Enhanced Object-Based Production Conference

From NCOR Wiki
Revision as of 15:03, 4 December 2019 by Dgo2112 (Talk | contribs)

Jump to: navigation, search

Event Date and Venue

May 13 – 14, 2020

SAIC Rosslyn Office 1901 Ft. Myer Drive Arlington, VA 22207

Organizer and Contact

David G. Limbaugh dglimbau@buffalo.edu

Conference Description

Conference Purpose: This conference aims (1) to identify the lessons learned from the Referent Tracking methodology, (2) to explore how these lessons might be translated to the domain of intelligence analysis, (3) to identify potential benefits relating to semantic technology, and (4) to explore how these benefits can enhance OBP.

The Problem: Ever increasing quantities of disaggregated data pose a problem for intelligence analysis. The problem is magnified when much of the data is sparse, obscure, or ever-changing. A key contributor to this problem is the inconsistency of data management policies. Mutually incompatible data management solutions have been and continue to be adopted not only by organizations but also by different departments within organizations. Data, as a result, is difficult to aggregate, and difficult to discover and to interpret, sometimes even difficult to interpret by those who created the data in the first place. This problem of too-much data inconsistently handled has an analogue in medicine in the realm of patient data. Electronic Health Records, for example, are collections of data about patients which grow and change in ways which make it difficult to track the medical state of a patient as it changes over time, for example as patients move between hospitals.

The Research: Intelligence data is useful only if it is available to decision makers when they need it. The subject of this conference is an Intelligence Community (IC) inter-organizational data architecture that has the potential to rapidly handle the enormous amounts of data collected continuously by the IC. The data architecture –– called ‘Enhanced Object-Based Production’ (E-OBP) –– is based on the Referent Tracking (RT) approach developed and tested in the medical domain over some 15 years [1]. E-OBP takes the object-oriented approach of Object-Based Production (OBP) but expands ‘object’ to any salient portion of reality. This allows traditional OBP to be transformed into an expressive, flexible, and scalable, data architecture.

The governing principle of E-OBP is to structure data that objectively mirrors reality in a way that allows tracking. Reality is made of unique entities with shared features and relationships indexed to locations and times. E-OBP uses 1) unique identifiers to refer to unique entities, 2) terms from a controlled vocabulary to represent features, relationships, times, and places, and 3) time-indexed, first-order logic expressible, assertions to represent when an entity has some feature or some relationship to other entities.

E-OBP applies not only to data about first-order reality – tanks, people, missions, economic transactions, and so on – but also to data about these data, which it tracks using the same information infrastructure. It tracks when data become available, who made it available, the methods by which it was obtained, and whether it should be trusted [2]. It also records when data is discovered to be inaccurate, in a way that allows for more sophisticated error tracking. All of these data are brought together by the system to form a gigantic evolving graph, which forms a comprehensive and continuously adjusted picture of reality structured to allow zooming on identified threats, sensitive areas, government actions, and so forth.


Barry Smith (University at Buffalo)

David G. Limbaugh (University at Buffalo)

Ron Rudnicki (CUBRC, Inc.)

Timothy Lebo (Air Force Research Laboratory)

Werner Ceusters (University at Buffalo)

William Hogan (University of Florida)

William S. Mandrick (SAIC)