Advanced Self-Learning Ontologies

This proposal proposes a system, CLEO – Collective Learning Environments for Ontologies, for supporting adaptive, self-learning of ontologies. It addresses three challenges: the design of an adaptive system, tracking lineage of new information, and ontology fusion. The system will implement a collective learning environment for developing new ontologies and extending existing ontologies. Collective learning harnesses the learning abilities of multiple learning cells organized topologically to conform to the natural meshlike structure of domain concepts. CLEO will be a rule-based system that will use rules to reason about new and existing concepts, determine relationships between them, and create or extend the mesh representing the problem domain’s concepts. It will also track the lineage of new or refined concepts and their attributes as they are added to the ontology KB and provide for versioning and visualization of the ontology mesh over its evolution.