Publication
Modelling Episodes with Generic Ontology Design Patterns
Bernd Krieg-Brückner; Mihai Codescu; Mihai Pomarlan
In: Karl Hammar; Oliver Kutz; Anastasia Dimou; Torsten Hahmann; Robert Hoehndorf; Claudio Masolo; Randi Vita (Hrsg.). Proceedings of the Joint Ontology Workshops. Joint Ontology Workshops (JOWO-2020), SKALE: Workshop on Scalable Knowledge Graph Engineering, located at Bolzano Summer of Knowledge (BOSK 2020), August 31 - October 7, Bolzano & Virtual, Italy, No. 2708, CEUR-WS.org, 2020.
Abstract
Developing knowledge-driven applications requires a mix of
competencies; however, ontology experts are often not available to handle
the foundational aspects of knowledge modelling, and end-users may
be prone to introducing mistakes. As a solution, Generic Ontology Design
Patterns, GODPs, encapsulate complex semantics; during reuse, instantiations
provide handles for checking arguments against structural
and semantic constraints stated in ontology parameters. Development
is divided according to expertise: ontology experts develop GODPs while
domain experts concentrate on their application in a domain; the safety
of modelling and of data input for end-users is significantly increased.
Safe ontology engineering with GODPs is demonstrated with episodes,
a significant extension of DUL patterns, for a use case in service robotics:
GODPs for narratively enabled episodic memories provide logical consistency,
and safe population with data substantially scales up.