Metagraph Approach as a Data Model for Cognitive Architecture

Abstract

There is no doubt that cognitive architecture model must be capable of describing very complex hierarchical systems. The basic lower-level “lingua franca” model is required for cognitive architecture description. We propose to use a metagraph model as a basic model. The key element of the metagraph model is metavertex which may include inner vertices and edges. From the general system theory point of view, a metavertex is a special case of the manifestation of the emergence principle, which means that a metavertex with its private attributes and connections become a whole that cannot be separated into its parts. The metagraph agents are used for metagraph transformation. The distinguishing feature of the metagraph agent is its homoiconicity, which means that it can be a data structure for itself. Thus, a metagraph agent can change the structure of other metagraph agents. The structures similar to neural networks are typical elements of cognitive architecture. It is shown that proposed metagraph approach is suitable enough to represent neural networks.