Use of Natural Language Processing to Extract Technical Competency Frameworks from Maintenance Task Analyses
(Room 320A)
03 Dec 19
5:00 PM
-
5:30 PM
Tracks:
Full Schedule, Tuesday Schedule
Technical competencies focused on maintenance, operations, and troubleshooting tasks on engineered systems must be derived from trusted, authoritative data sources, such as system tasks analyses. Those trusted, authoritative data sources are constantly changing caused by engineering design changes. In this type of dynamic enterprise data environment, competencies must reliably link to systems, people, and their work. These data types linked together are prime sources for human and system performance analysis. In the Navy, technical competencies are not linked to authoritative data sources causing technical curriculum to become latent and disconnected from the supported system. The heavy impact on readiness is a costly effect that has resulted in untrained sailors and mismanaged content. To address this issue, the Navy and Credential Engine signed a Cooperative Research and Development Agreement (CRADA) to map the GEIA 0007 and the S3000L logistical support analysis specifications to the Credential Transparency Description Language (CTDL) specification to tie maintenance requirements to competency models. Through the use of natural language processing, the CRADA team developed software that converts the specifications and the inherent content into Linked Data formats, extracts key information from the specification, then molds that content into the syntax of terminal learning objectives (audience, behavior, condition, degree) structured in the CTDL. The resulting technical competency framework mirrors the product structure and the associated tasks in the logistics specification and is linked through unique system identifiers forming a "digital thread". The software then analyzes the competency framework in the CTDL and ouputs a corresponding course outline in S1000D, an international technical manual specification. The time saved on manual job duty task analysis allows for the same analysts to be part of an iterative cycle of reviews and approvals of competency frameworks from authoritative sources. The learning is binded to the work through data standards, which in turn allow a faster identification of curriculum impacted by engineering design changes. This paper describes the process.