There are numerous factors that can diminish the effectiveness of training within the military. Until recently we have often thrown away or passed up valuable data that could have provided massive advantages when making training decisions.  

This brings up the question, can training effectiveness be improved through the implementation of a new data strategy approach that complements existing training?  Developing training from scratch is time demanding, labor intensive, and costly. However, teaming Artificial Intelligence (AI) with the development of data doppelgangers can provide real-time visibility into learner proficiency during a course and throughout their career. A data doppelganger (also known as data double, digital twin, digital shadow) is the conversion of human bodies and minds into data flows that can be figuratively reassembled for the purposes of personal reflection and interaction (Ruckenstein, 2014). A data doppelganger extends the typical data collected and integrates more granular interactions from multi-modal sources. These rich data sets are then translated into comprehensive context aware models built from course objectives, Subject Matter Expert (SME) experience, and domain constructs. Furthermore, teaming AI algorithms designed for these types of data sets allows dynamic training recommendations to target specific types of instructional strategies (e.g., desirable difficulties, contrasting cases, stress exposure) and even account for knowledge and skill decay.  

For example, actions within a virtual environment can be measured and assessed in tandem with real-time physiological data to compute a dynamic representation of the learner’s states, or data doppelganger. With this information, the intensity of a scenario delivered via a stress exposure instructional strategy can either be dialed up or down depending on the performance and physiological state of the learner. In this paper, a conceptual framework for developing data doppelgangers will be presented alongside several possible use cases, recommendations, and considerations.