Some aspects of artificial personality architecture

Abstract

This paper examines the possibility of artificial personality as a form of strong AI and as a model of the human psyche. It is argued that connectionism is not relevant for creating an artificial personality – brain is a hybrid system, which, in addition to self-learning circuits, contains algorithmic procedures. A hypothesis of the functional role of the awareness of self-consciousness property as a reflexive rank, since genuine intelligence arise, is proposed. AI as a model can be built as a hybrid multi-agent system. Some circuits may be analog, others – digital; some systems may be self-learning, others may operate under algorithmic rules and mathematical functions. The success of perceptronical inclines to idea that sensation modeling can be self-learning. The level of emotions can be realized as primary representations in the analog form. The later may be abstracted in judgments treated as secondary representations. The level of self-consciousness or meta-psychological states implies certain possibility of learning to self-learning (as an analogue of awareness of self-awareness) using secondary representations at the input.