fy17 utilizing machine learning and artificial intelligence for medical training needs (mach learning) award
the fy17 jpc-1/msis mach learning award supports research to determine, define, and validate predictive machine learning modeling systems that predict early stages of training needs in order to eliminate the current culture of just-in-time training or not being able to offer refresher training in an efficient manner. the intended outcomes are relevant for military and general public training and educational purposes.
the fy17 jpc-1/msis mach learning award seeks metrics/evaluation criteria, the definition of those metrics/evaluation criteria, and specific measuring tools to obtain (collect) and potentially analyze the metrics/evaluation criteria needed to develop a sufficient algorithm for use in the prediction of training needs, both in the skill acquisition phase and in refresher training. it is also seeking a proof-of-concept and/or prototype (depending on award vehicle) in which the proposed predictive model works together with incoming data/information from different, well-evidenced sources. the models need to use real data/information/knowledge and need to compare against known guidelines within the relative medical discipline and a comparable, peer-level group.
the outcomes of the research will allow data/information/knowledge into the proof-of-concept and/or prototype (depending on award vehicle) model. the model needs to be tested in a laboratory-type environment, preferably by a subawardee with no active participation in the development of the model. actual interfaces will need to be described and defined in the outcome, but functionality of the entire defined interface does not need to be demonstrated in the delivered proof-of-concept and/or prototype (depending on award vehicle).
submissions to the fy17 jpc-1/msis mach learning award should not only include the design and the methodologies to create the mach learning proof-of-concept/prototype, but should also include data/information that will or could be used as input into or output from a machine learning and artificial intelligence model.