Publications

This report presents the findings of NATO Information Systems Technology (IST) Exploratory Team 111 (ET-111) regarding the status of Knowledge Representation and Reasoning (KRR). Knowledge representation is the expression of knowledge that is computer-tractable, able to be reasoned over, and exploitable – especially in the age of big data. This report provides a technical introduction to the field of KRR. The capabilities of knowledge systems are discussed, as well as how they can be created, and how they may then be applied to real-world military problems. The distinction is made between KRR (‘symbolic’ AI) and machine learning (‘sub-symbolic’ AI), along with a consideration of how the two approaches can be complementary. We discuss some of the challenges for NATO member nations, and how KRR can influence these areas. We present specific examples of knowledge representation including the Multilateral Interoperability Program (MIP) Information Model (MIM), the Rich Event Ontology (REO) and the Defense Intelligence Core Ontology (DICO). Moreover, we review areas of research including causality, explainability, and trust in KRR systems. Finally, we discuss implications for NATO, and present our key recommendations for future work involving technical activities and exploratory teams.