The report discusses approaches to the creation and development of modern intelligent systems based on cognitive approach. Most of the existing intelligent systems focused on the use of the computer paradigm. The basis of this paradigm laid the symbolic information representation and processing with algorithms focused on the logic and combinatorics, proceedings. This computer approach to solving some intellectual task is inferior to the man. Many experts began to consider new approaches to the solution of intellectual tasks. One such approach is the cognitive approach. Currently, he is associated with the results of many research areas and is difficult to understand and use. The author discusses the main steps of the creation and development of computer cognitive systems for simple model problems. The possibility of applying categories, the role and place of the Gestalt, the formation of characteristic values, which are important elements of cognitive semantics. Emphasizes the need for research in the field of knowledge of cybernetic systems, approaches to their representation, storage and processing, methods of interpreting and creating new knowledge.
Current approaches to the creation and development of computer cognitive systems Developed and applied intelligent systems, for the most part, focused on the application of computer paradigm. This is due to the fact that the emergence and development of computer technology has led to creating on its basis the methods and approaches for solving intellectual tasks. The basis of this computer paradigm includes symbolic representations of reality and their processing with algorithms and procedures that use logics and combinatorics. While solving a lot of such intelligent tasks computer systems are inferior to man. Experts in the field of artificial intelligence consider a number of approaches to solve complex problems for the computer tasks, one of which is the cognitive approach.
The basis of the cognitive approach focuses on the processes that are associated with the representation, storage, processing, interpretation, and the creation of new knowledge. For computer cognitive systems the important issues comprises the application of mechanisms common for developed biological systems headed by a man, and which are required for the solution of intellectual tasks.
Now many experts investigate the approaches connected with construction and
development of computer cognitive systems. These approaches focus on the basic
elements of cognitive science: concepts, prototypes, categories, Gestalt. Studying,
application and development of the systems with these elements demands the methods
taking into account cognitive aspects in processes of perception, thinking, knowledge,
explanation and understanding. The article [
It is considered that it is vital to start the research on these issues as it is noted in
the work [
Structure of cognitive system for simple model tasks One of a simpler model task can be a subject of the applied approaches to investigate cognitive computer systems focused on the solution of complex challenges.
The use of a simple model task is necessary to understand, to research, to create and to develop cognitive computer systems. The understanding of a system operation during the decision of a simple task is the result of studying and reconstructions of cognitive mechanisms used by a human while deciding such a task. Research is necessary for studying new structures of systems, their behavior and their development.
As a simple model task we can consider the situation in which it is necessary for the object with a cognitive computer system to find a subject from any point of the stated location, to reach it and to perform some operations.
Perceived Real World Subsystem perception of the Real world Subsystem current view of the Real world
Planning subsystem Knowledge base Subsystem impact on the Real world
Managed Real World
The structure of a cognitive computer system for the solution of the stated model task is shown in the figure 1. The system contains: • subsystem of Real world perception; • subsystem of a current Real world concept; • planning subsystem; • knowledge base; • subsystem of Real world impact.
The subsystem of the Real world perception represents a set of sensors measuring parameters of the surrounding world.
The subsystem of a current Real world concept collects the results of all measurements, transforms them to a format, universal for the system, and at the request of a planning subsystem represents the current values of corresponding signs (parameters). It represents a prototype of telemetric system.
The planning subsystem forms a great number of managing influence impacts to transform the object from a current to a demanded state . Integrated approach to represent knowledge is used.
The knowledge base contains knowledge necessary for planning managing influence impacts.
The subsystem of impacts on the Real world carries out the realization of the managing impacts (commands) of an object of the Real world. 3
The creation of computer systems with the use of cognitive mechanisms The main approaches to the creation of computer systems using cognitive mechanisms to solve simple model problems are the following:: 1. A system based on knowledge of the classical type. It is assumed that you are using the ideas about the objects of reality as concepts generated by the subsystem of perception of the Real world. Subsystem perceives reality through a sensual image. These concepts are presented in the format of an integrated approach to knowledge representation. 2. Creating and using object subsystem of formation and application of concepts. The concept of the authors Z. D. Popova and I. A. Sternina refers to: "...discrete mental education, as the basic unit of mental human code possessing a relatively orderly internal structure, which is a result of cognitive (cognitive) activities of the individual...". In computer cognitive system, mental concepts are represented by units in structure using both real and mental nature. Mental entities are generated based on the single units, and then by general (abstract) concepts of the real world. In the development of cognitive systems at first single units, and then abstract (generic) mental concepts and signs can be applied. 3. Creating and using object subsystem of formation and application of categories. In computer systems, cognitive categories are initially formed with the participation of a human. On the basis of machine learning, for example, deep learning of neural networks a model mechanism for the creation of categories is formed based on the experience of the inner world of a person (team of specialists). After you have successfully created a model of the mechanism of the formation of the categories it replicated for future application without human intervention. 4. Creating and using for the object the subsystem of formation and application of new concepts based on the measuring characteristics (parameters), concepts and categories. This subsystem must ensure the development (evolution) of computer cognitive systems. The proposed cognitive mechanisms can serve as the basis for building new applications. The provisions of this approach can be supplemented and changed depending on the tasks assigned. In the author's opinion, the use of prototypes at one of the stages of creating computer cognitive systems is promising.