System analysis, management and information processing
Classification and conditions of application of algorithms
for automatic ontologization of the state space of a general artificial
intelligence agent under the control of neurocognitive architecture
Abstract:
The purpose of the study is to form an algorithmic base for ontologizing the state space of
a general artificial intelligence agent. The main objectives of the work are the classification of
algorithms for the autonomous synthesis of ontologies of the “agent-environment” system and the
selection of conditions for their application. The form and contents of the main ontologization
algorithms for the state space of an intelligent agent of artificial general intelligence are determined and
their classification is constructed. The mechanisms for the formation of concepts of functional
representation of entities in the state space of an intelligent agent based on dynamic generation “on
demand” are substantiated. A meta-algorithm for selecting private ontologization algorithms has been
constructed depending on the selected application conditions.
Keywords:automatic construction of ontologies, artificial general intelligence, multi-agent systems,neurocognitive architectures