What is hidden behind the term modeling

Introduction

I am an analyst, I am engaged in modeling domain areas, and I know that each area has its own individual methods for identifying and describing accounting objects. Thanks to these methods, specific objects appear in each specific area of ​​knowledge. Ideally, all these objects should be in strict accordance with each other, that is, they should not contradict each other.

There are licked subject areas in which there are no problems with the selection of objects. The most harmonious in this sense, mathematics. Modeling mathematical constructions is a pleasure. All because mathematicians paid a lot of effort and attention to creating a consistent axiomatic basis and rules of inference. Thanks to this, mathematics is modeled extremely easily. The same goes for physics (not counting unsolved problems).

In addition to mathematics and physics, there are other areas of knowledge in which methods for identifying objects have been formulated for centuries and have a high price paid in blood — this is medicine (in terms of describing a person’s structure), construction, and martial art. Separately are areas related to commodity-money relations, because we know how people relate to their property. Due to the urgent need to obtain unambiguous and correct interpretations of these areas of knowledge look decent.

But there are subject areas over which there is no strict control, or they are still young enough to reach maturity. The selection of objects in these areas is quite problematic.
')
For example, a process in business analysis can be called both a function and a functional structure, an operation, a typical operation, a scenario, or a typical scenario.

In system engineering, an object and its construction are denoted by one word — the system, the type of functions is called a function, and the description of the system is a set of solutions.

It is all the same, as if in accounting, one term called both the general ledger and the book of accounting for income and expenses. This inevitably leads to collisions, models lean on each other and fall apart, which I often observe in real cases.

In order to cope with the modeling of such problem areas, I had to go into the formalization of the modeling process. This allowed to approach the issue of creating models only formally. Since the method is formal, it allows you to identify contradictions in the subject areas, too, in a formal way. In my articles, I gradually reveal this method, and the one who carefully reads them should eventually learn this method.

Modeling "reality"

Representation

There is existence, and there is our consciousness. Consciousness is the field in which the process of presenting what we perceive is unfolded. In addition to the representations associated with the perceived being, there may arise:

• representations associated with existing hypotheses: for example, ideas about places where we do not exist now, ideas about the past, ideas about the future, ideas about something that we want to create (projects).
• representations associated with an imaginary being: unicorns, branch beavers,
• representations of representations: types, models of representations, and so on.
• other views

Moreover, the ideas connected with the perception of the reality occupy an insignificantly small part of our consciousness in comparison with other ideas. Perhaps this is a feature of human consciousness.

Validation of views

In order not to go away into the imaginary world, we use the methods of validating our ideas with existence.

The processes of representation and validation go synchronously and are called the synchronization of our representation with the existing.

Description of views

To transfer experience, we must be able to describe our ideas. To do this, use the methods for describing views Therefore, the work on modeling things goes in two stages: first, on the basis of things, using the methods of representation, we create the representations in our consciousness, then on the basis of the created representations, using the methods of description, we create their descriptions.

Sometimes there is the task of describing views that are not related to the real, for example, projects. Then we are talking about the description of representations related to hypotheses. This is how the design work looks like - imaginary hypotheses are described.

Recreation of representations

Reverse process: re-creating views based on descriptions. It looks like this: first, based on the description, using the methods of interpretation, we create the representation, and then we try to validate the presentation with the real.

The processes of description and interpretation go in parallel, and generally form a cycle of synchronization of the description and presentation.

Perception cycle

In general, a complete cycle with knowledge sharing looks like this:

Essence - method of representation - representation in consciousness - description method - description - description interpretation method - representation - representation synchronization method - the real.

In this article I will talk about the presentation and description, about the processes that we together call modeling, but I will not touch on the reverse processes - interpretation and validation.

Synchronization of views and descriptions

The presentation and description are related to each other. The way we imagine the world, and the way we describe it, is synchronized. If out of sync occurs, we are talking about either an altered state of consciousness or a loss of reason. Therefore, usually we cannot imagine what we cannot describe, and we cannot describe what we cannot imagine. Stools exist as long as there is a stool presentation method and a description method synchronized with it. Disappear one thing, disappear stools around the world. And vice versa, if a new way of representing reality appears, and a way of describing it consistent with it, entities of a new type will appear. At the heart of our ideas is the idea of ​​the three-dimensionality of the world and time, which “flows” in one direction.

