Post-Non-Classical Artificial Intelligence and its Pioneer Practical Applications

摘要

Artificial intelligence (AI) approaches and methods have been developed since the beginning of the last century. This paper identifies three stages in the history of AI development: classical, non-classical, and post-non-classical, where each stage includes the capabilities of the previous one. Classical and non-classical approaches include logical constructions, deep learning, genetic and ant algorithms, denotative (formalised) semantics and ontologies. Post-non-classical approaches aim to embrace latent parts of consciousness by indirect methods, including convergent approaches, quantum semantics, topologies and category theories, inverse problem solving, cognitive modelling and the like. The convergent approach takes into account the collective unconscious and can help to enrich conceptual models with cognitive (non-formalised) semantics. The continual power of cognitive semantics is tens of orders of magnitude greater than that of denotative semantics, but there are as yet no methods for using this potential. Examples of the real practical implementation of the components of post-non-classical AI approaches are given, and a design for a digital platform is suggested.