Название: Cognitive Semantics of Artificial Intelligence: A New Perspective
Автор: Alexander Raikov
Издательство: Springer
Год: 2021
Страниц: 130
Язык: английский
Формат: pdf (true), epub
Размер: 10.1 MB

This book addresses the issue of cognitive semantics' aspects that cannot be represented by traditional digital and logical means. The problem of creating cognitive semantics can be resolved in an indirect way. The electromagnetic waves, quantum fields, beam of light, chaos control, relativistic theory, cosmic string recognition, category theory, group theory, and so on can be used for this aim. Since the term Artificial Intelligence (AI) appeared, various versions of logic have been created; many heuristics for neural networks deep learning have been made; new nature-like algorithms have been suggested. At the same time, the initial digital, logical, and neural network principles of representation of knowledge in AI systems have not changed a lot. The researches of these aspects of cognitive semantics of AI are based on the author's convergent methodology, which provides the necessary conditions for purposeful and sustainable convergence of decision-making.

A human sees AI as such a cognitive system. Despite a relatively long history of this term, AI has not yet encroached on many of the cognitive abilities of humans. The reason is that modern AI does not go beyond formalized systems and discrete logic, even the most complex one entangled in deep artificial neural networks. AI is immersed in an electronic computer and an iron robot, both having neither human soul nor ability to suffer and empathize.

The concept of cognitive semantics can be adopted to differentiate “strong AI” from “weak AI.” Weak AI systems only process symbols without a non-formalizable understanding of what they mean. This is a traditional look at AI systems, which can also be “narrow” and “general.” A “narrow AI” system solves a particular problem better than humans would do. An artificial “general” intelligence (AGI) system can solve problems in different application domains, with better results than humans. Sometimes AGI and strong AI are considered as synonyms. Therefore, weak AI uses only denotative semantics. Strong AI processes symbols, as well, but it can understand what they mean owing to cognitive semantics. Previously, the difference between “weak AI” and “strong AI” has been most important to philosophers and practitioners irrelevant to AI. With the concept of cognitive semantics, strong AI systems can become a kind of cyber-physical systems, which have engineering character and can be used in different sectors of the economy.