Example 1

Example: Let us face the reality to which the method of presentation can be applied: “leaflet with text”. If we can do this, then an image of a “leaf with text” is created in the mind. To this image, you can apply the method "read aloud." As a result of applying this method, a sequence of sounds is born in consciousness. If the Speech Recognition method can be applied to this sequence, then as a result of applying this method an image is born. Thus, the path from reality to image is rather long. To reduce this path, you can master the speed reading. Those who practiced it know that learning fast reading consists in building images from the text, bypassing the pronunciation of the text. Those who have not learned to read fluently spell, adding one more step to this sequence. For them there are no groups of characters, but there is a set of letters. From them they add sounds, from sounds - words.

Example 2

Let someone say: look, there the horse gallops!
Each word has its own method.
Look - choose the visual perception channel.
Look there - direct your gaze in that direction.
Horse - take the method of interpretation appropriate to this term and try to apply it to what you perceive.
Jumps - take the method of interpretation corresponding to this term and try to apply it to what you perceive.
It may be that by the time when I turned my eyes in the right direction, the method of the horse creates an image, and the method of jumping is no longer. How will I interpret perception?
Then I will go through other methods: sitting, standing, grazing, and so on. When I find the method that is well applicable, for example, it is worth it, I will say that the horse is standing.
It may happen that the subject made a mistake and applied the horse method to a space to which this method is not applicable.
If I start using the horse method, I will not be able to do it. I will say - there is no horse there.
But, if I start to apply the method first, I can interpret the jumping.
Then I will try to solve the problem using other methods, for example, using the kangaroo method. If this method works, then I will say: jumping, but not a horse, but a kangaroo.
The horse method, applicable to one space, creates one image in consciousness, the same method, applicable to another space, returns another image. Therefore, one horse is not the same horse, but similar.
The “ride” method, applied to one space, creates one image in consciousness, the same method, applicable to another space, returns another image. Therefore, one jumping is not the same jumping, but similar.
The same applies to the methods: red (one red image is not another red image) 10 meters (one image at 10 meters is not another image at 10 meters) and so on.

Method and Object

The second thing to remember is that the gallop does not belong to the horse. Often make the mistake of attributing to the horse the ability to gallop. Jumping does not belong to the horse, it is the result of applying the “jumping” method, which is different from the “horse” method. It is possible to say this: a horse, but not to say a horse, or a kangaroo. Other examples: windy, dark, and so on. The fact that a horse gallops means that the same image can be described as a horse and as a steed. But you can say the opposite: the horse is a horse.

If we are talking about the creation of AI, then for him similar methods of representation look like weights in a neural network. Different AIs can share neural networks and coefficients to transfer the presentation methods to each other. Each such method can generate the type of object “in the mind of a computer” unlike anything in our understanding. If the AI ​​needs communication, then the neural networks and weights will “settle”, they will be given the names and the results of the operation of these networks will acquire specific names. This will be the stage of creating your own language in the framework of AI. And not necessarily at the same time that the AI ​​will need only those representations that are placed in our minds, it can do something different.

View classification

Submissions are divided into:

In space:

1. Objects
2. Groups of objects, each of which can be described (constructions)
3. Groups of objects whose individual description is impossible (heaps)

In time:

1. Operations
2. Groups of operations, each of which can be described (processes)
3. Groups of operations whose individual description is impossible (functions)

Description of views

Now consider the task of describing the presentation. Methods are also used to describe representations.

Each description method is a description of the view from a separate point of view. The fuller the description, the more points of view it takes into account. Rich descriptions take into account many different points of view. The poor are few.

The result of the application of the description method to the presentation is as follows: this representation is of this type.

For example, the description method “horse”, applicable to a representation, returns the word “horse”, which means that the representation is of the horse type.

The description method “red” returns the word “red”, which means that the representation is of the red type.

The method of describing "10 meters" returns the word "10 meters", which means that the image is of a type of 10 meters.

Object and its property

A common error often occurs. Due to the peculiarities of the Russian language, the interpretation in the form of a static object (horse) is endowed with the ability to have properties (red). But this is only true in language. In the model with the same success it could be said that the horse is a property of red. The asymmetry of the description laid down in the language, obmusolena philosophers, and successfully migrated to the modeling languages, which leads to known collisions. To avoid this inconsistency, a so-called attribute-free modeling language was invented.

