How intelligence works (a single algorithm for distinguishing and generalizing)

Do you know how exactly you know something? No one knows!

I want to reveal to you a few secrets of natural intelligence, and at the same time tell you about how I design artificial intelligence.

A small disclaimer. The article will describe very ambitious ideas. Most of the ideas presented can be expanded into independent article cycles. Therefore, ideas are presented here only for initial familiarization. I have no illusions that there will be many who grasp the idea of ​​the rally. Therefore, questions are welcome, I will try to clarify. And yes, I know that this is all very similar to a million other ideas, algorithms, etc. The only difference is that this design of ideas claims to be a completely complete simulation of the work of natural intelligence in all aspects that you can or cannot imagine . No black holes, unreasonable problems or unknown technical solutions.

Yes, it’s all about the most complete and strongest artificial intelligence. I foresee the greed of some researchers and the neglect of some collectors of ideas. But nevertheless, now, after a warning, we proceed. There will be a lot of information and its density is very high, so - hold on to something stronger. You may have to re-read dozens of times and ask thousands of questions. I am ready to answer them, since it is time to take solitary research to a more practical level, requiring the involvement of several hundred specialists.

To get started, let's see how the mind works when something finds out. It would seem that we orient ourselves so quickly in the situation, a few moments are enough to find out what objects are visible to us, how they are located, what kind of behavior they have. This illusion gives the impression that the brain works very, very fast.

Why is this an illusion? I’ll try to explain.

Imagine that you suddenly found yourself in a very unfamiliar situation. Where never been. For example, in the jungle, or vice versa, in the desert. Or maybe it will be just a thick ragged fog? Or you see someone else's starry sky.

What's happening? What distinguishes the mind? He has nothing to rely on. We have to classify the situation slowly and painfully. Search for reference clusters of features in order to orientate relative to them.

Do you know what a newborn sees? The chaotic movement of colored spots without content and meaning. But he hears something that allows him to begin to navigate. He hears the mother’s familiar voice. He was already accustomed to touch her body, warm, fragrant, giving delicious milk.

The process of understanding begins with the formation of the senses.

When the mind finds itself in an unfamiliar situation, it does not understand it. He has nothing to rely on. There is not something that he can already predict.
And he begins a consistent and deep classification of the situation

In order to somehow move from images to algorithms and mathematics, we need only four terms and the relationship between them.

The first term is a sign. A sign is something that can be perceived. For example, black color. Or salty taste. Or round shape.
The second sign is presentation. The view is based on the recurrence of features. If two signs are perceived together, a view arises generalizing these signs.
The third term is subview. Subview is the relation of a sign to a representation. For all representations, its attributes are sub-representations.
The fourth term is a supermark. Supercharacter is the attitude of the submission to the supercharacter. For all the signs of the representation they enter into, these are super signs.

All together it is combined into a proactive universal classifier

What does proactive mean? This means that he classifies all the time. This is extremely important. All the behavior of the classifier is determined only by where it wants to look, what to check.

What does it mean - universal? This means that he classifies everything that he can reach.

What does the classifier mean? This means that he builds a connected network of ideas about situations, about trends in their changes, about the most stable representations of the situation, thereby forming the most stable idea about the outside world.

And how does all this work?

The classifier algorithm is not complicated, but rather difficult to understand. First, I will describe the requirements for the algorithm, and then I will give a couple of illustrative examples.

1. The sensory organs are drivers
1a specific to each sphere of perception . No matter how complicated the sphere of perception, it unfolds into a one-dimensional set, in which the address (index) and the value of the attribute according to this index
1b correspond to the attribute . The classifier, referring to the driver at a specific address, receives the measured value as a result. This can be a symbol in sequential text, the color of an area in space, the height and amplitude of sound, temperature, etc.
1c. Importantly, the driver uses a relative addressing model, from index to index, without preserving the time sequence. For example, in the line “MOTHER SOAP FRAME”, the index indicates an offset. If the current index points to the first letter of the second word, then the offset +2 points to the letter “L”. If you again set the offset +2, we get a space.

