What is important for the creation of AI 2: “Dependence. View from the neural network ”

Addiction

Every time there were programs on television about addiction to the Internet or computer games, I felt uneasy. These dependencies have always been equated with a narcotic one, emphasizing that, practically, there is no difference between these dependencies. And what could be worse? I played computer games often and a lot, but most of the time I devoted to studying the PC and software itself, spending a lot of time experimenting on hardware and software, which I still enjoy. And, moreover, I had no dependence on the Internet. He was very expensive, he used it only to search for information on an abstract, course, diploma. But often he caught the slanting glances of his parents, when on television they once again talked about these dependencies. I was always surprised how exactly “scientists” (I’m not surprised that, in fact, not scientists, but the devil knows who) searched and measured the total in these dependencies. In this article I will try to prove that each addiction "is practically no different from drug addiction," and that this similarity does not prove the addiction at all. I will show that addiction, habit and hobby "is practically no different from drug addiction." I will also “prove” that AI can be dependent and habits will be inherent in it.
Since gambling, drug addiction and Internet addiction are completely different things, the only thing in common is psychological dependence. Therefore, I will consider the dependence on the part of the “nervous” system, namely, neurons and neural networks (I take the dependencies for the sake of example only, and what kind of dependence does not matter at all). I will not describe what a neural network is - materials on this topic on the Internet simply fell into pieces. The only thing I’ll clarify is that the brain is a self-organizing neural network, so this network will be discussed.

Attention this is not an article where you can read and chew popcorn!
Caution! A lot of hard conclusions to understand! Therefore, do not translate in vain tomatoes and rotten eggs ...
If you are ready, please cut!

First terms and abbreviation:

GNS is a global neural network (the entire neural network).
LNS is a local neural network (part of a GNS), the neurons of which can be either connected together or scattered across a global neural network. It is understood that the neurons of a given network are united by some property, for example, this network is some kind of skill, or memory.
NS - a neural network (it doesn’t matter, GNS or LNS).
Negative signal - an error signal in the NS (weakens synapses).
Positive signal - confirmation signal (strengthens synapses).
Interference - third-party signals that complicate the work or lead to incorrect operation of the system on which they act.

Skill ~ Habit?

To begin with, let's start from the bottom - let's consider what a skill is:
Skill - an activity formed by repeating and bringing to automatism.
And also, what is a habit: A
habit is an established way of behavior, the implementation of which in a certain situation acquires the character of a need for an individual that prompts him to perform some actions or actions.
(Both definitions are taken from Wikipedia) To perform any action you need to be able to do it. Thinking logically, we can conclude that a habit is a skill that works (“turns on”) “in a certain situation”. It turns out that a habit is a skill to select a situation in order to use another skill. It follows that from the side of the neural network, the habit from the skill, in fact, does not differ - it is the same skill, the same “local” network of neurons, organized in the learning process. On the other hand, the entire neural network consists of the interconnection of such “local” networks - some were from the very beginning, others were built in the learning process, I think this was no secret to anyone. What else can a complex neural network consist of if not with smaller networks combined into one big one? Mechanically closing an important window or message,

