The theory of creating artificial intelligence

Creation of artificial intelligence - myth or reality?

Theory

Working with modern PIC or AVR type microprocessors, an idea came to my head — a theoretical idea. And I can not say that the same idea did not occur to people specifically involved in the practical creation of AI. The essence of this idea is to recreate the human brain and its work by connecting microcontrollers to processors and data storage.
I will not go into the details of the device and technology of the microcontroller - you can do it yourself by going to the Wikipedia website.
The idea is that the group of microcontrollers is responsible for certain parts of the human brain.
Sounds silly and unreal you say. But having really studied the work of the human brain, having thrown out (for a time) the brain sections that are unnecessary for the robot and correctly programmed a few basic ones, it is realistic to achieve a result in which artificial intelligence will be similar to ours - human.
Let us give an example: sight - having connected a camera (eyes), people have already created a program code for working with a camera. That is, we launch the robot -> the camera turns on and in front of it is a room with furniture, computers, etc. Our human brain instantly calculates the length and width, where you can go and where you don’t, WHERE OUTPUT is task one. - That is, exit and entry visually the first task.

What will the robot do?
Answer: NOTHING.
Now, a little about our "thinker" - I read in a book, I do not remember the name. I won’t write how many billion neurons are connected and involved, but the fact is that the cerebral cortex is responsible for all types of memory: auditory, taste, visual, olfactory, tactile. Example: you go home from work and walk five blocks, all this time you’ve come across people, cars, buildings, houses, etc. Our brain records all the information. If you are asked, have you seen a girl in a bright red dress in winter along the way? You didn’t even pay attention to this girl, but she passed next to you - and your visual organ has already transmitted information. You think and answer - NO. But this girl and her image has already been saved in one of the parts of the brain. If I'm wrong, correct me. But this is one of the examples and I cite it for the logical creation of the algorithm of vision. (Although I said that it is impossible to create AI with algorithms, you can only solve some tasks at first)
Our main task: to create an algorithm in which all information entering the visual organs of the robot will be classified and recorded in the data warehouse. The main place of course we take to the house (the workplace where the machine will be used) secondary to his walks.
Yes, I understand that this is an unimaginable data stream and to distinguish it you will have to write billion code and the microcontroller will not help here - but it will be responsible for the eyes themselves (food, changing the order of things, range of objects, input and output, etc. )
First of all, I thought to recreate the musculoskeletal system, but after reflecting I realized that it is even more complicated than the visual one.

The main question that arises is: how will he think? How will he choose?
Main answer: A person is born and also does not know how to walk, write, read, speak, etc. Man learns everything along the way. We just need to create technology and design the main parts of the brain, write down the basic human actions and realize the ability to learn - RECORDING INFORMATION to a permanent repository, from where the robot could take answers and asking new questions would preserve possible new solutions. Why not? Just digging through habrhabr I came across an article where they already did something similar in IBM - an article called “Functional computer model of the brain” (if I had seen it before I would not have written this article, but it's too late) The guys are trying to completely simulate the brain and even the solution of all tasks that would go the way of a person.
The idea of ​​this post is quite simple - to describe the really working principle of creating artificial intelligence, that is, reason. Create a machine that can not only programmatically respond to the tasks set by man by the method of algorithms (equations) and the machine (the robot, without using cybernetics as a method to achieve the goal) able to think, choose, plan their lives and of course help the person.

Previous attempts of mankind

Many of you watched the popular science television program Discovery Science, where Japanese, Chinese, Americans, and even Russians tried and created robotics. They invested millions of dollars and spent years developing what the science fiction writers called the “terminator”, “robocop”, etc. Without mentally listing all the authors and directors who created this miracle, but mankind is still at the stage of development of robotics for just one big reason: I have never succeeded in creating artificial intelligence similar to the human brain. No robot can consciously make a choice.
The Internet is full of videos, where Japanese technologists (Honda, if my memory serves me) tried to create a real machine, exactly like a person, according to external data. What did the "scientists" not take into account? - This is a feature of the structure of the human brain - as the most important processor of the human body. The created robot from metal alloys was programmed for certain commands and when the operators and the whole plant gathered for a presentation of the launch of their miracle machine, something that none of them could predict happened. The robot took one step and fell to the ground like a pile of metal. I do not blame its creator, this video was filmed in 2000 and uploaded to YouTube recently, a lot of things happened in the past 13 years, probably in the same Japan or China already created prototypes of working humanoid machines and are waiting in the wings. But I am writing this book for those who not only wondered but also for those who have been looking for an answer to it for a long time: is it possible to create artificial intelligence? If so, how? If not, how to make it possible. In this book, I will also devote some chapters to issues related to personal security and the real threat to a person who has such a robot on the farm. But all in order.
')

Prehistory

The theory of creating artificial intelligence originated in my head more than ten years ago, during the boom of computer games. To be more precise - Counter-Strike is one of the first games that millions of teenagers played, held championships and tournaments, created clans. Thousands of hours of my life were given to this masterpiece by programmers. The only thing that depressed was the artificial intelligence of the bots (opponents). Playing against the computer - it was possible to hone their skills to such a level that 10,20,30 opponents turned into meat. But having come to the computer club and playing with real people - all the training went on a smack. I could defeat 20 computer enemies alone. But in reality, we sometimes in five could not defeat one person.
Of course, developers created and optimized bots (artificial enemy), developed new tactics, tricks, techniques for each card, but one hour was enough for the notion of order of action and algorithm of the enemy - this is not artificial intelligence - this is a problem that all programmers encounter creating AI ( artificial intelligence)
Playing modern games, we can see the progress of the developers. Modern opponents not only worry about their lives, but also behave like people — they hide behind obstacles, throw grenades, smoke you out of their shelters, fire from guns, covering comrades. All this is of course very good, BUT - you guessed what kind of “BUT” it is.

If this idea is already used, then I apologize for the time spent.

If not, then I hope to soon provide more information about the idea and its implementation.