How to make artificial intelligence? Artificial intelligence technologies

Table of contents:

How to make artificial intelligence? Artificial intelligence technologies
How to make artificial intelligence? Artificial intelligence technologies

Artificial intelligence like natural intelligence is a reality. At the end of the last century, there were several projects that not only substantiated this possibility, but also demonstrated concrete results.

The ubiquitous introduction of computing technology has created a moral obligation for noble academics, professors and eminent specialists (the elite) to take an active part in the process and "lead the wave". The authority of the elite is a strong thing, everything below the rank bows, and the young scientific and technical generation (students, graduate students, engineers) goes underground so that when everything is over, a little later, they can live and work.

To fall under the hot hand of elite knowledge, inspired by the rosy prospects of a blitzkrieg in anticipation of world recognition, means never to do anything, and it will be possible to say goodbye to a career forever.

Fantasy and startingfoundation

Fiction artists, fans, skilled engineers, really respected scientists and educated people have always thought about artificial intelligence, wrote, sighed and fantasized about how to make artificial intelligence.

Fantasists and engineers
Fantasists and engineers

When computers gained strength and began to demonstrate stable performance, when programming got stronger, and programming languages accepted the status of a practical and reliable tool. When database management systems appeared and specialists saw practical results… Then even an ordinary qualified specialist saw what power a program that processes information can give. I thought: yes, she can learn, she can draw conclusions and solve problems, probably, it is already possible to create artificial intelligence, and this will be a step into the future.

Objective reasons for the first failure

All scientific centers and schools have begun to design expert systems, create knowledge bases, design intelligent systems, and write popular articles, generating public interest.

The elite led the wave

Historically there were three scientific centers: Novosibirsk, Moscow and Kyiv. Scholars and specialists from St. Petersburg, Minsk, Kazan, Perm and other cities took part from time to time.

There were no results, but there was a lot of talk and fuss. Meanwhile, computers have already been introduced at large enterprises, automated control systems and databases have already been created.

The First All-Union Conference on Artificial Intelligence(Pereslavl-Zalessky, November 1988), having absorbed the voluminous works of the elite, showed zero intellectual and energy potential.

From the technical side, there were all the possibilities for creating artificial intelligence. It's just that the elite did not have the necessary knowledge and skills

The elite had not yet moved away from the creation and management of large enterprises for the production of software and were captivated by the ideas of automated process control.

The program will do everything - the elite claimed. But the program is not capable of development. The program only works as defined by the author

No intelligence could even originate in the program or in the elite knowledge.

Real projects of the last century

If we understand artificial intelligence (AI) as similar to natural, then its creation requires public sanction, legal support and state protection. This is too serious for the individual and society. Artificial intelligence is a state-wide program.

It is in demand, interesting and practical, if AI is understood as an intelligent system:

  • enterprise management;
  • spacecraft trajectory calculation;
  • managing public transport;
  • ensure proper nutrition at school;
  • sporting events, etc.

The emergence of the theory of inventive problem solving (TRIZ) and the creation of a community of TRIZ specialists is an excellent opportunity to create an intelligent system. But the project "Inventing Machine" (late 80syears) was never implemented.

Reason for failure. Orientation to the language of artificial intelligence "Prologue", the lack of a reasonable organization of work and the ambitions of the elite led to intellectual anarchy in the labor collective, in which, moreover, there was not a single qualified specialist.

But the result was still there. Attention was drawn, new knowledge appeared among specialists, experience was gained.

Inventions, effects, patterns
Inventions, effects, patterns

By and large, the task was primitive: there is a database of inventions, physical, chemical and other effects (patterns), you just had to competently write a dialogue with the user-engineer and help in realizing his creative potential at the enterprise.

Another interesting project was built on a less rich theoretical foundation and no external environment at all. The theory of primitive reflexes is the result of a generalization of developments in the psychology of behavior, the logic of thinking and particular tasks.

Task 1. There is a mouse and there is a maze. Any labyrinth, anyone builds it, but not the author of the program. Result: the mouse learned, remembered mistakes and passed any maze. The implementation was done on an Atari computer, the language is Basic.

