RSS feed Add to your Facebook page Watch us on Youtube

Activity title

Knowledge Representation and Reasoning

Activity Reference

IST-ET-111 (AI2S)

Panel

Information Systems Technology

Security Classification

NATO UNCLASSIFIED

Status

Planning

Activity type

ET

Start date

2019

End date

2020

Keywords

Decision Making, Graph Theory, Knowledge Management, Semantic Information Processing

Background

As the number of sensors, feeds and other data sources available to coalition operations increase, the focus of technology development is turning to ensuring such sources process data in-situ and provide only useful structured/pertinent observations, not least to deal with limited and intermittent network connectivity. Recently, IST-144-RTG has successfully demonstrated the potential for ‘content based analytics’, including the automated recognition of military hardware in video feeds and the extraction of significant entities in open reporting. The destinations for many of those observations will be screens and dashboards that seek to provide situational awareness for human analysts. However, as the number of observations increase and the interconnections of those observations become more abstract the capacity for human analysts to make good decisions is likely to continue to be overstretched. For example, as operations look to routinely apply full spectrum effects, the interconnection of observations from the physical, social and cyber domains of operations will be increasingly important, but connections between observations in each may not be easily discernible without supporting analytical capabilities. The integrated exploitation of observations, and therefore the representation of information and knowledge, is at the heart of concepts like “Information Advantage” and in achieving the higher level data fusion identified in the JDL Fusion Model . The real challenge now, therefore, is to be ready to integrate those observations to support situational awareness and understanding across an entire operation, enabling human/machine teams at all levels of command and across multi-national coalitions. In a coalition context this complexity also includes cultural and linguistic dimensions.

Objectives

Graph based methods, including so called “knowledge graphs” now offer the capacity to integrate multiple observations, supporting concepts like Activity Based Intelligence. However, the concept in Defence is immature and open to wide interpretation and differing approaches. Nonetheless, in other domains technology development and application continue at pace. Centres of excellence on knowledge engineering and reasoning, many with a focus on non-Defence applications, exist across NATO nations. If such approaches/technologies are to support future coalition operations, joined up effort is required across a number of themes to ensure capabilities are aligned.

Topics

This includes: • Development of domain specific ontologies • Building effective architectures for event driven processing • Deployment and assessment of inferencing capabilities • Handling uncertainty and ambiguity in observations • Information sharing and observation provenance • Approaches for federated deployment and coping with scale • Standards and interoperability

Contact Panel Office