LNS destruction mechanisms

Any of the LNS can be destroyed, for example, if someone hits you on the fingers when you automatically close the desired window again (at that moment I remembered Demosfen), as a result of numerous negative signals, this local network will begin to weaken the synapses and, in the end finally, will begin to break ties. As a result, of course, you can “build” a network that will “force” you to squint every time you close a window, but it will already be a completely different network (squinting in anticipation of a hit on the fingers, or remembering the sensations of a blow ) It will not be so easy to destroy the network if it has strong enough connections, because of which you will squint and close the windows at the same time.
Let's look at the “natural” destruction of LNS. Suppose you drowned your old mobile phone, which you are used to and with closed eyes know what is in it, where and how, but you had to buy a new one. A lot of habits and skills acquired from using an old phone are not suitable for him. From time to time, your fingers do something, repeating what they did on the old phone, but on this phone these actions do not give the desired results, as a result there are a lot of mistakes and stupid actions. Your LNS often receive negative signals and begin to collapse, break ties, create new ones. New connections can be completely wrong, leading to illogical actions. LNS data from ordered networks turns into chaos, creating interference in the STS. The skill is gained - a new LNS is built, but, for some time, there is an old one, which, from time to time, forces us to do old actions not applicable to the new phone. In other words, it creates “interference” in the STS, bringing us discomfort, wrong actions and a lot of trouble. The old LNS will look for its place in the STS for some time, destroying and creating new connections both inside and outside. It will either be reorganized into new actions, which is unlikely (except that the old and new actions will be very similar), or it will crumble into chaos and neurons will be “forced” to look for their place in the STS in other LNSs, inevitably creating interference in their work. destroying and creating new connections both inside and outside. It will either be reorganized into new actions, which is unlikely (except that the old and new actions will be very similar), or it will crumble into chaos and neurons will be “forced” to look for their place in the STS in other LNSs, inevitably creating interference in their work. destroying and creating new connections both inside and outside. It will either be reorganized into new actions, which is unlikely (except that the old and new actions will be very similar), or it will crumble into chaos and neurons will be “forced” to look for their place in the STS in other LNSs, inevitably creating interference in their work.
In any case, the destruction of the LNS, there is a high probability of death of neurons. There is no guarantee that each neuron will find a new place in the GNS, especially if it consists of self-sufficient LNSs that no longer need to be trained. Any neuron added to it will only create interference and will rather interfere than help. Such a network will either collapse (in extreme cases, it will be damaged) if it is weak or simply does not “let in” a new neuron. A free neuron not in contact with stimuli will simply die. It is difficult to imagine what kind of chaos is created during the destruction of the LNS; one can still take into account that each such LNS is associated with other LNS and its destruction leads to the fact that other LNS will be “forced” to look for new connections. The destruction of the LNS can be equated to a local explosion in the GNS.Why can we conclude that part of the neurons of the destroyed LNS will be attached to the free conclusions of other LNS. Free conclusions will create and destroy synapses for a long time, until there are new correct synapses. You can also mention that in the learning process, in any case, some neurons can be released, but this can not be compared with destruction.

Some lyrics

A little digression before continuing to reflect on the topic. Imagining the destruction of the LNS, it dawned on me: "so this is why every part of the brain has its own place." This, of course, does not explain why it is exactly where it is located, but it explains why. One has only to imagine what happens if the destruction of some LNS affects the LNS responsible for breathing, heartbeat, or any equally important LNS. Death is what happens (the released neurons from the destroyed LNS will look for their place in other networks, creating interference and, as a result, the malfunctioning of the networks into which they will be introduced). Another thing that I thought about is that the destruction of the network occurs only with a sufficient number of negative signals. How does a person understand that he is doing something wrong? This happens in different ways. In one case, it is pain in others, it is emotions (of him or someone from those around him), inferences of the person himself - a result is obtained, but, for example, the opposite of what was expected, because of which the person received stress. Stress itself, in this case, is a negative signal. Stress can be so strong that this person may be speechless - the LNS responsible for speech has collapsed (albeit not completely, only part of the synapses). And here is speechless before his actions? Stress is a negative signal, it spread along the GNS (albeit not evenly), “falling under the distribution” synapses burst. This is a normal reaction to a negative signal in the NS. As already mentioned, the destruction of the neural network is unlikely to do without the death of neurons. Recall the expression: “Do not be nervous! Nerve cells are not restored. ” Here I also notice that in a natural NS, the mechanism of propagation of a negative signal (error) is not completely transparent, as in an artificial one. Firstly, it does not act so locally and clearly; secondly, in addition to signals, there are still a lot of processes in the NS catalysts / inhibitors, etc. etc. I’m not going to get into the Vety wilds - this is another topic and not one ... Okay, enough lyrics, let's get back to the topic ...