Task 2. A child under three years old (figuratively). The task is to teach the little man situational behavior: mom, dad, ball, toy, Masha is good, Petya is a boy, a round object, water flows, a ball rolls and flies, what is your name, etc. - primitive dialogues from pink childhood. The implementation was performed on the DVK computer(PDP/11). Assembly language.

Outcome 1: The program did what it was designed to do. Understood the question and answered. I studied and answered such questions adequately

The second result is especially important!

Outcome 2: A stranger passed by, but I wanted to laugh at the intellectual program. The topic of childhood was of little interest to him. But he himself taught this program elementary arithmetic: add, multiply, divide. This is the first sign that the program can be called intelligent: a real step into the future - it solved a problem for which it was not programmed

There were not so many such projects in the late 80s, but they were. For various reasons, they did not take place and did not reach the practical result.

Perhaps this was facilitated by perestroika, but there is a characteristic moment: on the other side of the ocean and in the high-tech Land of the Rising Sun, where it was quiet, calm, a lot of money and technology - the result also failed, and the question is how to make artificial intelligence remained open.

Object-oriented programming

From the pinnacle of modern knowledge, it can be argued that none of the projects of the last century to create an intelligent system could take place, and if it did, it would be subject to a radical alteration. This is due to civil strife: programming languages fought for the palm, and the development of programming technologies did not bother anyone.

Specialists did not know how to work quickly, did not have the necessary knowledge and skills.

Only in the early 90sthe contours of the transition to object-oriented programming were outlined and a real opportunity to achieve real results began to appear.

Supercomputer technology and artificial intelligence is not the right party. Computers need to be fast and reliable, but that's not enough. Classical programming on a computer of any power will not give the desired result. The result of a programmer's work is a program, and this is the frozen knowledge of the author.

Supercomputers and intelligence
Supercomputers and intelligence

The question is how to make artificial intelligence, in fact, the question is how to implement dynamics: the program should develop outside the will of the author and solve problems that the author did not subscribe to.

HTML & XML Hypertext

Object-oriented programming (OOP) is a technology, a tool, a concept. At the time of its birth, the PLO fell into the wrong environment. Programmers used OOP as a tool for creating interfaces, programming constructs, communication channels, file systems, but not for real objects.

A real object is, for example, a word, a phrase, a sentence.

Text analysis
Text analysis

Instead, hypertext has entered the world of information technology. The Internet is grateful for a new approach to the formalization of information, but this is not a development to artificial intelligence, but a slowdown in it. A characteristic feature of the mentality of a skilled programmer is to keep the consumer (user) at a distance and improve the level of misunderstanding by the latter of the real situation.

Medics do the same. If, say a lotto the patient, it means that he will think a lot and be sick longer. But medicine cares about human he alth and can be justified.

Programming is hard to justify. We are talking about the processing of information of the consumer (user). The programmer has no right to restrict the rights of the owner (owner) of information and reserve the exclusive right to own what does not belong to him.

Object and object system

Object-oriented programming (OOP) is real: an object and systems of objects, and this is the real answer to the question of how to make artificial intelligence. This is a working technology and a guaranteed result.

To create a simple intelligent system, it is enough to implement objects: "word", "phrase", "sentence". Formulate the "symbol" object and define the systems of interaction between the created objects. The next step is to implement natural language grammar and syntax. Next, you need to start teaching the intelligent system the ability to solve a problem.

Semantics and Syntax
Semantics and Syntax

Outwardly everything looks simple. Technologically, this is solvable, but this is not yet artificial intelligence. A program cannot be called intelligent as long as it functions at will (based on the author's algorithm).

Everything that is called intellectual can function only on the basis of knowledge (moreover, knowledge acquired personally). Only when a program can accumulate and apply knowledge (learns) can it be considered almost intelligent. If the program can transformacquired knowledge for solving related problems - the goal is almost achieved.

The elite is gone, the Oracle has arrived

The new age wasn't particularly fond of artificial intelligence. The public consciousness was puzzled by social and economic problems. But leading IT companies and successful corporations in the field of information technology have accumulated enough knowledge and skills to generalize and use them at an expert level.

