Modeling reflection in artificial intelligence systems: state of art and prospects

Abstract

Abstract. The paper is devoted to analyzing the current state of research and assessing the prospects for modeling reflexive processes using artificial intelligence systems, in particular, multi-agent systems for relevant simulation of collective problem solving. Reflexive modeling of each other by agents will ensure the development of a coherent model of the control object, the purpose of collective work and norms of interaction, as well as the developing of effective interaction between agents. This will allow the system, by self-organizing, to adapt to the characteristics of arising problems, in particular, to take into account their complex structure, the network nature of conditions and goals, opacity, subjectivity and dynamism.