Some more terms

Having considered two simple concepts (skill and habit), you can take the slippery path to begin to consider the third concept - dependence.
Dependence (addiction) is an obsessive need that moves a person to a certain activity. (from Wikipedia)
Here I returned to the definition of habit (and highlighted the word need). We can say that addiction is the same habit and we will be right. But this concept is a little more complicated than it would seem. Imagine the situation that you are thirsty. The body generates some kind of irritating signals, which as a result should lead to the satisfaction of the need to drink. You can drink a glass of water without even thinking, and not even remember that you did it.
And now let's delve deeper into the question and consider several terms (from Wikipedia):
Addiction (habituation) is a gradual decrease in the response as a result of ongoing or repeated stimulation under normal conditions (not accompanied by reinforcement of the stimulus).
A receptor is a complex formation consisting of the terminals (nerve endings) of dendrites of sensitive neurons, glia, specialized formations of intercellular substance and specialized cells of other tissues, which in a complex provide the transformation of the influence of factors of the external or internal environment (stimulus) into a nerve impulse.

Addictive

Receptors are able to perceive not only the stimulus, but also its intensity. Let's get used to the addiction experiment. We put a coin in our hands on a brick and sit for a while without moving it. Over time, we will stop feeling the coin. If we put a brick on our hand, then we would hardly be accustomed to it, the mass of nerve endings would "yell" about damage, i.e. changes would occur at the micro level, over time this would be visible to the naked eye: redness of the hand, etc.The reason for this is not receptor fatigue at all. Imagine a neural network that received the same signals. What happened to the network? She changed until she learned from these signals. And suddenly the signals changed. What will happen? Yes, the same as in the previous case - the network will be rebuilt until it learns. Presumably, our nervous system is able to "feel" only changes in stimuli. A hand pressed by a brick can no longer feel or move, until the blood circulation resumes.As soon as we move our hands (change the stimulus), we will again begin to feel the coin. We may not even feel the pain of a wound until we move it or touch it. Now let's take, for example, listening to music while listening to the same song for a long time, we will no longer receive the pleasure that we received when we first listened. Listening to music, we “train” the LNS, receiving a lot of positive signals. At some point, the network learns and stops emitting positive signals. If you continue to listen to music that is already tired, then the training will not stop, it will continue only in other neural networks and if the network starts to receive completely “unacceptable" signals it will signal very negative signals, it may even begin to break down, in other words, you can go crazy. Imagine a network which is already sufficiently trained and then she is “told” that her every action is erroneous. That is why when we are sad we want to listen to sad music. Because music stimulates those neural networks in which the search for the right solutions takes place. Free conclusions and neurons create “interference” in the NS, when listening to music, these neurons and conclusions connect and receive positive signals - confirmation of the correct actions. Hence the interest in various things, the desire to learn, to receive new sensations. There is never equilibrium in a system — there are always free conclusions that interfere with the NS — forcing us to act, in search of a new one, to create and consolidate new synapses. On the other hand, if signals contribute to breaks, first of all, at first, to negative signals, then such signals are a big discomfort for NS, such signals would be simply unacceptable. In other words, the more signals signaling that the synapse is superfluous and the stronger the synapse, the more discomfort for the NS, the synapse was 100% sure that it works correctly, and they tell him that everything that he does wrong is fundamentally a disaster .
All this is nothing but addiction. An addiction through learning but beyond nature is not the only type. Consider another example. We are sitting in a car that rides on paving stones. If we sit in it for a while , we just grit our teethwe will cease to feel the same irritation as at the beginning of the trip (this is addiction through training). And when we stop the car, for some time we will feel that we are shaking, as if we're still driving. What does it mean? This means that some processes are taking place in the National Assembly, which make us feel shaking in its absence. It is possible that the addiction is the same as in the previous examples, but this type of grafting is impossible without the internal activity of the NS. The irritants are inside the network itself. Maybe these stimuli are those neurons that have free conclusions that are looking for the right compounds, but I suspect that these are not the only stimuli. The presence of stimuli inside is the main difference between natural and artificial neural networks. The artificial NS acts only in the presence of external stimuli. I dare to notice

Thinking about addiction ...

Addiction ... I can’t say exactly what it is, but there are several versions of what is happening (I think they are all correct):

Version one "Change"

We feel only those actions that lead to the creation or rupture of synapses. Not so ... Actions that weaken or strengthen the synapses, and we just feel the intensity of the synapse changes. With any impact, neither the network, i.e. the arrival of signals into it, the synapses will change in it (synapse scales, breaks, connections of new ones, etc.). When the network is trained, the synapses practically do not change, we are used to the irritant and feel it.

