The question of the nature of the distributed memory of neural networks is considered. Since the memory capacity of a neural network depends on the presence of feedback in its structure this question requires further study. It is shown that the neural networks without feedbacks can be exhaustively described based on analogy with the algorithms of noiseproof coding. For such networks the use of the term "memory" is not justified at all. Moreover, functioning of such networks obeys the analog of Shannon formula, first obtained in this paper. This formula allows to specify in advance the number of images that a neural network can recognize for a given code distance between them. It is shown that in the case of artificial neural networks with negative feedback it is really justified to talk about a distributed memory network. It is also shown that in this case the boundary between distributed memory of a neural network and information storage mechanisms in such elements as RS-triggers is diffuse. For the given example a specific formula is obtained, which connects the number of possible states of outputs of the network (and, hence, the capacity of its memory) with the number of its elements.

Multivalued logics are becoming one of the most important tools of information technology. They are in great demand for creation of artificial intelligence systems that are close to human intelligence, since the functioning of the latter cannot be reduced to the operations of binary logic. At the same time, the problem of improving the efficiency of using the results of research in multivalued logics, as well as the problem of interpreting variables of multivalued logic, is acute. These problems create certain interdisciplinary barriers and make it difficult to implement the results of research in the field of multivalued logics in other fields of knowledge. It is shown that the problem of interpreting multivalued logic variables can be removed by establishing correspondence with fuzzy logic variables. Improving the efficiency of using of operations of multivalued logics and their variables can be provided by using their close connection to Galois fields. This connection, among other things, makes it possible to reduce any operations of multivalued logics, the number of variables in which is equal to a prime number, to algebraic functions whose arguments take values in Galois fields. This allows, among other things, to eliminate the very cumbersome constructions used in works on multivalued logic and make its apparatus convenient for use in related scientific disciplines in information technology. Direct verification of the adequacy of algorithms based on the use of Galois fields can be carried out by means of radio-electronic circuits, examples of which are presented in the present paper.

No abstract

No abstract