Enabling contextualized ontology modeling with a collaborative multi-view system

Abstract

Ontology modeling is an essential task for developing symbolic AI in which domain experts or knowledge engineers define taxonomies of concepts and semantic relationships to represent a particular domain's knowledge. The effective integration of symbolic knowledge with non-symbolic content enables richer knowledge representation and reasoning, and more explainable and efficient training of machine learning models used across multiple AI applications. This demo's main goal is to present how the Knowledge Explorer System (KES) multiview architecture enables contextualized ontology modeling over a hybrid knowledge representation by making it easier to create and visualize taxonomies and relationships defined in a specific context. KES ontology modeling view supports efficient modeling of contextualized ontologies while its structural view supports their exploration, curation, and integration with non-symbolic content.