Models are abstract representations of things of interest. Knowledge models, also known as ontologies, are made up of categories, properties and relationships (technically called classes, data properties and object properties). These knowledge models capture understandings, beliefs or system definitions.
Knowledge models are primarily used for machine-based reasoning in the artificial intelligence community. In the Thetus Publisher world, models are used throughout the system to articulate structure and beliefs, filter views and perspectives, capture processes and workflow, and drive the user interface. The Publisher’s model-driven architecture facilitates rich inference at search time, enabling the discovery of emerging and novel relationships and patterns.
Knowledge models are essential to attaining an understanding of complex systems in which interconnected participants see systems differently, interact at varying levels of detail and are motivated to influence the system in ways that could impact other participants. Because of the complex and shifting nature of discovery, knowledge models need to change to adapt to shifting dynamics and circumstances. For this reason, changes to the model are tracked to show how understanding is evolving. Since models are constantly evolving, they are put into dictionaries (technically called namespaces) where policy is articulated. Policy facilitates productive collaboration and preserves lines of reasoning by capturing users’ roles in modifying the models.