Abstract:
The search for theoretical bases for modeling the semantics of attributive relations is an urgent task,
the solution of which will contribute to the creation of a universal formal description of concepts and relations between them. The application of a formal approach based on neurocognitive architectures opens the
prospect of investigating the possibilities of automatic ontologies construction on the basis of multi-agent
self-organization.
A hypothesis is proposed and a formal description of the neurocognitive architecture is developed, oriented on imitation modeling of the lingvosemantic function of understanding the word combinations of a
nominal group on the basis of multiagent self-organization and interactive learning.