webbanner infodatamodeling

Controlled Natural Language-based Knowledge Inter-exchange Network (CKI-NET)

The CKI-NET initiative is researching, designing and implementing a natural language-based knowledge inter-exchange network that will minimize the effort needed to gather and analyze raw intelligence data and to produce a fused, fully vetted intelligence assessment.

KBSI’s CKI-NET will use controlled natural language (CNL) and system wide ontologies as the primary communication scheme for exchanging information among warfighters and intelligence service providers. These communications occur across nodes in a network that, because of their dynamic warfighter context, can be ad-hoc, disconnected, intermittent, and limited (DIL).

CNLs make use of simplified syntactic and semantic structures that are taken from the discourse of the domain and the taskings required of the in-play technology.CNL’s simplified structures retain the expressiveness of the root natural language while removing ambiguity in the linguistic structures allowed within the language. Removing the ambiguity allows the CNL structures to be deterministically processed by computer systems, addressing the challenges inherent in automatically processing sensor data and automatically and efficiently processing and routing intelligence requests within these erratic networks. These capabilities will allow the CKI-NET to produce actionable information automatically from the Information Exchange Requirements (IERs) and make this information usable by the intelligence analyst at the time of receipt.

The current intelligence gathering practice requires that analysts manually generate IERs, including specifying the operation task requiring the intelligence; the Intelligence, Surveillance, and Reconnaissance (ISR) assets available to fulfill the request; the proper information description characterizing the type of ISR request; and the attributes of the data that will support the analysis. The collected intelligence data is sent to the originating analyst who must next manually (i.e., using stand alone applications) perform the analysis and writes the intelligence report.

In CKI-NET, analysts construct a CNL statement that expresses the required intelligence with the aid of a predictive editor that checks and refines the analyst’s CNL statements with respect to the CKI-NET grammar. The refined CNL statements abstract the remaining information, making it unnecessary for analysts to know the details of operational elements, IER information descriptions, and other IER attributes. Having IERs and ISR sensors defined using the CKI-NET ontology (via CNL statements) provides machine understandable expressions of IERs and ISR sensors. The CNL representations make the CKI-NET semantic model easy to work with and understand, and make it possible for nodes in the CKI-NET network to communicate with other nodes about their own state and the state of the network. A single, unified CNL grammar will allow ISR products to be understood by humans, be interoperable, and be synchronized across all DoD forces and other mission partners.

The CKI-NET ontologies and CNL grammar are the central enablers supporting a move from the current net-centric strategy to a data-centric strategy. In a data-centric strategy, data drives the virtual network topology and the application of resources to achieve outcomes. The CKI-NET semantic model provides the mechanisms needed to optimize network bandwidth usage adjusting the fidelity of ISR requests. The CKI-NET semantic model enables the communication of the information needed by fidelity adjusting algorithms to choose how to execute ISR requests and achieve optimization of network throughput while satisfying analyst requests. CKI-NET’s machine understandable representations for ISR sensors and IER structures will make this type of real-time network optimization a reality.

In addition, the use of CNL also addresses the greater influx of data that results from (1) new sensors being added to the network and (2) technological upgrades in the sensors that provide a higher fidelity output. CKI-NET provides technology for reducing the amount of data fed to the network.In addition to improving data processing for the intelligence analyst, the CKI-NET will also provide a high-level of flexibility in accommodating network growth and evolution to meet future intelligence gathering challenges.