of this topic: everything is clear with imagination - these are model runs in various possible variants, and with creativity it is more interesting! I believe that the process of creativity is an attempt to fix a part of your model in some material physical form, with the aim of either transmitting it to another conscious being or to get the opportunity to more fully embrace the simulated one (nevertheless, the brain resource is finite in this regard).
• Offtop, but in the continuation of the topic: magicians and visionaries. Tarot cards, runes and other fortune telling on the coffee grounds. I believe the pioneers in this business used these systems to visualize / physically implement the models that were in their head. So it’s easier to work with them. And the location of these in space was far from accidental. Just ignorant people did not understand the essence of the process and thought that by means of these magical objects, fortune tellers communicate with spirits. And over time, the fortune tellers themselves also crushed and lost their initial analysis skills.
• In general, I believe that due to the presence of mechanisms of generalization and classification, as well as the search for patterns, consciousness must strive to rationalize the world. Those. something that has an internal structure should be perceived more positively than a chaotic and poorly predictable - not fitting into the model. I fully admit that the feeling of beauty, harmony - the sense of beauty - is a consequence of this striving (when it comes to a work of art). Moreover, the order can be quite complicated - not necessarily a cube, but it can very well be a fractal. And the higher the level of intelligence, the more complex categories of structuredness can be learned.
• Someone would argue that, they say, but what about the beauty of “wild nature”, people, animals, and the like ... Well, here it’s rather relevance / conformity / authenticity - that’s all. The perception of other people in general can be based on nested instincts.
• And yet - the author also puts some promise in his work. Those. This is part of his model. It is obvious that those who directly perceive his work may have different options: from “not gone”, when it is impossible to embed the author’s model in his model, to catharsis, insight and other states - when it’s not just “gone” and “coincided”, but also "put everything in its place" ...
• By the way, this article is also a creative work ... Has it gone? ;)