Combining methods

Sometimes methods are combined into groups and attributes are obtained. The 10 meter method and the 11 meter method are different methods, but from the same class that forms the attribute “size”.
The methods included in such groups can be combined (exactly as we dump heaps of sand into one - as a result of a heap again) and get one method. Therefore, the same analogy works here with roles: we can consider the group of methods “Sizing”, we can consider the type of method “Sizing” and we can combine all the methods into one “Sizing”.

Why is it possible to combine methods? Because of all the methods of this group, only one is applicable to the image (the methods are mutually exclusive). For example, it is impossible to merge the method round with the yellow method, because it is possible to apply these methods simultaneously to one image: the yellow circle.

If mutually exclusive methods give an answer on one image, then in this case we are talking about the multiplicity of points of view. That is, in the past example, if the method of 10 meters and the method of 11 meters give the result on the same image, then we say either: this is a mistake, or - these are different points of view.

Different points of view in the description of the presentation

In this case, it’s right to speak not about different points of view, but about mutually exclusive ones. This is a question that I often hear as if we are modeling an object from one point of view. This is wrong, we model from a variety of points of view, just these points of view are not mutually exclusive. As soon as we find that the points of view are in antagonism, we say that these are mutually exclusive points of view, or, as we used to say, different points of view.

All descriptions from all non-exclusive points of view can be combined and get one big description. This description will be a description in which all points of view are coordinated among themselves, but they will be different points of view!

One more important point about which it is necessary to mention. I said for a reason that the description of the view is being made, not of the objects of the real world. Because everyone has his own ideas. For example, the image of a horse in the consciousness of one subject is not the same image as in the consciousness of another.

Description Method Groups

Description methods are used in groups. For example, a horse gallops - two methods in a group. Each method allows you to more or less specialized to describe the presentation. The more specialized the methods, the more they are required. Therefore, the methods should be sufficiently specialized for practical needs, but no more.

Methods can be "orthogonal" to each other, or "parallel". For example, the horse method and the method gallops are orthogonal to each other when the horse can stand with the same probability. If the horse always rides, and only the horse gallops, then the horse method is parallel with the method of galloping. The use of parallel methods in one model leads to redundancy, and we say tautology. Ideally, it is good when the methods are “orthogonal” to each other, that is, they are not related to each other.

For each branch of knowledge has its own group of methods. For each of them have their own words. The more words in the language of the domain, the more methods in this field.
It is clear that the description methods for the operating company differ from the description methods adopted by the construction company. Thus, the task of mapping the results of applying various methods in various spheres of human activity arises. That's what I do - I take methods from different areas of human knowledge and map them to each other.

There is a method in which an electron is a particle, there is a method in which an electron is a wave, there is a method in which an electron is a matrix. They often forget about the matrix, but about the particle and the wave they often write this way: electron-particle and electron-wave. Probably because it is difficult for a person who is far from physics to imagine an electron - a matrix.

View operations

I said that red as a representation is not one as red as representation two. This is a different red. But these different reds can be combined and get one red. It's like two heaps of sand to fall into one and get one heap. A pile of sand can be divided into two piles. So red can be divided into two red. We can combine views and share them.

Is it possible to divide the car in the same way? In space - no, because dividing a machine in space will not give us two cars, but we can divide the car in time. And then we will have two cars, each of which belongs to different members of the same family. Such a division is counterintuitive, but not barred logically. It is like a mixture of two gases - they interpenetrate each other, mix, form a denser substance than each of them separately. In the same way, two cars belonging to different family members penetrate into each other, mix and make matter more dense.

Is it possible to divide the operation in the same way? Let there be an operation to drive a nail. This operation can be divided in time into two with the same hammer and nail divided into parts. I could not imagine how to divide the operation into two sub-operations in space in the same way as to divide the car. If anyone can, please write!
Is it possible to divide the function in the same way? Take the function "show time". This function as well as the machine can be divided into parts in time. But sometimes it can be divided in space. For example, if the function has two dials. Then they can be divided into two functions with one dial. You can combine several functions “show time” into one function, and get one function with multiple dials.

On the objects and groups, you can enter a variety of measures, but this topic is still beyond the scope of the current story.

Results

Modeling lies in the fact that we know the various methods of presenting and describing things that we use here and there. Where the method is suitable, clarification is obtained, and reality acquires forms and features in the subject’s consciousness. Where this cannot be done, reality remains blurred. The effectiveness of modeling is related to the effectiveness of these methods. They must be complete, detailed in moderation and orthogonal.