Main Subcommittee Category
Policy, Standards, Management, and AcquisitionHow would you label your submission?
GovernmentAbstract
Data standards govern how digital data are formatted, organized, and stored to facilitate later use. The training value gained through the application of data standards has long-been acknowledged Standards have enabled real-time performance feedback and rigorous training research, and the application of standards has poised the operational training community to benefit greatly during the current explosion of artificial intelligence (AI) capabilities. Doing so, however, assumes that training systems are built with the capability to support seamless data manipulation and export, which analyses have shown is not the case (NAWCTSD & Katmai, 2023). Data trapped within a training management or after-action review system negates the potential of current computational advances. Complexities in the requirements process and coordination of training system requirements for acquisition, including cross-service collaboration, are acknowledged across the services (Marler et al., 2021; NAWCTSD & Katmai, 2023). Organizations must proactively identify data requirements during the design and development of simulation-based training and after-action review systems to realize the full value of their digital training data. Program managers considering open systems architecture and various data strategies have resources to guide them (e.g., Defense Acquisition University, 2013; Guertin & Hurt, 2013), however, these resources fall short of articulating needs for specific data types and analyses goals to support data-driven learning analytics. This paper addresses this gap. Specifically, we discuss how audio, video, simulation trace, and other multimodal data should be collected and formatted to support training analytics with emerging AI tools and techniques. Illustrated in the context of military medical training, these considerations are applicable in other domains. These recommendations can be leveraged by program managers looking to avoid roadblocks preventing efficient and effective use of their individual and team-level human performance data to inform training and operational decisions.