Grounding Crisp and Fuzzy Ontological Concepts in Artificial Cognitive Agents

摘要

The grounding of crisp and fuzzy concepts is defined for the class of artificial cognitive agents equipped with means for recognizing states of external objects and language signs. Each concept is created in the internal cognitive space of the agent in the context of social communication. This communication makes it possible to correlate cognitive agent's perceptions with language signs generated by other members of the same population of agents. Related processes yield the so called semiotic relation consisting of recognized instances of semiotic triangles. A simple measure between concepts is suggested and used to define higher level relations of semantic synonymy, similarity, generalization and contradiction. In consequence an approach to modeling ontology creation for a particular class of agents and a particular language is defined.