The elite remained in the last century. In the new century, information technology monsters, developers of large databases and expert systems, entered the arena. Technological bottlenecks remain, but programming today is a remarkably fast paced process. Everyone needs and needs to do everything quickly.

Oracle is the undisputed leader and undisputed authority. This is an indicator of the dynamics of the information industry. Throughout its product line, the word intelligence in action can be traced. This is not really a practical application of artificial intelligence, but it is really a demanded application of the knowledge of Oracle experts in a real situation.

Information Technology Leaders
Information Technology Leaders

Virtually all information companies have embraced and applied expertise to their products. The accumulated experience went into the system of standard reactions to user questions. All questions that the company's product could not answer on its own flew to the call-center, info-center or support, where the specialist gave a human answer.

Excellent and practical technology. Programming artificialintelligence on the fly! The user of the company's product does not care how he received the answer to his question, how exactly the problem was solved. He got the result. The consumer does not care whether artificial intelligence is in the product of this company or not. The user (consumer) needs a result.

Chinese boats

Cat-or dog-robot with artificial intelligence that Tamagotchi. The idea is chic, but entertaining. It's not about the intellect, but about the emotions. The buyer gets the satisfaction of owning a smart product of technological progress and having something to show to friends. Peace idea.

Robot dog
Robot dog

Combat version of the robot with artificial intelligence in a frightening performance.

As a thunder in broad daylight, the media presented the success of Chinese scientists. A couple of hundred drones, made in the form of boats, drew aerobatics at sea.

Beautiful pictures were painted by boats. They walked in an even formation, then split into groups, described circles and converged back. A great many aerobatics were performed, but the information flashed through the air and disappeared into space.

It would be interesting to see how these Chinese robots will be able to keep the formation and how many figures they can unsubscribe in the safest Belarusian swamp.

Flying, crawling, floating

The fashion for drones of all types of troops in the military of all countries has turned into a disease. But here is the question. If you block communication, eliminate the operator or turn off the electricity, whichwill make the "iron warrior" in the air, on land and at sea?

But the manufacturer declares undoubted fighting qualities, armed with artificial intelligence. And if the oil runs out in the engine or the propeller jams? You never know, the elephant will block the view. There are many options and situations. There is no chance for such artificial intelligence.

If we talk about artificial intelligence, then any flying, crawling or floating combat intellectual product must receive a task, complete it and return to base on its own. Well, as soon as there is no such thing, then there will be no artificial intelligence in the next hundred years.

About the essence of things: what does a person need?

It is doubtful that society is ready for the arrival of smart robots. Smart homes already exist, but software technologies do not account for all options for the "life" of a smart home.

There are always many options, and a team of highly qualified programmers can only foresee what they know and what is written in the terms of reference.

It will not work in any way to implement the learning option for a smart home. Any mistake (the smart home has learned a little), but what will the household say about this? What if a smart home learned how to restore power or clean sewers, and the learning process took only one week?

Smart home is learning
Smart home is learning

The second failure and the second attempt to learn something will be fraught with the elimination of smart home software, and the owner will be right. You need to study on the side, and not in real life.

Modern technologies are not ready to realize artificialintelligence in full, which provides for the consideration of all possible situations. However, it is not possible to cover all situations.

By limiting the use of software to specific "intellectual" limits, one can find a person's interest in using them. Practice shows: you cannot provide everything, do what you are absolutely sure of and the consumer will be satisfied.

About the reality of artificial intelligence

Artificial intelligence like natural intelligence is a reality. At the end of the last century, this was proven in practice. If you can talk about several successful projects in one place on Earth, then in reality there are a few more of them.

Nothing ever happens in a single copy, nature always duplicates everything and the results of human hands too.

The essence of artificial intelligence is knowledge, not a program. Knowledge should be the basis for acquiring and using new knowledge. Man is the crown of nature, but his "engine" in a logical sense is technically reproducible.

artificial intelligence programming
artificial intelligence programming

One should not attach importance to emotions, soul, ideas about morality and morality. Not much is required from artificial intelligence: just help a person in solving everyday problems.

Popular topic

Editor's choice