Algorithm for modeling the cognitive function of the emotional assessment of situations based on the training of multi-agent neurocognitive architectures

To create a common artificial intelligence, it is necessary to create an apparatus of cognitive function of emotional assessment, which would be a driving force that can initiate, direct and regulate the system. The paper presents an algorithm for modeling the cognitive function of emotional assessment of situations based on the training of multi-agent neurocognitive architectures. A negative emotional assessment of the starting situation mobilizes the system to find the optimal solution to achieve the result, i.e. the desired situation will be formulated in order to avoid getting into a similar situation the next time, which will lead to training by changing knowledge in the conditional part and contractual relations between agents in the cognitive blocks of multi-agent architecture. A positive assessment is accompanied by increased “release” of additional energy to the agents, which leads to the strengthening of ties in the form of contracts that were concluded to achieve the goal in the cognitive blocks of multi-agent architecture.