Artificial intelligence as a set of questions

When we talk about strong artificial intelligence, we understand that this is not an isolated question, not a thing in itself, but a question the answer to which implies an explanation of all phenomena that are associated with a person’s thinking. That is, answering the question about the nature of intelligence, we will inevitably have to answer such questions as:

• What is information?
• How does the brain represent knowledge?
• What is language?
• What is the role of language in thinking?
• How are things done?
• How is planning done?
• What is the nature of fantasy and memories?
• What is motivation?
• What is the nature of emotions?
• Where does the variety of emotional evaluations come from?
• What is the point?
• How is thought born and what is its nature?
• What is attention?
• What is love?
• What is harmony and beauty?

A lot of questions give rise to the obvious, one can say “naive”, the idea that the brain is a set of systems, each of which is responsible for solving a specific task, and the listed questions relate either to the actions of certain systems or to the processes of information exchange between them. This approach is also pushed by the engineering experience of designing computer or other systems, and the human body itself, which consists of organs with a pronounced specialization. And everything would be fine, but the structure of the real brain is well studied and there are no systems with isolated functions resembling those that arise from the questions asked.

Because of the need to comply with the explanations given by neurophysiology, nothing else remains but to assume that all these processes are a consequence of some kind of unified architecture of the brain, and there are its manifestations in various situations. But what is this amazing architecture? What do we know about her?
')
The basic opinion regarding the current state of searching for the nature of thinking, which, by the way, was formed not so much by scientists, but mainly by popularizers who want to understand the problem, is that: “We know a lot about brain physiology, there is a huge theory of neural networks and neural network management, There are many psychological theories based on an abstract view of the brain, but at the same time there is no explanation for the mystery of thinking, an explanation of how the brain operates with information and which ones otsessy responsible for the ability to think. " To this is sometimes added the assertion - a variation of the “Chinese room” paradox, that: “the creation of a strong AI is impossible in principle without modeling biological consciousness”. And this leads to the argument that all “rational” models should be treated only as attempts to model a certain “technical” part of the brain, and one should not expect a complete clue of the main secret from them. But is it?

A huge number of very intelligent people tried to give answers to the above questions. There are many good theories. True, the difficulty is that many of these theories exist in isolation, not trying to dock with the answers to other questions. The reason for the complexity is clear: the questions relate to different scientific disciplines, and playing in a foreign field is difficult, at least due to the difference in terminology. That is, for each question there are many possible answers, among which it is possible that there are correct ones, but we need a concept that fits into the known data on the structure of the real brain, which will allow in each area to choose the “matching” theories.

So, the most interesting thing is that such a concept, and the answers to all the above questions all already exist and, strange as it sounds, but the question of creating a strong AI is in many ways ideologically clear. The remaining questions are rather technical in nature and relate more to implementation algorithms.

I have prepared several lectures that allow me to penetrate the understanding of the basic principles and, as I hope, move to a new level of understanding of the issues of AI. Unfortunately, it turned out a long, very long time, but many great minds had time to work on the question, and if it’s shorter, it turns out to be either completely incomprehensible, or the logic of reasoning breaks down too much. A prerequisite - if you listen, then consistently. It is useless to start from the middle, looking for the answer to the question of interest, understand the words and intonation, but surely miss or misunderstand the main thoughts. Even if it seems to you that some topic is familiar to you, do not rewind - my task was not to explain once again well-known principles, but to focus attention on those global ideas that stand behind them.

The common name of the cycle: "The logic of thinking."
Part 1.
Intro of a strong AI.
Behaviorism.
Perseptron Rosenblatt and his connection with the ideas of behaviorism.
Training with reinforcements.
Assessment of the quality of the situation as a signal of reinforcement.
Adaptive critics. Q and V critics.
Motivation. Models of explicit and implicit motivation.
The structure of the brain. Bark.
The work of the cortex on the example of the primary visual cortex.
The neocognitron Fukushima, as a model of the visual cortex.
Generalization Analogy with factor analysis. The essence of generalization.

Part 2.
The basis of the projection system.
Principles of self-organization.
Cognitive brain architecture.
The mechanism of formation of emotions.
Emotions as an assessment of the quality of the situation.
The reduction of ratings to fears and anticipations.
Emotions and natural selection. The formation of emotions in ontogenesis.
Examples of the formation of emotions.

Part 3
The principle of a mad programmer. Emotional basis. Emotional focus.
Thought and awareness.
Record thoughts.
The paradox of the "Chinese room".
Associative blurring.
Neural correlates. Brain rhythms.
Thinking.

Part 4
Evolutionary sequence of brain formation.
Knowledge.
Information.
Language and thinking.
Philosophy of language.

Part 5.
The concept of emotion.
Waiver of predestination.
Love is like an emotion.
Change of emotions.

Part 6.
Social emotions.
The meaning of the word beauty.
The classic explanation of beauty. Explanation of beauty through the nature of emotions.
Harmony.
The beauty of man. Clear beauty. Sexuality. "Mysterious" beauty.

Part 7
The beauty of information transfer.
Social status.
Emotion is funny.
Confusion of signs and essence of the funny.
Humor.


Playlist