Universal expert system based on ontoepisosociophylogenetic training of federations of intelligent neurocognitive agents

The work is devoted to solve a scientific problem of developing a conceptual justification for the possibility of autonomous training of intelligent expert systems based on ontoepisociophylogenetic training of neurocognitive agents. The aim of the study is to develop basic principles of creating universal expert systems based on ontoepisociophylogenetic training of federated intelligent neurocognitive agents. The basic principles of ontoepisociophylogenetic training of universal federated expert systems have been developed. It is shown that the functional specialization of intelligent agents within a federation, subject to their cooperation in order to maximize the combined increment of the values of the target functions, allows overcoming efficiency limitations. The use of epigenetic algorithms for fixing ontological knowledge of intelligent agents within a federation in generations of evolutionary optimization is substantiated. The possibility of constructing multi-generational populations in order to increase the overall efficiency of a universal expert federated system is substantiated.