Philosophical foundations of intelligence collection and analysis: a defense of ontological realism

ABSTRACT

It is a common belief within the Intelligence Community (IC) that data residing in disparate information systems can be mined in useful ways by means of artificial intelligence (AI) and natural language processing (NLP) methods working alone, which is to say, without the aid of some kind of integrating framework. Here, in contrast, we argue that the sort of integration and analysis that is required if we are to connect data and information deriving from heterogeneous sources in useful ways needs semantic integration, in other words integration that rests on the ability to identify shared meanings across different bodies of data. To achieve such integration requires what we shall call an Integrating Semantic Framework (ISF). A framework of this sort is based on ontologies, which are controlled structured vocabularies designed to foster interoperability in the collection and curation of data and thereby to prevent the sorts of siloing of information that arise where there is inconsistency in the use of terms.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. A data silo is a collection of data held by one group that is not easily or fully accessible by other groups in the same organization (“What are Data Silos?”, https://www.talend.com/resources/what-are-data-silos Accessed February 21, 2022).

2. The Oxford English Dictionary defines Artificial Intelligence as the development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.

3. The Common Core Ontologies (CCO) comprise eleven ontologies that aim to represent and integrate taxonomies of generic classes and relations across all domains of interest. Accompanying these ontologies is a rule-based method for representing the content of any data source whatsoever through constructing domain ontologies as extensions of CCO. See National Institute of Standards and Technology (NIST) “Document: An Overview of the Common Core Ontologies”.

4. Győri, “Basic Level Categories,” 149.

5. Ibid., 154.

6. Ibid., 149.

7. Smith and Ceusters, “Ontological Realism,” 139–188.

8. Smith, “Beyond Concepts,” 1.

9. Johnston et. al., “Transforming Defense Analysis,” 13.

10. Ibid., 13.

11. Deputy Secretary of Defense, “Defense Data Strategy”, Executive Summary.

12. Information Management (IM) is defined in Joint Publication 3–0, Joint Operations and Joint Publication 3–33 Joint Task Force Headquarters.

13. Chairman of the Joint Chiefs of Staff, Joint Operations, GL-10.

14. Chief Information Officer, DoD Net-Centric Data Strategy, 2003.

15. Referent Tracking, the counterpart of OBP developed in the field of biomedicine, attempts to organize medical and other data around the real-world entities – such as patients, doctors, images, tumors, clinical notes, etc. This is achieved by providing unique identifiers for such entities along the lines described in Ceusters and Smith “Ontological Realism”. Implications for intelligence analysis are outlined in Limbaugh et al. “Warranted diagnosis”.

16. Johnston, Catherine, “Transforming Defense Analysis,” 13.

17. Armstrong, “What is Digital Twin?” Derived from IBM Definition.

18. Digital twins work well for isolated items, say factories; but not where the twinned items get involved with each other. If we create say an instance digital twin for each Main Battle Tank instance, and then the tanks get involved in different virtual environments (e.g., virtual battles), each digital twin will have to be maintained separately, which will be very hard to do.

19. Cox et. al., “The Space Object Ontology”.

20. Smith et. al., “Information Artifact Ontology,” 33.

21. The digital thread refers to the communication framework that allows a connected data flow and integrated view of the asset’s data throughout its lifecycle across traditionally siloed functional perspectives. See: Leiva, “Demystifying the Digital Thread and Digital Twin Concepts” https://www.industryweek.com/technology-and-iiot/systems-integration/article/22007865/demystifying-the-digital-thread-and-digital-twin-concepts Accessed February 21, 2022.

22. Boone et. al., “Interoperability and Integration,” 2.

23. The W3C Web Ontology Language (OWL) is a Semantic Web language designed to represent rich and complex knowledge about things, groups of things, and relations between things. OWL is a computational logic-based language such that knowledge expressed in OWL can be exploited by computer programs, e.g., to verify the consistency of that knowledge or to make implicit knowledge explicit. See: https://www.w3.org/OWL.

24. On best practices for the formulation of definitions in ontologies see Seppälä and Smith, “Guidelines for writing definitions in ontologies.”

25. Arp et. al., Building Ontologies, glossary.

26. Ibid., 38.

27. Ibid.

28. The Common Core Ontologies are described in Rudnicki, Ron, “An Overview of the Common Core Ontologies”, https://www.nist.gov/system/files/documents/2019/05/30/nist-ai-rfi-cubrc_inc_004.pdf.

29. Deputy Secretary of Defense, Data Sharing, E1.1.17.

30. Gollu and Deshpande, Object Management Systems, 5494.

31. The Five Eyes brings the UK, the United States, Canada, Australia and New Zealand into the world’s most complete and comprehensive intelligence alliance. See: The Five Eyes – The Intelligence Alliance of the Anglosphere (ukdefencejournal.org.uk).

Additional information

Notes on contributors

Bill Mandrick

Bill Mandrick is an Intelligence Analyst Chief and Senior Ontology Advisor at Science Applications International Corporation (SAIC). He is also a retired U.S. Army Officer and former Senior Research Fellow at the Joint Special Operations University (JSOU) at U.S. Special Operations Command.

Barry Smith

Barry Smith is an ontologist working at the University of Buffalo. He is Director of the National Center for Ontological Research and principal editor of Basic Formal Ontology (ISO/IEC 21838-2), which is used as integrating architecture in some hundreds of ontology initiatives throughout the world.