2. The classifier saves transitions, counting the probability of confirmation and refutation of values ​​during transitions from a key attribute by index offset.
2a. If the current value of the sign is already on the network, then the classifier selects the most probable representation for this sign and predicts the most probable transition to the next sign, after which it turns to the driver with an offset.
2b. The entire network of representations is constantly adjusted and sorted according to the most probable representations of the current features
2c. Second-order representations interconnect first-order representations. Thus, in order to check the second-order representation, it is necessary to check the first-order representations included in it. It turns out a recursive algorithm.

3. The instant state of the classifier is represented by a hierarchy of ordered lists of representations of different orders.
3a. The algorithm always tries to increase the order of presentation in the first place. The highest order of presentation reflects the highest level of understanding. So, for example, for the lid and legs, the next level of presentation is the table, and for the table, chair and sink, the next level is the kitchen.
3b. The classifier contains simple numbers as representations. This is the internal language of the classifier. An internal language in a subset corresponds to the language that people use for communication (expression of representations and their recognition). The algorithm for comparing representations of expressions and perceptions is exactly the same, but a hack is possible. You can choose a certain order of representations and increase the priority of those that make sense to a person. Then the classifier will quickly explore precisely those ideas about the world that will help it quickly and better navigate the world perceived by people and communicate with us in one language

4. The classifier does not work with infinite memory and in infinite time. Therefore, he ignores quite a few signs and perceptions as unpromising for a stable classification. Later we will examine how multidimensional spheres of perception are packed into a network of representations, including movement and measure of movement - time.
4a. Preliminarily, a million transitions between subrepresentations of the previous order are allocated for each order of representations, while only twenty thousand of the most probable are taken into account for the formation of the next order. It is important here that both the most probable confirmations and the most probable refutations are useful, therefore we consider them separately
4b. The number of orders of representations is quite large, and is almost equal to the number of representations of the first order. In order to reduce the number of orders (after all, for each order we assign about a million transitions between subviews) from a million to the same 20 thousand, the idea of ​​the highest context is applied. The representations confirming the highest order representation after ranking in probability are thinned out, preserving only the shortest chains to the primary signs.

5. The direction of classification is always determined from the presentation of the highest order. The context contains representations of the highest order ordered by probability, constantly updating this context, checking the occurrence and termination of signs that confirm or deny this context most quickly.
5a. The content of this context is the most complete understanding of the studied situation and trends
5b. The less in the context of understanding, the higher the assessment of clarity and accuracy of understanding
5c. It is in the context of the context that the issues of communication, distinguishing between subjects and objects, their groups and interaction, goals and desires, relationships and emotions are determined. This issue can be considered for a long time, for a start there should be a good idea of ​​the main cycle of the classifier and the main cut of the classification network state.

Why all these details?

Of course, there is a simpler representation of the basic idea of ​​the classifier. This idea sounds like this: “what does it belong to?” This can be difficult to understand. When we speak of a cognitive act, we usually mean the question "what is this?"
In practice, this question has two aspects: “what does it own?” and "what does it belong to?" These two questions are invariant, it is enough to simply expand the direction of the search. These are parallel processes of perception / behavior. Perception moves towards the search for the owner (the most stable support for further classification), and behavior moves towards the search for the owner.

What next?

For those who are interested, we can join forces in the study of the model and the development of useful and relevant applications that do not require significant resources, but work with the most important AI industry - with understanding.
Not much where you will find something specific on the subject of understanding. What is understanding? How to define it? How to model it?

Therefore, I propose a very deeply elaborated model, covering both metaphysical and philosophical issues, as well as pragmatic technical issues of product implementation and implementation.

Welcome to the new world - the world of real artificial intelligence and artificial personalities.

Your move, harazhiteli.