Basic concepts and definitions. The functional structure of the use of AI

Basic concepts and definitions.

Artificial Intelligence (AI) is the science of concepts that allow a VM to do things that people look reasonable. But what is human intelligence? Does this ability to reflect? Does this ability to absorb and use knowledge? Does this ability to operate and exchange ideas? Undoubtedly, all these abilities are part of what is the intellect. In fact, it seems impossible to define in the usual sense of the word, because intelligence is a fusion of many skills in the field of information processing and presentation.

The central tasks of AI are to make the VM more useful and to understand the principles underlying intelligence. Since one of the tasks is to make VMs more useful, scientists and engineers specializing in computing technology need to know how AI can help them in solving difficult problems

1. Application area.

   1. Verbal conceptual learning.
   2. Learning networks (neural networks);
   3. Natural language processing;
   4. Writing machine music;
   5. Adaptive programming;
   6. Making decisions;
   7. Pattern recognition;
   8. Games;
   9. Proofs of theorems;

Future plans for AI applications: In agriculture, computers must protect crops from pests, prune trees and provide selective care. In the mining industry, computers are designed to work where there are too dangerous conditions for people. In the production of VM should perform various types of tasks for the assembly and technical control. In the institutions of the VM are obliged to create schedules for groups and individuals, to make a brief summary of the news. In educational institutions, VMs should consider the problems that students solve in search of errors, just as errors are looked for in the program, and eliminate them. They should provide students with superbooks stored in the memory of computing systems. In hospitals, VMs should help diagnose, refer patients to appropriate departments, monitor the course of treatment. In the household, the VM should help with tips on cooking food, purchasing food, monitor the condition of the floor in the apartment and the lawn in the garden. Of course, at present, none of these things is possible, but research in the field of AI can contribute to their realization.

A brief historical overview of the development of works in the field of AI.

The beginning of research in the field of AI (late 50s) is associated with the works of Newell, Saiman and Shaw, who investigated the processes of solving various problems. The results of their work were such programs as "LOGIC-THEORETIC", designed to prove theorems in the calculus of statements, and "GENERAL SOLVENTOR OF TASKS". These papers marked the beginning of the first stage of research in the field of AI, associated with the development of programs that solve problems based on the use of various heuristic methods.

The heuristic method of solving the problem was considered as peculiar to human thinking "in general", which is characterized by the appearance of guesses about the way to solve the problem and then check them. He was opposed to the algorithmic method used in the computer, which was interpreted as the mechanical implementation of a given sequence of steps, deterministically leading to the correct answer. The interpretation of heuristic methods for solving problems as a purely human activity led to the emergence and further spread of the term AI. Thus, in describing their programs, Newell and Simon cited as arguments confirming that their programs simulate human thinking, the results of comparing evidence records of theorems in the form of programs with reasoning records of a person. In the early 1970s, they published a lot of data of this kind and proposed a general methodology for developing programs that simulate thinking. Around the time that the work of Newell and Simon began to attract attention, MIT research groups also formed at the Massachusetts Institute of Technology, Stanford University and the Stanford Research Institute. In contrast to the early works of Newell and Simon, these studies were more related to formal mathematical concepts. Methods for solving problems in these studies evolved based on the expansion of mathematical and symbolic logic. The modeling of human thinking was of secondary importance.

Further research in the field of AI was greatly influenced by the emergence of the Robinson resolution method, based on the proof of theorems in the logic of predicates and being an exhaustive method of proof. The definition of the term AI has undergone a significant change. The purpose of research conducted in the direction of AI, was the development of programs that can solve "human problems". So, one of the prominent AI researchers of the time, R. Benerjee, wrote in 1969: "The field of research, commonly called AI, can probably be represented as a set of methods and tools for analyzing and designing machines that can perform tasks that could be handled until recently only a man. At the same time, the speed and efficiency of the machine must be comparable with the person. " The functional approach to the focus of research on AI has been preserved mainly to the present, although a number of scientists, especially psychologists, are still trying to evaluate the results of work on AI from the point of view of their correspondence to human thinking.

The research ground for the development of AI methods at the first stage was all sorts of games, puzzles, and mathematical tasks. Some of these tasks became classical in the literature on AI (problems about monkeys and bananas, missionaries and cannibals, the Hanoi Tower game at 15, and others). The choice of such tasks was determined by the simplicity and clarity of the problem environment (the environment in which the solution of the problem unfolds), its relatively small cumbersomeness, the possibility of a fairly easy selection and even artificial design "under the method". The main flowering of this kind of research came at the end of the 60s, after which the first attempts were made to apply the methods developed for problems solved not in artificial, but in real problem environments. The need to study AI systems in their operation in the real world led to the formulation of the task of creating integral robots. Carrying out such work can be considered the second stage of research on AI. Experimental robots operating under laboratory conditions have been developed at Stanford University, Stanford Research Institute and some other places. Conducting these experiments showed the need to address fundamental issues related to the problem of representing knowledge about the environment of functioning, and at the same time insufficient research of such problems as visual perception, building complex plans of behavior in dynamic environments, communication with robots in natural language. These problems were more clearly formulated and posed to researchers in the mid-seventies, associated with the beginning of the third stage of research of AI systems. Its characteristic feature was the shift of attention of researchers from the creation of autonomously functioning systems, independently solving the tasks assigned to them in a real environment, to the creation of man-machine systems that integrate human intelligence and VM capabilities into a single whole - solving the task integral human-machine solving system. Such a shift was caused by two reasons:

1. By this time, it turned out that even simple at first glance problems that arise before an integrated robot during its operation in real time cannot be solved by methods developed for experimental tasks of specially formed problem environments;
2. It became clear that the combination of complementary capabilities of a person and a computer allows you to circumvent sharp corners by shifting to the person those functions that are not yet available for computers. The focus was not on the development of separate methods for machine problem solving, but on the development of methods for the tools that ensure close interaction between a person and the computing system during the entire process of solving a problem with the possibility of prompt changes by the person during this process.

The development of research on AI in this area was also due to a sharp increase in the production of computer equipment and their sharp reduction in price, making them potentially accessible to wider circles of users.

The functional structure of the use of FIC.

This structure consists of three sets of computing tools (see figure). The first complex is a set of tools that execute programs (executive system), designed from the standpoint of effective problem solving, in some cases has a problem orientation. The second complex is a set of intelligent interface tools that have a flexible structure that provides the ability to adapt to a wide range of interests of end users. The third set of means by which the interaction of the first two is organized is the knowledge base, which ensures that the first two complexes use a complete and system of knowledge about the problem environment that is independent of the processing programs. The executive system (IS) combines the entire set of tools that ensure the implementation of the established program. Intelligent interface - a system of software and hardware that provides the end user with the use of a computer to solve problems that arise in the environment of his professional activity, either without intermediaries or with little help from them. The Knowledge Base (BR) - occupies a central position in relation to the rest of the components of the computing system as a whole, through the BRs the integration of the means of the Armed Forces involved in solving problems is carried out.