Multi-Valued Logics as an Advanced Basis for Artificial Intelligence (As an Example of Applied Philosophy)

Abstract

It is shown that the creation of artificial intelligence systems, gradually approaching human intelligence, cannot be limited by binary logic and the classical interpretation of the category of truth. The nature of human thinking is variable, which requires the use of flexible algorithmic platforms that operate with multivalued logic. It is shown that the existing approaches to the interpretation of the essence of intelligence, going back to the Turing test, are not consistent. More adequate is a criterion built on the ability of the intellect to lie, including creative particulars. It is shown that using multi-valued logic is a promising tool for constructing the algorithmic basis of artificial intelligence. It is shown that the use of just such logic is of interest from the point of view of studying self-organization processes in a telecommunication environment, resulting in the appearance of “spontaneous intelligence”, too. An analogue of such intelligence, for example, is the phenomenon of bureaucracy. It is shown that the question of the essence of the intellect is a prerequisite for the further development of non-trivial logical systems since the functioning of the intellect cannot be reduced to the operations of classical formal logic.

Formation of Hydrophobic–Hydrophilic Associates in the N-Vinylpyrrolidone and Vinyl Propyl Ether Copolymer Aqueous Solutions

“…The outcomes delineated in [33], forming the foundation upon which polymer solutions may similarly be regarded as neural network analogues, complement the postulates set forth in [37][38][39]41]. Evidently, the inference can be drawn that intricate systems spanning a spectrum of diverse realms (in the philosophical sense of the term [42]) can be construed as akin to neural networks.…”

Convolutional neural networks from the perspective of the problems of multivalued logic

“…The objective of this work is to develop tools to overcome this difficulty, i.e., to develop a methodology that would allow to bring the calculation of the convolution operation to algebraic operations that exclude the need to use fractional numbers. In fact, we are talking about making it possible to use the digital convolution theorem [22], which cannot be used directly, since the physical meaning of the convolution operation in Galois fields does not coincide with the meaning of operation (1).…”

“…This problem, which can be described as a problem of overcoming the logical opacity of neural networks, has a pronounced methodological aspect [22,23]. As it was emphasized in [23], already at the present stage of research, it is permissible to raise the question of decrypting those algorithms according to which the neural network created by the human brain functions.…”

“…Moreover, in order to understand what "artificial intelligence" is, it is necessary to understand what intelligence as such is [26,27]. At the same time, there is no doubt that these algorithms cannot be reduced to binary logic [22,23]. There is also no doubt that convolutional neural networks are also of interest from these positions, since they largely rely on an analogy with human vision and the nature of visual information processing by the brain [28,29].…”

Features of digital signal processing algorithms using Galois fields GF(2n+1)

“…The development of electronic circuits operating in Galois fields is also of interest from the point of view of improving artificial intelligence (AI) systems. Namely, as it was shown in the reports [ 11 , 12 ] the further development of AI, assuming its gradual approach to the biological prototype, cannot exclude the transition to multivalued logic, since human thinking is irreducible to binary logic. It is appropriate to emphasize that the problem of creating of AI approaching a biological prototype is of considerable interest, including from the point of view of revealing the essence of intelligence as such [ 13 , 14 ].…”