Outline of a Brain Model for Self-Observing Agents

Abstract

Brain-inspired models for conscious robots should refer to the cellular double structure of the brain, consisting of the neuronal system and the glial system, embodying two ontological realms. Therefore, a purely neurobiological approach to machine consciousness is biased by an ontological fault in exclusively referring to the neuronal system. The brain model for self-observing agents outlined in this paper focuses on the glial-neuronal synaptic units (tripartite synapses). Whereas the neuronal component of the synapse embodies objective subjectivity processing sensory information, the glial component (astrocyte) embodies subjective subjectivity generating subjective behavior (intentions, consciousness) in its interactions with the neuronal part of the synapse. The elementary principle of the implementation of self-observing agents is this: a brain is capable of self-observation, if the concept of intention to observe something and the concept of the observed are located in different places. Based on a formalism of qualitative information processing, the architecture of self-observation is described in increasing complexity, building networks. It is suggested that if a robot brain is equipped with a network of modules for self-observation, the robot may generate subjective perspectives of self-observation indicating self-consciousness.