Program Summary
Military and civilian organizations have deep interest in human performance optimization (HPO). A key challenge for optimizing human performance, however, is the “tyranny of averages:” a common experimental approach that uses between-subject outcomes and group averages (means) to make conclusions about the efficacy of a given intervention. This approach frequently (mis)characterizes individual variance as statistical “noise,” “residuals,” or “error.” The resulting interventions (e.g., diet, physical training regimen, brain stimulation) are, at best, suboptimal and, at worst, deleterious for each person. Current AI capabilities for identifying, characterizing, and modeling human performance interventions are limited for similar reasons as most approaches likewise treat individual differences as residual error, particularly because accounting for individuals can significantly increase dimensionality of the datasets.
To address this challenge, the TAILOR program is exploring whether and, if so, to what extent AI methods can enhance prediction about which HPO intervention(s) will be most effective for different individuals and teams. In particular, TAILOR seeks to test third wave AI tools and approaches that can use contextual reasoning to make counterfactual predictions about HPO interventions and better leverage individual variability (i.e., if person X had been given intervention Y, then they would have had outcome Z). By incorporating biological, psychological, and social factors that give rise to individual differences, TAILOR hypothesizes that successful approaches will be able to reason over multiple factors in order to better “tailor” individualized HPO outcomes, adapt if those factors change, and continue to help the DoD leverage diversity as a strength.