Version two “Search for a nook”

The LNS begins to search for the position of bonds in which the effect on it will have the least intensity.

Version Three “Action-Counteraction”

Each effect on the LNS is suppressed by counteraction in the NA, i.e. The NS gives signals to suppress impacts on it. As a result of which we are still shaking after we got out of the car, which was traveling along the cobblestones.
In my opinion, these are not separate versions, but parts of a single whole.

Addiction

Consider a typical addict. Having tried the drug and having enjoyed it, he takes it again and again. Each time, he gets less and less pleasure, which makes him evade the dose. Finally, he wants to throw a harmful occupation, but he can not throw it - breaking makes you take the drug again.
It's no secret that the nervous system is a complex computer that controls all processes in the body. Receiving signals about the need, the nervous system sends signals to its satisfaction. For example, you feel thirsty - a whole chain of skills and habits is launched (what? Where? How?), Because the National Assembly “knows” what to do (it is trained in this). With natural needs, everything is clear, but how things happen with the "experienced".
Take gaming addiction for example. You sit down, play a game, enjoy it, as a result of which the LNS responsible for this is built.
What happens then? We have LNS, in which there are quite strong synapses (we played a lot of games). If you do not use this LNS, it will begin to collapse.
Imagine a large LNS built to a drug addiction, or a game, or some other activity ... This LNS begins to break apart synapses. The free conclusions of neurons again begin to look for their place in the network, connecting with other free conclusions. In all this chaos in the STS there is a huge amount of negative signals and interference. This, of course, also depends on the size of the LNS, which is being destroyed: the more neurons in it, the more it creates errors and interference. How can we conclude that a small LNS cannot be an addiction. (This, by the way, differs from a habit or skill), and that addiction is just a big LNS. We can also conclude that any frequent activity can turn into addiction and this is not always bad. Hard work and learning something will inevitably entail the creation of a large LNS, if it does not collapse with stress and bad emotions, as a result of training this LNS contrary to desire. It turns out that if you make an effort, then contrary to reluctance, the LNS can grow and "drown out" reluctance itself (or the LNS of "reluctance" will drown it, you can think and argue forever here). I dare to assume that everyone has their own dependencies and there’s no getting away from it, the main thing is that they don’t interfere with our life.

Again some lyrics ...

I dare to suggest that the mechanism of counteraction can arise as a simple training of some LNS perceiving pleasure signals or pains as negative signals and learning how to develop counter-active signals. And under the circumstances, an anti-signal starts to be generated, even in the absence of the signals themselves, which leads to severe discomfort of the NS. This is just a preview of my version. Here we can conclude that for one network the signal can be positive and at the same time be negative for another.
Here it is worthwhile to deal with the mechanisms of obtaining pleasure and discomfort. Getting pleasure is a positive signal for the LNS, discomfort is negative. Most likely, with the mechanisms for spreading errors, not everything is as simple as we think.
The mechanisms of the emergence and spread of positive and negative signals are not fully understood. It is not clear why the structure of the brain remains almost unchanged throughout life. In other words, what controls the neurons without letting them roam anywhere, but to remain on their territory.

findings

- Once again I became convinced of the importance of emotions (stress, joy, etc.) for the development of intelligence and it is not even clear what it is from the side of the National Assembly.
- Once again it was brought that the signals are received by the NS not only from the outside, but also from the inside.
- The same signal can be positive for one LNS and negative for another at the same time.
- We still do not fully understand the principle of the natural NS. Error propagation mechanisms are not fully understood.
- A modern ANN will never become an AI, as it is quite primitive (based on previous conclusions).
- If everything was clear, there would not be this article, because AI working instances would have long been there.
- Addiction is just a big LNS.
- Everyone has their own addictions and this is not always bad.
- It became clear why it is so hard to do nothing. Because as a result of inaction, synapses break and free neurons or their conclusions begin to look for where to connect, closing different sections of the NS, which forces us to do something or suppress the actions of other LNS. This makes us think that, perhaps, this is the “generator” of signals inside the NS and thinking itself.
- A neuron is not just a cell, but a kind of unicellular organism with its own needs and behavior (which no one took into account when modeling ANNs).

The source of thought began to be a video:


Thank you all for your attention. Waiting for interesting comments.